Release [run release-20261109T032054Z]

#2
by Battam - opened
This view is limited to 50 files because it contains too many changes. See the raw diff here.
Files changed (50) hide show
  1. CHANGELOG.md +98 -0
  2. NOTICE +12 -0
  3. README.md +68 -58
  4. README.zh.md +68 -58
  5. campaign_files/2026-10-b-disk-images-3/btrfs-progs-convert-needs-recovery/README.md +0 -3
  6. campaign_files/2026-10-g-solver-instances-2/2dpacklib/README.md +0 -6
  7. campaign_files/2026-10-g-solver-instances-2/asiekierka-free-game-levels/README.md +0 -40
  8. campaign_files/2026-10-g-solver-instances-2/berthier-sudoku-classif-dob39c/README.md +0 -4
  9. campaign_files/2026-10-g-solver-instances-2/berthier-sudoku-classif-eleven-te2/README.md +0 -4
  10. campaign_files/2026-10-g-solver-instances-2/berthier-sudoku-classif-royle-17c/README.md +0 -4
  11. campaign_files/2026-10-g-solver-instances-2/biq-mac-library/README.md +0 -4
  12. campaign_files/2026-10-g-solver-instances-2/biqmac-library/README.md +0 -16
  13. campaign_files/2026-10-g-solver-instances-2/borroot-akari/README.md +0 -12
  14. campaign_files/2026-10-g-solver-instances/cblib/README.md +0 -4
  15. campaign_files/2026-10-gt-browser-tests-2/allenporter-ical-recurring-single-change/README.md +0 -4
  16. campaign_files/2026-10-gt-browser-tests-2/apache-poi-formula-eval-testdata/README.md +0 -4
  17. campaign_files/2026-10-gt-browser-tests-2/apache-poi-function-testdata/README.md +0 -44
  18. campaign_files/2026-10-gt-browser-tests-2/biweekly-issues/README.md +0 -6
  19. campaign_files/2026-10-gt-browser-tests-2/browsertime-har-unit-tests/README.md +0 -6
  20. campaign_files/2026-10-gt-browser-tests-2/browsertime-trace-unit-tests/README.md +0 -22
  21. campaign_files/2026-10-gt-browser-tests-2/browsertime-visualmetrics-tests/README.md +0 -4
  22. campaign_files/2026-10-gt-browser-tests-2/calamine-tests-ods/README.md +0 -4
  23. campaign_files/2026-10-gt-browser-tests-2/calamine-tests-xlsx/README.md +0 -14
  24. campaign_files/2026-10-gt-browser-tests-2/chrome-devtools-mcp-heap-fixtures/README.md +0 -6
  25. campaign_files/2026-10-gt-browser-tests-2/chrome-har-perflogs/README.md +0 -4
  26. campaign_files/2026-10-gt-browser-tests-2/chrome-recorder-cypress-export-tests/README.md +0 -4
  27. campaign_files/2026-10-gt-browser-tests-2/chrome-recorder-nightwatch-export-tests/README.md +0 -4
  28. campaign_files/2026-10-gt-browser-tests-2/chrome-recorder-webdriverio-export-tests/README.md +0 -4
  29. campaign_files/2026-10-gt-browser-tests-2/chromium-bidi-unit-tests/README.md +0 -10
  30. campaign_files/2026-10-gt-browser-tests-2/chromium-policy-pref-mapping/README.md +0 -18
  31. campaign_files/2026-10-gt-browser-tests-2/coach-core-merge-tests/README.md +0 -4
  32. campaign_files/2026-10-hw-eda-tests-2/bosl2-the-belfry-openscad-library-v2/README.md +0 -8
  33. campaign_files/2026-10-hw-eda-tests-2/buserror-simavr-tests-github/README.md +0 -14
  34. campaign_files/2026-10-hw-eda-tests-2/cadquery-plugins-community-plugin-collection-for-cadquery/README.md +0 -12
  35. campaign_files/2026-10-hw-eda-tests-2/camotics-gcodetool-regression-tests-github-cauldrondevelopmentllc-camotics/README.md +0 -16
  36. campaign_files/2026-10-hw-eda-tests-2/camotics-tplang-regression-tests-github-cauldrondevelopmentllc-camotics/README.md +0 -6
  37. campaign_files/2026-10-hw-eda-tests-2/cocotb-examples/README.md +0 -6
  38. campaign_files/2026-10-hw-eda-tests-2/cocotb/README.md +0 -4
  39. campaign_files/2026-10-m-issue-trackers-2/academy-bus-nj-gtfs/README.md +0 -4
  40. campaign_files/2026-10-m-issue-trackers-2/agl-gerrit-changes/README.md +0 -4
  41. campaign_files/2026-10-m-issue-trackers-2/altlinux-devel/README.md +0 -4
  42. campaign_files/2026-10-m-issue-trackers-2/aomedia-gerrit-changes/README.md +0 -4
  43. campaign_files/2026-10-m-issue-trackers-2/aosp-gerrit-changes/README.md +0 -4
  44. campaign_files/2026-10-m-issue-trackers-2/arin-arin-ppml/README.md +0 -4
  45. campaign_files/2026-10-m-issue-trackers-2/atlassian-public-jira/README.md +0 -8
  46. campaign_files/2026-10-m-issue-trackers-2/atoc-gtfs-2021-archive/README.md +0 -4
  47. campaign_files/2026-10-m-issue-trackers-2/big-blue-bus-gtfs/README.md +0 -4
  48. campaign_files/2026-10-m-issue-trackers-2/bitcoin-dev/README.md +0 -4
  49. campaign_files/2026-10-m-issue-trackers-2/blender-bf-committers/README.md +0 -4
  50. campaign_files/2026-10-m-issue-trackers-2/bluestar-passenger-gtfs/README.md +0 -4
CHANGELOG.md CHANGED
@@ -2482,3 +2482,101 @@ Changes to the content of this index, oldest first. There is one entry for each
2482
  ## 2026-11-05: batches added to the tables
2483
 
2484
  - open-food-facts: batch inc added to meta/sources.json and meta/batches.parquet
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2482
  ## 2026-11-05: batches added to the tables
2483
 
2484
  - open-food-facts: batch inc added to meta/sources.json and meta/batches.parquet
2485
+
2486
+ ## 2026-11-09: source card format version 4 to 5
2487
+
2488
+ - The cards in cards/ now use card_version 5 (previously 4).
2489
+
2490
+ ## 2026-11-09: daily update
2491
+
2492
+ - Sources added (5): california-code-sections, habs-haer-hals, miplib2017-solution-checker, rules-as-code-trees, zbmath-open
2493
+ - Sources removed (0): none
2494
+ - Rows added and removed, by source:
2495
+ - appsec-trees: 375,901 rows added (118,350 to 494,251)
2496
+ - audio-processing-trees: 857 rows added (32,253 to 33,110)
2497
+ - bioinformatics-trees: 52 rows added (11,304 to 11,356)
2498
+ - california-code-sections: 2 rows added (0 to 2)
2499
+ - chemistry-materials-trees: 24 rows added (46,793 to 46,817)
2500
+ - crypto-library-trees: 163 rows added (19,287 to 19,450)
2501
+ - ctan: 13 rows added (5,840 to 5,853)
2502
+ - ctftime: 11 rows added (75,445 to 75,456)
2503
+ - data-gov: 3,111 rows added (819,809 to 822,920)
2504
+ - engine-source-trees: 8,563 rows added (168,057 to 176,620)
2505
+ - european-parliament-plenary: 28 rows added (1,238,704 to 1,238,732)
2506
+ - evidential-deep-learning-repro: 1 rows removed (51 to 50)
2507
+ - fact-ai-dawlau-repro: 1 rows removed (72 to 71)
2508
+ - fact-ai-imandrealombardo-repro: 1 rows removed (28 to 27)
2509
+ - fairness-attacks-repro: 1 rows removed (31 to 30)
2510
+ - fatcat-files: 1,297 rows removed (54,916,041 to 54,914,744)
2511
+ - fedora-koji-build-logs: 3,290 rows removed (647,268 to 643,978)
2512
+ - forensics-tool-trees: 29 rows added (3,808 to 3,837)
2513
+ - formal-math-trees: 2,656 rows added (33,021 to 35,677)
2514
+ - gerrit-code-review: 978,994 rows added (3,082,751 to 4,061,745)
2515
+ - github-release-assets: 64 rows removed (200,503 to 200,439)
2516
+ - govinfo-bulk-data: 106 rows added (420,895 to 421,001)
2517
+ - gpu-kernel-trees: 1,079 rows added (24,497 to 25,576)
2518
+ - habs-haer-hals: 2,102 rows added (0 to 2,102)
2519
+ - hamiltonian-nn-repro: 1 rows removed (12 to 11)
2520
+ - hf-training-records: 960 rows removed (491,135 to 490,175)
2521
+ - homebrew-packages: 3,122 rows removed (81,274 to 78,152)
2522
+ - inference-serving-trees: 6,635 rows added (31,924 to 38,559)
2523
+ - kios-research-leakdb: 109 rows removed (112 to 3)
2524
+ - legacy-interop-trees: 1,539 rows added (20,052 to 21,591)
2525
+ - lichess: 3 rows removed (1,382 to 1,379)
2526
+ - mendeley-data: 273 rows added (16,054,592 to 16,054,865)
2527
+ - miplib2017-solution-checker: 12 rows added (0 to 12)
2528
+ - music-notation-trees: 177 rows added (30,614 to 30,791)
2529
+ - numerical-library-trees: 301 rows added (19,294 to 19,595)
2530
+ - oeis: 2,806 rows added (400,947 to 403,753)
2531
+ - open-food-facts: 8 rows added (4 to 12)
2532
+ - openneuro: 5,822 rows added (7,520,249 to 7,526,071)
2533
+ - osv: 28,782 rows removed (541,288 to 512,506)
2534
+ - owid-charts: 2,763 rows added (967 to 3,730)
2535
+ - pmc-open-access: 5,330,477 rows added (18,904,034 to 24,234,511)
2536
+ - query-engine-trees: 426 rows added (25,715 to 26,141)
2537
+ - robotics-control-trees: 1,046 rows added (28,458 to 29,504)
2538
+ - rules-as-code-trees: 380 rows added (0 to 380)
2539
+ - sec-edgar-filings: 7,977 rows added (18,854,067 to 18,862,044)
2540
+ - smithsonian-open-access: 45,664 rows removed (9,359,112 to 9,313,448)
2541
+ - typesetting-trees: 219 rows added (11,755 to 11,974)
2542
+ - ubuntu-irc-logs: 88 rows added (457,017 to 457,105)
2543
+ - uk-parliament-hansard: 63 rows added (1,371,597 to 1,371,660)
2544
+ - usgs-earthquakes: 668 rows added (4,917,150 to 4,917,818)
2545
+ - zbmath-open: 1,484,538 rows added (0 to 1,484,538)
2546
+ - From this entry on, rows withheld from removed sources are counted as distinct rows (each row once, however many copies of it were withheld); earlier entries counted copies. Distinct rows withheld so far: 0
2547
+ - Rows withheld because they copy an evaluation benchmark: 48
2548
+
2549
+ ## 2026-11-09: batches added to the tables
2550
+
2551
+ - appsec-trees: batch inc added to meta/sources.json and meta/batches.parquet
2552
+ - oeis: batch inc added to meta/sources.json and meta/batches.parquet
2553
+ - openneuro: batch inc added to meta/sources.json and meta/batches.parquet
2554
+
2555
+ ## 2026-11-09: sources held back
2556
+
2557
+ - Each source below keeps its newest version that passed every check; its newer batch waits until it passes them:
2558
+ - adflib-adf-deleted: changed pointer group file data/adflib-adf-deleted/2026-10-b-disk-images-3/part-00000.parquet: 1 key(s) of the old version are in no data file of source adflib-adf-deleted after the commit (row 2 from batch 2026-10-b-disk-images-3); the removal record holds none of them, and none fails a publish check now
2559
+ - adflib-adf-malformed: changed pointer group file data/adflib-adf-malformed/2026-10-b-disk-images-3/part-00000.parquet: 2 key(s) of the old version are in no data file of source adflib-adf-malformed after the commit (row 4 from batch 2026-10-b-disk-images-3, row 5 from batch 2026-10-b-disk-images-3); the removal record holds none of them, and none fails a publish check now
2560
+ - akretion-factur-x-libreoffice-extension: changed pointer group file data/akretion-factur-x-libreoffice-extension/2026-10-a-spreadsheets-with-answers-2/part-00000.parquet: 1 key(s) of the old version are in no data file of source akretion-factur-x-libreoffice-extension after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers-2); the removal record holds none of them, and none fails a publish check now
2561
+ - austender-ict-procurement-statistics: changed pointer group file data/austender-ict-procurement-statistics/2026-10-a-spreadsheets-with-answers-2/part-00000.parquet: 2 key(s) of the old version are in no data file of source austender-ict-procurement-statistics after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers-2, row 2 from batch 2026-10-a-spreadsheets-with-answers-2); the removal record holds none of them, and none fails a publish check now
2562
+ - australia-government-workbooks: changed pointer group file data/australia-government-workbooks/2026-10-a-spreadsheets-with-answers-2/part-00000.parquet: 2 key(s) of the old version are in no data file of source australia-government-workbooks after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers-2, row 2 from batch 2026-10-a-spreadsheets-with-answers-2); the removal record holds none of them, and none fails a publish check now
2563
+ - australia-government-workbooks-noassertion: changed pointer group file data/australia-government-workbooks-noassertion/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 2 key(s) of the old version are in no data file of source australia-government-workbooks-noassertion after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers, row 2 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
2564
+ - basseyisrael-inventory-system: changed pointer group file data/basseyisrael-inventory-system/2026-10-a-spreadsheets-with-answers-2/part-00000.parquet: 1 key(s) of the old version are in no data file of source basseyisrael-inventory-system after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers-2); the removal record holds none of them, and none fails a publish check now
2565
+ - bayesian-siac-filter-repro: changed pointer group file data/bayesian-siac-filter-repro/2026-10-c-repros-2/part-00000.parquet: 1 key(s) of the old version are in no data file of source bayesian-siac-filter-repro after the commit (row 5 from batch 2026-10-c-repros-2); the removal record holds none of them, and none fails a publish check now
2566
+ - bbmh-ap-imex-repro: changed pointer group file data/bbmh-ap-imex-repro/2026-10-c-repros-2/part-00000.parquet: 1 key(s) of the old version are in no data file of source bbmh-ap-imex-repro after the commit (row 2 from batch 2026-10-c-repros-2); the removal record holds none of them, and none fails a publish check now
2567
+ - bc-government-workbooks: changed pointer group file data/bc-government-workbooks/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 21 key(s) of the old version are in no data file of source bc-government-workbooks after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers, row 2 from batch 2026-10-a-spreadsheets-with-answers, row 3 from batch 2026-10-a-spreadsheets-with-answers, 18 more); the removal record holds none of them, and none fails a publish check now
2568
+ - bhardwaj2024-excel-gst-billing-invoice: changed pointer group file data/bhardwaj2024-excel-gst-billing-invoice/2026-10-a-spreadsheets-with-answers-2/part-00000.parquet: 1 key(s) of the old version are in no data file of source bhardwaj2024-excel-gst-billing-invoice after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers-2); the removal record holds none of them, and none fails a publish check now
2569
+ - bkp-2015-heath-street: changed pointer group file data/bkp-2015-heath-street/2026-10-b-disk-images-3/part-00000.parquet: 1 key(s) of the old version are in no data file of source bkp-2015-heath-street after the commit (row 2 from batch 2026-10-b-disk-images-3); the removal record holds none of them, and none fails a publish check now
2570
+ - bonndata-workbooks-cc0: changed pointer group file data/bonndata-workbooks-cc0/2026-10-a-spreadsheets-with-answers-2/part-00000.parquet: 1 key(s) of the old version are in no data file of source bonndata-workbooks-cc0 after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers-2); the removal record holds none of them, and none fails a publish check now
2571
+ - borealis-dataverse-workbooks-cc-by: changed pointer group file data/borealis-dataverse-workbooks-cc-by/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 7 key(s) of the old version are in no data file of source borealis-dataverse-workbooks-cc-by after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers, row 2 from batch 2026-10-a-spreadsheets-with-answers, row 3 from batch 2026-10-a-spreadsheets-with-answers, 4 more); the removal record holds none of them, and none fails a publish check now
2572
+ - borealis-dataverse-workbooks-cc0: changed pointer group file data/borealis-dataverse-workbooks-cc0/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 10 key(s) of the old version are in no data file of source borealis-dataverse-workbooks-cc0 after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers, row 2 from batch 2026-10-a-spreadsheets-with-answers, row 3 from batch 2026-10-a-spreadsheets-with-answers, 7 more); the removal record holds none of them, and none fails a publish check now
2573
+ - btrfs-progs-fsck-raw-images: changed pointer group file data/btrfs-progs-fsck-raw-images/2026-10-b-disk-images-3/part-00000.parquet: 6 key(s) of the old version are in no data file of source btrfs-progs-fsck-raw-images after the commit (row 1 from batch 2026-10-b-disk-images-3, row 2 from batch 2026-10-b-disk-images-3, row 3 from batch 2026-10-b-disk-images-3, 3 more); the removal record holds none of them, and none fails a publish check now
2574
+ - btrfs-progs-fuzz-images: changed pointer group file data/btrfs-progs-fuzz-images/2026-10-b-disk-images-3/part-00000.parquet: 15 key(s) of the old version are in no data file of source btrfs-progs-fuzz-images after the commit (row 1 from batch 2026-10-b-disk-images-3, row 2 from batch 2026-10-b-disk-images-3, row 3 from batch 2026-10-b-disk-images-3, 12 more); the removal record holds none of them, and none fails a publish check now
2575
+ - canada-government-workbooks: changed pointer group file data/canada-government-workbooks/2026-10-a-spreadsheets-with-answers/part-00000.parquet: 2 key(s) of the old version are in no data file of source canada-government-workbooks after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers, row 2 from batch 2026-10-a-spreadsheets-with-answers); the removal record holds none of them, and none fails a publish check now
2576
+ - cdi-import-test-images: changed pointer group file data/cdi-import-test-images/2026-10-b-disk-images-3/part-00000.parquet: 3 key(s) of the old version are in no data file of source cdi-import-test-images after the commit (row 3 from batch 2026-10-b-disk-images-3, row 4 from batch 2026-10-b-disk-images-3, row 5 from batch 2026-10-b-disk-images-3); the removal record holds none of them, and none fails a publish check now
2577
+ - cfpb-complaints: 1 new name hit(s)
2578
+ - cleancooking-protocols-noassertion: changed pointer group file data/cleancooking-protocols-noassertion/2026-10-a-spreadsheets-with-answers-2/part-00000.parquet: 1 key(s) of the old version are in no data file of source cleancooking-protocols-noassertion after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers-2); the removal record holds none of them, and none fails a publish check now
2579
+ - codek11-supply-chain-inventory-analysis: changed pointer group file data/codek11-supply-chain-inventory-analysis/2026-10-a-spreadsheets-with-answers-2/part-00000.parquet: 1 key(s) of the old version are in no data file of source codek11-supply-chain-inventory-analysis after the commit (row 1 from batch 2026-10-a-spreadsheets-with-answers-2); the removal record holds none of them, and none fails a publish check now
2580
+ - concat-stepsize-repro: changed pointer group file data/concat-stepsize-repro/2026-10-c-repros-2/part-00000.parquet: 1 key(s) of the old version are in no data file of source concat-stepsize-repro after the commit (row 1 from batch 2026-10-c-repros-2); the removal record holds none of them, and none fails a publish check now
2581
+ - hf-training-records: 18 new name hit(s)
2582
+ - npm: 82 new name hit(s)
NOTICE ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ This dataset computes its license tiers and reason codes from the license categories of ScanCode LicenseDB.
2
+ They are in the column license of meta/catalog.parquet, the license line of every cards/<source>.md,
3
+ projections.license_unknown of meta/coverage.json, and the two license tiers of meta/sources.json,
4
+ meta/batches.parquet and the rows of data/ (license_training_ok, license_redistribution_ok).
5
+
6
+ ScanCode LicenseDB: https://scancode-licensedb.aboutcode.org/
7
+ Index: https://scancode-licensedb.aboutcode.org/index.json
8
+ Version: the index served with Last-Modified Tue, 03 Nov 2026 16:11:08 GMT; copied unchanged on 2026-11-08; sha256 9b0bdc5cc63a53bcc73bd90c87755a5db6da0096258642832ba54be78e3ad40b
9
+ Copyright (c) nexB Inc. and others. ScanCode is a trademark of nexB Inc.
10
+ License: CC-BY-4.0, the Creative Commons Attribution 4.0 International license (https://creativecommons.org/licenses/by/4.0/)
11
+ ScanCode LicenseDB states: the database and its contents are provided on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND,
12
+ either express or implied; no content from ScanCode LicenseDB should be considered or used as legal advice.
README.md CHANGED
@@ -24,7 +24,7 @@ configs:
24
  <!-- artifact-pointer-layout: v2 -->
25
  <!-- Generated by artpool.index_projection from meta/, cards/*.yaml, the field registry and its templates. Every release regenerates it, so edits to this file do not last. -->
26
 
27
- ![Public web, in two layers (task material in collections; corpora), one row per file (address, size, SHA-256, license), your selection, fetched and verified](https://huggingface.co/datasets/Battam/agentic-env-index/resolve/release-2026-11-05.3/docs/banner.svg)
28
 
29
  # Agentic Env Index
30
 
@@ -32,45 +32,46 @@ An index of public files for building agent environments: where each one is, its
32
 
33
  English | [简体中文](README.zh.md)
34
 
35
- [![license: CC0-1.0](https://img.shields.io/badge/license-CC0--1.0-4F6D80)](https://creativecommons.org/publicdomain/zero/1.0/) [![material sets: 1,799](https://img.shields.io/badge/material%20sets-1%2C799-4F6D80)](docs/sources.md) [![corpora: 133](https://img.shields.io/badge/corpora-133-4F6D80)](docs/sources.md) [![release: release-2026-11-05.3](https://img.shields.io/badge/release-release--2026--11--05.3-4F6D80)](https://huggingface.co/datasets/Battam/agentic-env-index/tree/release-2026-11-05.3)
36
 
37
  ## Dataset Details
38
 
39
  The index holds two layers over the same public web, at two depths and for two uses;
40
  their counts are never added together, their sizes are.
41
- A material set is a batch of files picked from one source (a site, a repository or a dataset), for building tasks; material sets are grouped into collections by task family; corpora are whole upstreams read in bulk; material sets are not taken from the corpora.
42
 
43
- - **Task material:** 1,799 material sets in 13 collections, with 4,183 units, 787 of them with an answer file.
44
  A material set (task-ready) is a batch of files picked from one source (a site, a repository or a dataset), for building tasks;
45
- its files are grouped into units, often with answers.
46
- - **Corpora:** 133 corpora with 1,227,714,425 rows, to search when no material set fits.
47
  A corpus (plural corpora; raw) is one whole upstream that a collector keeps reading, such as a package registry, a bug tracker, a mailing list or a web archive;
48
- its rows point to files or records and carry no units and no answers.
 
49
  - **What it is:** one row per file, with where to download it, its size, its hashes and its license.
50
  The Dataset Viewer has two subsets: `catalog`, one row per material set or corpus with its `kind` and `collection`, and `preview`, the first 10 rows of each.
51
- One material set or corpus is read from its folder `data/<source>/`, as in the quick start below and with `tools/fetch.py`.
52
- - **Size of the whole index:** Recorded size 4.50 PB; deduplicated size 4.09 PB; size check: 99.6% match (48,396 of the 49,407 rows read compared); not compared: address unreachable 477, file gone 131, login needed 331, no size from the server 68, other error 4; estimated size of 309,522,457 of the 309,787,645 rows without a recorded size: 657 TB (an estimate; 95% range 512 TB to 1.05 PB, the sum of its sources' ranges, wider than a joint interval; from 17,054 rows read), the other 265,188 with too few rows read for an estimate.
53
  The recorded size is the sum of `size_bytes` over the rows that have one (catalog column `bytes`). The check reads a sample of rows of every data file (`meta/sizes.parquet`): 20 of its rows with a recorded size, and of its rows without one its share of its source's sample, sized for their mean size within 25% at 95% confidence by the spread of the sizes read so far (at most 400 rows per source per round, 20 for a source with none read yet); its share is the rows that match their recorded size out of the rows it compared (status `match`, `size_mismatch` or `hash_mismatch`), and the rows it could not compare, which never reached their file, are counted beside it by status (`unreachable`, `gone`, `needs_auth`, `no_length`, `error`; catalog column `size_check`). The estimate for a source's rows without a recorded size (catalog columns `unsized_rows` and `unsized_estimate`) is their number times the mean measured size of such rows read, with a 95% BCa bootstrap range (`scipy.stats.bootstrap`, its default resample count); a source with fewer than 10 such rows read has none. Above one source, the estimate is the sum of its sources' estimates and its range the sum of their ranges, which is wider than a joint interval. The figures of every level are in `meta/size_figures.json`.
54
  - **Who it is for:** people and agents that pick material for terminal tasks and then fetch the files themselves.
55
  - **What it is not:** a list of ready tasks or environments, and not a copy of the files.
56
 
57
  ## Quick start
58
 
59
- Everything below is pinned to the release tag `release-2026-11-05.3`.
60
 
61
  1. Fetch files with one command ([uv](https://docs.astral.sh/uv/) reads the script's dependencies):
62
 
63
  ```sh
64
- uv run https://huggingface.co/datasets/Battam/agentic-env-index/resolve/release-2026-11-05.3/tools/fetch.py --form spreadsheets-with-answers --sample 3 --out ./data
65
  ```
66
 
67
- This step returns units: 3 units from each material set of the form `spreadsheets-with-answers`, with their answers apart under `answer/`.
68
 
69
  2. Choose material sets from the catalog with [DuckDB](https://duckdb.org/):
70
 
71
  ```sql
72
- SELECT source, collection, title, units, answers, files, top_types, license.tier AS license_tier, needs_auth
73
- FROM 'hf://datasets/Battam/agentic-env-index@release-2026-11-05.3/meta/catalog.parquet'
74
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
75
  ORDER BY collection, units DESC;
76
  ```
@@ -81,13 +82,16 @@ with `kind = 'corpus'` it returns corpora instead, counted in rows (`files`).
81
  3. Choose files in one corpus:
82
 
83
  ```sql
 
84
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
85
- FROM 'hf://datasets/Battam/agentic-env-index@release-2026-11-05.3/data/if-archive/*/*.parquet'
86
- WHERE ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
87
  ORDER BY size_bytes LIMIT 20;
88
  ```
89
 
90
  This step returns rows of the corpus `if-archive`, one per file.
 
 
91
 
92
  A usage example video shows the index in use, in Chinese and in English, each with its subtitles: [play it on the data page](https://huggingface.co/spaces/Battam/agentic-env-catalog#usage-video).
93
  The data page plays the video and the subtitles of its own language, in its Get the data section.
@@ -100,56 +104,59 @@ only their sizes add up to the size of the whole index.
100
 
101
  ### Task material, by collection
102
 
103
- 1,799 material sets in 13 collections, with 4,183 units, 787 of them with an answer file.
104
- Recorded size 2.90 GB; deduplicated size 2.82 GB; size check: 99.7% match (9,111 of the 9,135 rows read compared); not compared: address unreachable 11, file gone 3, login needed 8, other error 2.
105
  Upstreams are the upstream works the material sets of a collection come from.
106
 
107
- | Collection | Material sets | Units | Units with an answer file | Upstreams |
108
  |---|---:|---:|---:|---:|
109
- | Spreadsheets with answers | 68 | 545 | 530 | 56 |
110
- | Disk images | 52 | 179 | 154 | 40 |
111
- | Reproduction packages and runs | 88 | 105 | 103 | 88 |
112
  | SEC filing reviews | 0 | unknown | unknown | 0 |
113
- | Solver instances | 63 | 187 | 0 | 60 |
114
- | Browser tests | 123 | 439 | 0 | 88 |
115
- | Hardware design (EDA) tests | 81 | 258 | 0 | 75 |
116
- | Issue trackers | 212 | 258 | 0 | 212 |
117
- | Media and documents | 276 | 391 | 0 | 269 |
118
- | Model serving configurations | 102 | 397 | 0 | 82 |
119
- | Rule calculations | 177 | 329 | 0 | 154 |
120
- | Security detection | 110 | 328 | 0 | 77 |
121
- | Science reference data | 161 | 342 | 0 | 157 |
122
- | Tests and exercises | 286 | 425 | 0 | 265 |
 
123
 
124
  ### Corpora
125
 
126
- 133 corpora with 1,227,714,425 rows.
127
- Recorded size 4.50 PB; deduplicated size 4.09 PB; size check: 99.6% match (39,285 of the 40,272 rows read compared); not compared: address unreachable 466, file gone 128, login needed 323, no size from the server 68, other error 2; estimated size of 309,522,457 of the 309,787,645 rows without a recorded size: 657 TB (an estimate; 95% range 512 TB to 1.05 PB, the sum of its sources' ranges, wider than a joint interval; from 17,054 rows read), the other 265,188 with too few rows read for an estimate.
128
  Each corpus is counted under its main domain.
129
 
130
  | Domain | Corpora | Rows | Distinct size |
131
  |---|---:|---:|---:|
132
- | Data science | 23 | 745,480,870 | 596 TB |
133
- | System administration | 29 | 185,535,882 | 182 TB |
134
- | Many domains | 16 | 109,251,330 | 224 TB |
135
- | Software engineering | 12 | 65,680,028 | 86.5 TB |
136
- | Science | 13 | 43,043,975 | 251 TB |
137
- | Operations | 4 | 39,396,283 | 1.78 TB |
138
- | Media | 9 | 12,795,782 | 876 TB |
139
- | Machine learning | 6 | 9,513,138 | 1.86 PB |
140
- | Hardware | 4 | 8,713,847 | 1.65 TB |
141
- | Security | 10 | 4,172,075 | 11.8 TB |
142
- | Unclassified | 4 | 3,890,105 | 286 GB |
143
  | Misc | 3 | 241,110 | 2.87 TB |
144
 
145
  - [Data page](https://huggingface.co/spaces/Battam/agentic-env-catalog): browse task material by collection, or corpora, one layer at a time, and build a fetch command.
146
  - [The `catalog` subset in the Dataset Viewer](https://huggingface.co/datasets/Battam/agentic-env-index/viewer/catalog): one row per material set or corpus, with its `kind` and `collection`.
147
- - [SQL Console](https://huggingface.co/datasets/Battam/agentic-env-index?sql_console=true&sql=SELECT+source%2C+kind%2C+collection%2C+title%2C+units%2C+answers%2C+files%0AFROM+train%0AORDER+BY+kind%2C+collection%2C+files+DESC%0ALIMIT+50): query the catalog in the browser.
 
148
  - [All material sets, by collection, and all corpora](docs/sources.md): what each one is, with its counts, size, license tier and login.
149
 
150
  ## Dataset Structure
151
 
152
- Each row of `data/<source>/<batch>/part-*.parquet` points to one file or record.
 
153
  The fields used most:
154
 
155
  | Field | Type | Meaning |
@@ -163,7 +170,7 @@ The fields used most:
163
  | `license_spdx` | text | The row's own license (SPDX); empty when it takes its batch license. |
164
  | `batch` | text | Batch the row was published in; its license is in meta/sources.json. |
165
  | `sensitive` | text | health or minors when the item is about health or about minors. |
166
- | `extra` | text | JSON with the row's other fields, such as its role and unit group in a collection. |
167
 
168
  Every row field and every column of the catalog: [docs/fields.md](docs/fields.md).
169
 
@@ -172,15 +179,16 @@ Every row field and every column of the catalog: [docs/fields.md](docs/fields.md
172
  - **Direct use:** choose sources and files for building agent tasks and environments, then fetch each file from its source and check it against the size and hashes in its row.
173
  - **Out-of-scope use:** this dataset is not a copy of the files, and not a training corpus by itself:
174
  the files stay at their sources under their own licenses.
175
- Files under `answer/` are there to check an agent's result;
176
- they never go into the environment an agent works in.
177
  - How to build tasks from the files, what the default profile of `tools/fetch.py` removes and what to keep out of an agent's environment: [docs/for-agents.md](docs/for-agents.md).
178
 
179
  ## Dataset Creation
180
 
181
  - **Collection:** rows are collected from the public listings, APIs and archives of each source and checked when added:
182
  the address must answer, and size and hash are fetched again and compared.
183
- Material sets come from the search and admission of candidates for each collection, not from the corpora.
 
184
  Every material set and every corpus with its counts: [docs/sources.md](docs/sources.md).
185
  - **Checks before release:** every source has its license recorded and grouped into a tier, personal information is read from sample files, and rows are compared with public benchmarks;
186
  rows that copy a public benchmark are removed, and units that do not meet a requirement are listed with the reason: [docs/access.md](docs/access.md).
@@ -195,6 +203,8 @@ Every row field and every column of the catalog: [docs/fields.md](docs/fields.md
195
  - Each file keeps the license of its source: see the license columns of each row (a row without its own license takes the license of its batch, `batches[].license` in `meta/sources.json`).
196
  - License tiers group each source's license by what it allows;
197
  the tier of every source and what each tier means: [docs/sources.md](docs/sources.md).
 
 
198
 
199
  ## Personal and Sensitive Information
200
 
@@ -202,19 +212,19 @@ Every row field and every column of the catalog: [docs/fields.md](docs/fields.md
202
  it contains no personal records of its own.
203
  - Some sources point to material that names its authors, such as mailing lists, bug trackers and forums.
204
  - Rows that point to data about health or about minors carry `health` or `minors` in the row column `sensitive`, so they can be filtered out.
205
- - Personal information read from sample files: 18 of 1,932 sources so far, each from 9 to 30 sample files read on 2026-10-04 and 2026-10-05; by level: `none_known` 3, `present` 14, `suspected` 1. The other 1,914 sources show `not judged yet`. Each source card gives its level with the sample files it rests on (`risk.pii` and `risk.pii_evidence` in the YAML card).
206
 
207
  ## Bias, Risks, and Limitations
208
 
209
- - Sizes: 1,886 of 1,932 sources have a size for every row. For the other 46, rows whose size the source does not publish are shown as unknown and left out of size totals, never counted as 0.
210
- - Judged fields: 0 of 1,932 sources have a summary and task fit read from sample files, and 18 of 1,932 have personal information read from sample files (the same count as in the section above); fields not judged yet say so. Counts, types, sizes and licenses of every source come from the index itself.
211
- - Growing sources: 23 of 1,932 sources are still being collected, and their row counts grow with each release.
212
 
213
  ## Updates and Versions
214
 
215
- - Latest release: `release-2026-11-05.3` (its newest CHANGELOG entry is dated 2026-11-05). Releases were published on 17 of the 30 days from 2026-10-07 to 2026-11-05.
216
  - Releases are listed in [CHANGELOG.md](CHANGELOG.md);
217
- each one is a commit with its own release tag (this one is [`release-2026-11-05.3`](https://huggingface.co/datasets/Battam/agentic-env-index/tree/release-2026-11-05.3)),
218
  so any earlier state stays readable at its tag.
219
  - `meta/projection.json` records the release tag and the version of the program and templates that generated this card.
220
  - Croissant metadata: [generated by the Hub](https://huggingface.co/api/datasets/Battam/agentic-env-index/croissant);
@@ -236,7 +246,7 @@ A removal covers the current version of this repository (the main branch) and th
236
  @misc{agentic_env_index,
237
  title = {agentic-env-index: an index of public materials for building agent environments},
238
  howpublished = {https://huggingface.co/datasets/Battam/agentic-env-index},
239
- note = {Release release-2026-11-05.3},
240
  year = {2026}
241
  }
242
  ```
 
24
  <!-- artifact-pointer-layout: v2 -->
25
  <!-- Generated by artpool.index_projection from meta/, cards/*.yaml, the field registry and its templates. Every release regenerates it, so edits to this file do not last. -->
26
 
27
+ ![Public web, in two layers (task material in collections; corpora), one row per file (address, size, SHA-256, license), your selection, fetched and verified](https://huggingface.co/datasets/Battam/agentic-env-index/resolve/release-2026-11-09.1/docs/banner.svg)
28
 
29
  # Agentic Env Index
30
 
 
32
 
33
  English | [简体中文](README.zh.md)
34
 
35
+ [![license: CC0-1.0](https://img.shields.io/badge/license-CC0--1.0-4F6D80)](https://creativecommons.org/publicdomain/zero/1.0/) [![material sets: 1,799](https://img.shields.io/badge/material%20sets-1%2C799-4F6D80)](docs/sources.md) [![corpora: 138](https://img.shields.io/badge/corpora-138-4F6D80)](docs/sources.md) [![release: release-2026-11-09.1](https://img.shields.io/badge/release-release--2026--11--09.1-4F6D80)](https://huggingface.co/datasets/Battam/agentic-env-index/tree/release-2026-11-09.1)
36
 
37
  ## Dataset Details
38
 
39
  The index holds two layers over the same public web, at two depths and for two uses;
40
  their counts are never added together, their sizes are.
41
+ A material set is a batch of files picked from one source (a site, a repository or a dataset), for building tasks; material sets are grouped into collections by task family; corpora are whole upstreams read in bulk; a group of corpus rows becomes a material set only by passing admission.
42
 
43
+ - **Task material:** 1,799 material sets in 13 collections, with 4,183 units.
44
  A material set (task-ready) is a batch of files picked from one source (a site, a repository or a dataset), for building tasks;
45
+ its files are grouped into units.
46
+ - **Corpora:** 138 corpora with 1,235,871,975 rows, to search when no material set fits.
47
  A corpus (plural corpora; raw) is one whole upstream that a collector keeps reading, such as a package registry, a bug tracker, a mailing list or a web archive;
48
+ its rows point to files or records, each row one unit.
49
+ - **Unit:** A unit is a group of files that its upstream publishes as one thing, delivered whole. Its key, `unit_id`, is the `unit_group` its rows share in `source_meta` (a test case, one workbook, one package), else the row's natural key written as its `uri`, then `@<range_offset>` for a byte range, then `!/<member_path>` for an archive member; it is unique within a source. The README.md and LICENSE the index writes for a collection source belong to the source, not to a unit. `tools/fetch.py` puts every unit in a folder of its own named after its `unit_id` and writes `unit_id` on every line of its manifests. A unit is material, not a task: one unit may give several tasks, and one task may use several units.
50
  - **What it is:** one row per file, with where to download it, its size, its hashes and its license.
51
  The Dataset Viewer has two subsets: `catalog`, one row per material set or corpus with its `kind` and `collection`, and `preview`, the first 10 rows of each.
52
+ One material set or corpus is read from the data files its catalog row names in `data_files`, as in the quick start below and with `tools/fetch.py`.
53
+ - **Size of the whole index:** Recorded size 4.50 PB; deduplicated size 4.09 PB; size check: 99.7% match (43,648 of the 44,629 rows read compared); not compared: address unreachable 466, file gone 119, login needed 331, no size from the server 61, other error 4; estimated size of 279,909,103 of the 311,970,416 rows without a recorded size: 583 TB (an estimate; 95% range 497 TB to 701 TB, the sum of its sources' ranges, wider than a joint interval; from 15,964 rows read), the other 32,061,313 with too few rows read for an estimate.
54
  The recorded size is the sum of `size_bytes` over the rows that have one (catalog column `bytes`). The check reads a sample of rows of every data file (`meta/sizes.parquet`): 20 of its rows with a recorded size, and of its rows without one its share of its source's sample, sized for their mean size within 25% at 95% confidence by the spread of the sizes read so far (at most 400 rows per source per round, 20 for a source with none read yet); its share is the rows that match their recorded size out of the rows it compared (status `match`, `size_mismatch` or `hash_mismatch`), and the rows it could not compare, which never reached their file, are counted beside it by status (`unreachable`, `gone`, `needs_auth`, `no_length`, `error`; catalog column `size_check`). The estimate for a source's rows without a recorded size (catalog columns `unsized_rows` and `unsized_estimate`) is their number times the mean measured size of such rows read, with a 95% BCa bootstrap range (`scipy.stats.bootstrap`, its default resample count); a source with fewer than 10 such rows read has none. Above one source, the estimate is the sum of its sources' estimates and its range the sum of their ranges, which is wider than a joint interval. The figures of every level are in `meta/size_figures.json`.
55
  - **Who it is for:** people and agents that pick material for terminal tasks and then fetch the files themselves.
56
  - **What it is not:** a list of ready tasks or environments, and not a copy of the files.
57
 
58
  ## Quick start
59
 
60
+ Everything below is pinned to the release tag `release-2026-11-09.1`.
61
 
62
  1. Fetch files with one command ([uv](https://docs.astral.sh/uv/) reads the script's dependencies):
63
 
64
  ```sh
65
+ uv run https://huggingface.co/datasets/Battam/agentic-env-index/resolve/release-2026-11-09.1/tools/fetch.py --form spreadsheets --sample 3 --out ./data
66
  ```
67
 
68
+ This step returns units: 3 units from each material set of the form `spreadsheets`, each in a folder of its own named after its `unit_id`.
69
 
70
  2. Choose material sets from the catalog with [DuckDB](https://duckdb.org/):
71
 
72
  ```sql
73
+ SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
74
+ FROM 'hf://datasets/Battam/agentic-env-index@release-2026-11-09.1/meta/catalog.parquet'
75
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
76
  ORDER BY collection, units DESC;
77
  ```
 
82
  3. Choose files in one corpus:
83
 
84
  ```sql
85
+ SET VARIABLE files = (SELECT list('hf://datasets/Battam/agentic-env-index@release-2026-11-09.1/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/agentic-env-index@release-2026-11-09.1/meta/catalog.parquet' WHERE source = 'if-archive'));
86
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
87
+ FROM read_parquet(getvariable('files'))
88
+ WHERE source = 'if-archive' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
89
  ORDER BY size_bytes LIMIT 20;
90
  ```
91
 
92
  This step returns rows of the corpus `if-archive`, one per file.
93
+ The first statement takes its data files from the catalog's column `data_files`, so the query reads exactly those files and lists no folder of `data/`.
94
+ A data file may hold the rows of several sources, so the query keeps this one's by its column `source`.
95
 
96
  A usage example video shows the index in use, in Chinese and in English, each with its subtitles: [play it on the data page](https://huggingface.co/spaces/Battam/agentic-env-catalog#usage-video).
97
  The data page plays the video and the subtitles of its own language, in its Get the data section.
 
104
 
105
  ### Task material, by collection
106
 
107
+ 1,799 material sets in 13 collections, with 4,183 units.
108
+ Recorded size 2.90 GB; deduplicated size 2.82 GB; size check: 99.7% match (8,122 of the 8,144 rows read compared); not compared: address unreachable 10, file gone 2, login needed 8, other error 2.
109
  Upstreams are the upstream works the material sets of a collection come from.
110
 
111
+ | Collection | Material sets | Units | Files | Upstreams |
112
  |---|---:|---:|---:|---:|
113
+ | Spreadsheets | 67 | 536 | 1,200 | 55 |
114
+ | Disk images | 52 | 179 | 470 | 40 |
115
+ | Reproduction packages and runs | 88 | 105 | 965 | 88 |
116
  | SEC filing reviews | 0 | unknown | unknown | 0 |
117
+ | Solver instances | 63 | 187 | 325 | 60 |
118
+ | Browser tests | 123 | 439 | 852 | 88 |
119
+ | Hardware design (EDA) tests | 81 | 258 | 651 | 75 |
120
+ | Issue trackers | 212 | 258 | 682 | 212 |
121
+ | Media and documents | 276 | 391 | 1,088 | 269 |
122
+ | Model serving configurations | 102 | 397 | 1,169 | 82 |
123
+ | Rule calculations | 177 | 329 | 727 | 154 |
124
+ | Security detection | 110 | 328 | 983 | 77 |
125
+ | Science reference data | 161 | 342 | 671 | 157 |
126
+ | Tests and exercises | 286 | 425 | 2,544 | 265 |
127
+ | In no collection | 1 | 9 | 19 | 1 |
128
 
129
  ### Corpora
130
 
131
+ 138 corpora with 1,235,871,975 rows.
132
+ Recorded size 4.50 PB; deduplicated size 4.09 PB; size check: 99.7% match (35,526 of the 36,485 rows read compared); not compared: address unreachable 456, file gone 117, login needed 323, no size from the server 61, other error 2; estimated size of 279,909,103 of the 311,970,416 rows without a recorded size: 583 TB (an estimate; 95% range 497 TB to 701 TB, the sum of its sources' ranges, wider than a joint interval; from 15,964 rows read), the other 32,061,313 with too few rows read for an estimate.
133
  Each corpus is counted under its main domain.
134
 
135
  | Domain | Corpora | Rows | Distinct size |
136
  |---|---:|---:|---:|
137
+ | Data science | 23 | 750,815,142 | 600 TB |
138
+ | System administration | 29 | 185,547,838 | 182 TB |
139
+ | Many domains | 16 | 109,251,421 | 224 TB |
140
+ | Software engineering | 12 | 65,693,542 | 86.5 TB |
141
+ | Science | 15 | 44,541,599 | 251 TB |
142
+ | Operations | 6 | 39,404,642 | 1.78 TB |
143
+ | Media | 9 | 12,751,384 | 874 TB |
144
+ | Machine learning | 6 | 9,520,852 | 1.86 PB |
145
+ | Hardware | 5 | 8,715,949 | 1.66 TB |
146
+ | Unclassified | 4 | 4,869,099 | 291 GB |
147
+ | Security | 10 | 4,519,397 | 11.9 TB |
148
  | Misc | 3 | 241,110 | 2.87 TB |
149
 
150
  - [Data page](https://huggingface.co/spaces/Battam/agentic-env-catalog): browse task material by collection, or corpora, one layer at a time, and build a fetch command.
151
  - [The `catalog` subset in the Dataset Viewer](https://huggingface.co/datasets/Battam/agentic-env-index/viewer/catalog): one row per material set or corpus, with its `kind` and `collection`.
152
+ - [SQL Console](https://huggingface.co/datasets/Battam/agentic-env-index?sql_console=true&sql=--+every+source+of+the+catalog+as+published%3A+the+default+profile+of+tools%2Ffetch.py+is+not+applied+here%0ASELECT+source%2C+kind%2C+collection%2C+title%2C+units%2C+files%0AFROM+train%0AORDER+BY+kind%2C+collection%2C+files+DESC%0ALIMIT+50): query the catalog in the browser;
153
+ it lists every source as published, without the default profile of `tools/fetch.py`, which the command applies.
154
  - [All material sets, by collection, and all corpora](docs/sources.md): what each one is, with its counts, size, license tier and login.
155
 
156
  ## Dataset Structure
157
 
158
+ Each row of the data files under `data/` points to one file or record;
159
+ the catalog's column `data_files` names the files of each source.
160
  The fields used most:
161
 
162
  | Field | Type | Meaning |
 
170
  | `license_spdx` | text | The row's own license (SPDX); empty when it takes its batch license. |
171
  | `batch` | text | Batch the row was published in; its license is in meta/sources.json. |
172
  | `sensitive` | text | health or minors when the item is about health or about minors. |
173
+ | `extra` | text | JSON with the row's other fields, such as its unit group in a collection. |
174
 
175
  Every row field and every column of the catalog: [docs/fields.md](docs/fields.md).
176
 
 
179
  - **Direct use:** choose sources and files for building agent tasks and environments, then fetch each file from its source and check it against the size and hashes in its row.
180
  - **Out-of-scope use:** this dataset is not a copy of the files, and not a training corpus by itself:
181
  the files stay at their sources under their own licenses.
182
+ What a task checks is decided when the task is built from the material;
183
+ the index holds material only.
184
  - How to build tasks from the files, what the default profile of `tools/fetch.py` removes and what to keep out of an agent's environment: [docs/for-agents.md](docs/for-agents.md).
185
 
186
  ## Dataset Creation
187
 
188
  - **Collection:** rows are collected from the public listings, APIs and archives of each source and checked when added:
189
  the address must answer, and size and hash are fetched again and compared.
190
+ Material sets come from the search and admission of candidates for each collection;
191
+ a group of corpus rows becomes a material set only by passing the same admission.
192
  Every material set and every corpus with its counts: [docs/sources.md](docs/sources.md).
193
  - **Checks before release:** every source has its license recorded and grouped into a tier, personal information is read from sample files, and rows are compared with public benchmarks;
194
  rows that copy a public benchmark are removed, and units that do not meet a requirement are listed with the reason: [docs/access.md](docs/access.md).
 
203
  - Each file keeps the license of its source: see the license columns of each row (a row without its own license takes the license of its batch, `batches[].license` in `meta/sources.json`).
204
  - License tiers group each source's license by what it allows;
205
  the tier of every source and what each tier means: [docs/sources.md](docs/sources.md).
206
+ - License tiers and reason codes are computed from the license categories of ScanCode LicenseDB, licensed CC-BY-4.0;
207
+ its attribution: [`NOTICE`](NOTICE).
208
 
209
  ## Personal and Sensitive Information
210
 
 
212
  it contains no personal records of its own.
213
  - Some sources point to material that names its authors, such as mailing lists, bug trackers and forums.
214
  - Rows that point to data about health or about minors carry `health` or `minors` in the row column `sensitive`, so they can be filtered out.
215
+ - Personal information read from sample files: 18 of 1,937 sources so far, each from 9 to 30 sample files read on 2026-10-04 and 2026-10-05; by level: `none_known` 3, `present` 14, `suspected` 1. The other 1,919 sources show `not judged yet`. The card of each source gives its level with the sample files it rests on (`risk.pii` and `risk.pii_evidence` in `cards/<source>.yaml`).
216
 
217
  ## Bias, Risks, and Limitations
218
 
219
+ - Sizes: 1,890 of 1,937 sources have a size for every row. For the other 47, rows whose size the source does not publish are shown as unknown and left out of size totals, never counted as 0.
220
+ - Judged fields: 0 of 1,937 sources have a summary and task fit read from sample files, and 18 of 1,937 have personal information read from sample files (the same count as in the section above); fields not judged yet say so. Counts, types, sizes and licenses of every source come from the index itself.
221
+ - Growing sources: 25 of 1,937 sources are still being collected, and their row counts grow with each release.
222
 
223
  ## Updates and Versions
224
 
225
+ - Latest release: `release-2026-11-09.1` (its newest CHANGELOG entry is dated 2026-11-09). Releases were published on 14 of the 30 days from 2026-10-11 to 2026-11-09.
226
  - Releases are listed in [CHANGELOG.md](CHANGELOG.md);
227
+ each one is a commit with its own release tag (this one is [`release-2026-11-09.1`](https://huggingface.co/datasets/Battam/agentic-env-index/tree/release-2026-11-09.1)),
228
  so any earlier state stays readable at its tag.
229
  - `meta/projection.json` records the release tag and the version of the program and templates that generated this card.
230
  - Croissant metadata: [generated by the Hub](https://huggingface.co/api/datasets/Battam/agentic-env-index/croissant);
 
246
  @misc{agentic_env_index,
247
  title = {agentic-env-index: an index of public materials for building agent environments},
248
  howpublished = {https://huggingface.co/datasets/Battam/agentic-env-index},
249
+ note = {Release release-2026-11-09.1},
250
  year = {2026}
251
  }
252
  ```
README.zh.md CHANGED
@@ -11,7 +11,7 @@ tags:
11
  ---
12
  <!-- Generated by artpool.index_projection from meta/, cards/*.yaml, the field registry and its templates. Every release regenerates it, so edits to this file do not last. -->
13
 
14
- ![公开网络,分两层(按合集的任务材料;语料库),每个文件一行(地址、大小、SHA-256、许可证),你的选择,下载并校验](https://huggingface.co/datasets/Battam/agentic-env-index/resolve/release-2026-11-05.3/docs/banner.zh.svg)
15
 
16
  # Agentic Env Index
17
 
@@ -19,45 +19,46 @@ tags:
19
 
20
  [English](README.md) | 简体中文
21
 
22
- [![许可证: CC0-1.0](https://img.shields.io/badge/%E8%AE%B8%E5%8F%AF%E8%AF%81-CC0--1.0-4F6D80)](https://creativecommons.org/publicdomain/zero/1.0/) [![材料集: 1,799](https://img.shields.io/badge/%E6%9D%90%E6%96%99%E9%9B%86-1%2C799-4F6D80)](docs/sources.zh.md) [![语料库: 133](https://img.shields.io/badge/%E8%AF%AD%E6%96%99%E5%BA%93-133-4F6D80)](docs/sources.zh.md) [![版本: release-2026-11-05.3](https://img.shields.io/badge/%E7%89%88%E6%9C%AC-release--2026--11--05.3-4F6D80)](https://huggingface.co/datasets/Battam/agentic-env-index/tree/release-2026-11-05.3)
23
 
24
  ## 数据集详情
25
 
26
  索引在同一片公开网络上分两层,深度不同、用途不同;
27
  两层的计数从不相加,大小可以相加。
28
- 材料集是从同一个来源挑出来的一批文件(一个网站、一个仓库或一个数据集),用来构建任务,按任务类别分到各个合集;语料库是把整个上游全部读下来;材料集不是从语料库里挑出来的。
29
 
30
- - **任务材料:** 13 个合集中的 1,799 个材料集,共 4,183 个单元,其中 787 个有答案文件。
31
  材料集(任务就绪)是从同一个来源挑出来的一批文件(一个网站、一个仓库或一个数据集),用来构建任务;
32
- 它的文件分组为单元,常带有答案。
33
- - **语料库:** 133 个语料库,共 1,227,714,425 行,在没有合适的材料集时用来搜索。
34
  语料库(原始)是采集器持续读取的一个完整上游,例如软件包仓库、缺陷跟踪系统、邮件列表或网页存档;
35
- 它的行指向文件或记录,不分单元,也没有答案。
 
36
  - **是什么:** 每个文件一行,写明下载地址、大小、哈希值和许可证。
37
  数据集查看器有两个子集:`catalog`,每个材料集或语料库一行,带有它的 `kind` 和 `collection`;`preview`,每个的前 10 行。
38
- 单个材料集或语料库从它的文件夹 `data/<source>/` 读取,见下面的快速上手和 `tools/fetch.py`。
39
- - **整个索引的大小:** 记录的大小 4.50 PB;去重后大小 4.09 PB;大小核对:99.6% 一致(读过 49,407 行,比对了 48,396 行);未比对:地址无法访问 477 行、文件已不存在 131 行、需要登录 331 行、服务器未给出大小 68 行、其他错误 4 行;没有记录大小的 309,787,645 行中 309,522,457 行的估计大小:657 TB(这是估计值;95% ���围 512 TB 到 1.05 PB,为其各来源范围之和,比联合区间更宽;依据读过的 17,054 行),其余 265,188 行读过的太少,无法估计。
40
  记录的大小是有大小的各行 `size_bytes` 之和(目录列 `bytes`)。核对读取每个数据文件中抽取的一组行(`meta/sizes.parquet`):有记录大小的行抽 20 行;没有记录大小的行按其所在来源的样本分摊,样本量按目前读到的大小的离散程度定,使其平均大小在 95% 置信度下误差不超过 25%(每个来源每轮最多 400 行,还没有读数的来源取 20 行);其比例是比对过的行(状态 `match`、`size_mismatch` 或 `hash_mismatch`)中与记录的大小一致的行所占的比例,没能比对的行(没有取到文件)按状态另行计数(`unreachable`、`gone`、`needs_auth`、`no_length`、`error`;目录列 `size_check`)。一个来源没有记录大小的行的估计(目录列 `unsized_rows` 和 `unsized_estimate`)是这些行的行数乘以读过的这类行的平均实测大小,并给出 95% BCa 自助法范围(`scipy.stats.bootstrap`,用其默认重抽样次数);读过的这类行少于 10 行的来源没有估计。在来源之上的层级,估计是其各来源估计之和,范围是各来源范围之和,比联合区间更宽。每个层级的数字都在 `meta/size_figures.json` 中。
41
  - **给谁用:** 为终端任务挑选材料、然后自己下载文件的人和智能体。
42
  - **不是什么:** 不是现成的任务或环境列表,也不是文件的副本。
43
 
44
  ## 快速上手
45
 
46
- 下面的一切都固定在发布标签 `release-2026-11-05.3` 上。
47
 
48
  1. 用一条命令下载文件([uv](https://docs.astral.sh/uv/) 会读取脚本声明的依赖):
49
 
50
  ```sh
51
- uv run https://huggingface.co/datasets/Battam/agentic-env-index/resolve/release-2026-11-05.3/tools/fetch.py --form spreadsheets-with-answers --sample 3 --out ./data
52
  ```
53
 
54
- 这一步返回单元:形式为 `spreadsheets-with-answers` 的每个材料集各 3 个单元,答案另放在 `answer/` 下。
55
 
56
  2. 用 [DuckDB](https://duckdb.org/) 从目录中挑选材料集:
57
 
58
  ```sql
59
- SELECT source, collection, title, units, answers, files, top_types, license.tier AS license_tier, needs_auth
60
- FROM 'hf://datasets/Battam/agentic-env-index@release-2026-11-05.3/meta/catalog.parquet'
61
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
62
  ORDER BY collection, units DESC;
63
  ```
@@ -68,13 +69,16 @@ ORDER BY collection, units DESC;
68
  3. 在一个语料库里挑选文件:
69
 
70
  ```sql
 
71
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
72
- FROM 'hf://datasets/Battam/agentic-env-index@release-2026-11-05.3/data/if-archive/*/*.parquet'
73
- WHERE ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
74
  ORDER BY size_bytes LIMIT 20;
75
  ```
76
 
77
  这一步返回语料库 `if-archive` 的行,每个文件一行。
 
 
78
 
79
  使用示例视频有中文版和英文版,各带字幕,展示了本索引的用法:[在数据页播放](https://huggingface.co/spaces/Battam/agentic-env-catalog#usage-video)。
80
  数据页在“获取数据”部分播放��页面语言相同的视频和字幕。
@@ -87,56 +91,59 @@ ORDER BY size_bytes LIMIT 20;
87
 
88
  ### 任务材料,按合集
89
 
90
- 13 个合集中的 1,799 个材料集,共 4,183 个单元,其中 787 个有答案文件。
91
- 记录的大小 2.90 GB;去重后大小 2.82 GB;大小核对:99.7% 一致(读过 9,135 行,比对了 9,111 行);未比对:地址无法访问 11 行、文件已不存在 3 行、需要登录 8 行、其他错误 2 行。
92
  上游是一个合集的材料集所来自的上游作品。
93
 
94
- | 合集 | 材料集数 | 单元数 | 有答案文件的单元数 | 上游数 |
95
  |---|---:|---:|---:|---:|
96
- | 带答案的电子表格 | 68 | 545 | 530 | 56 |
97
- | 磁盘镜像 | 52 | 179 | 154 | 40 |
98
- | 复现包与运行记录 | 88 | 105 | 103 | 88 |
99
  | SEC 申报审阅 | 0 | 未知 | 未知 | 0 |
100
- | 求解器实例 | 63 | 187 | 0 | 60 |
101
- | 浏览器测试 | 123 | 439 | 0 | 88 |
102
- | 硬件设计(EDA)测试 | 81 | 258 | 0 | 75 |
103
- | 问题跟踪系统 | 212 | 258 | 0 | 212 |
104
- | 媒体与文档 | 276 | 391 | 0 | 269 |
105
- | 模型服务配置 | 102 | 397 | 0 | 82 |
106
- | 规则计算 | 177 | 329 | 0 | 154 |
107
- | 安全检测 | 110 | 328 | 0 | 77 |
108
- | 科学参考数据 | 161 | 342 | 0 | 157 |
109
- | 测试与练习 | 286 | 425 | 0 | 265 |
 
110
 
111
  ### 语料库
112
 
113
- 133 个语料库,共 1,227,714,425 行。
114
- 记录的大小 4.50 PB;去重后大小 4.09 PB;大小核对:99.6% 一致(读过 40,272 行,比对了 39,285 行);未比对:地址无法访问 466 行、文件已不存在 128 行、需要登录 323 行、服务器未给出大小 68 行、其他错误 2 行;没有记录大小的 309,787,645 行中 309,522,457 行的估计大小:657 TB(这是估计值;95% 范围 512 TB 到 1.05 PB,为其各来源范围之和,比联合区间更宽;依据读过的 17,054 行),其余 265,188 行读过的太少,无法估计。
115
  每个语料库计入它的主要领域。
116
 
117
  | 领域 | 语料库数 | 行数 | 去重后大小 |
118
  |---|---:|---:|---:|
119
- | 数据科学 / 数据处理 / 查询 | 23 | 745,480,870 | 596 TB |
120
- | 系统管理 / DevOps / 环境构建 | 29 | 185,535,882 | 182 TB |
121
- | 多领域 | 16 | 109,251,330 | 224 TB |
122
- | 软件工程 | 12 | 65,680,028 | 86.5 TB |
123
- | 科学 | 13 | 43,043,975 | 251 TB |
124
- | 业务运营流程 | 4 | 39,396,283 | 1.78 TB |
125
- | 媒体 | 9 | 12,795,782 | 876 TB |
126
- | 机器学习 / 模型运维 | 6 | 9,513,138 | 1.86 PB |
127
- | 硬件 | 4 | 8,713,847 | 1.65 TB |
128
- | 安全 | 10 | 4,172,075 | 11.8 TB |
129
- | 未分类 | 4 | 3,890,105 | 286 GB |
130
  | 杂项 / 游戏 / 策略 / 约束调度 | 3 | 241,110 | 2.87 TB |
131
 
132
  - [数据页面](https://huggingface.co/spaces/Battam/agentic-env-catalog):按合集浏览任务材料,或浏览语料库,每次一层,并生成下载命令。
133
  - [数据集查看器中的 `catalog` 子集](https://huggingface.co/datasets/Battam/agentic-env-index/viewer/catalog):每个材料集或语料库一行,带有它的 `kind` 和 `collection`。
134
- - [SQL 控制台](https://huggingface.co/datasets/Battam/agentic-env-index?sql_console=true&sql=SELECT+source%2C+kind%2C+collection%2C+title%2C+units%2C+answers%2C+files%0AFROM+train%0AORDER+BY+kind%2C+collection%2C+files+DESC%0ALIMIT+50):在浏览器里查询目录。
 
135
  - [全部材料集(按合集)和全部语料库](docs/sources.zh.md):每个是什么,它的计数、大小、许可证等级和登录要求。
136
 
137
  ## 数据集结构
138
 
139
- `data/<source>/<batch>/part-*.parquet` 的每一行指向一个文件或一条记录。
 
140
  最常用的字段:
141
 
142
  | 字段 | 类型 | 含义 |
@@ -150,7 +157,7 @@ ORDER BY size_bytes LIMIT 20;
150
  | `license_spdx` | 文本 | 该行自己的许可证(SPDX);沿用所在批次的许可证时为空。 |
151
  | `batch` | 文本 | 该行发布时所在的批次;批次的许可证见 meta/sources.json。 |
152
  | `sensitive` | 文本 | 条目涉及健康或未成年人时为 health 或 minors。 |
153
- | `extra` | 文本 | 该行其余字段的 JSON,例如它在集合中的角色和单元分组。 |
154
 
155
  全部行字段和目录的全部列:[docs/fields.zh.md](docs/fields.zh.md)。
156
 
@@ -159,15 +166,16 @@ ORDER BY size_bytes LIMIT 20;
159
  - **直接用途:** 为构建智能体任务和环境挑选来源和文件,再从各自的来源下载每个文件,并按它所在行记录的大小和哈希值校验。
160
  - **不适用的用途:** 本数据集不是文件的副本,本身也不是训练语料:
161
  文件留在各自的来源,遵循各自的许可证。
162
- `answer/` 下的文件用来检查智能体的结果;
163
- 它们绝不放进智能体工作的环境。
164
  - 如何用这些文件构建任务、`tools/fetch.py` 的默认配置会去掉什么、哪些内容不能放进智能体的环境:[docs/for-agents.zh.md](docs/for-agents.zh.md)。
165
 
166
  ## 数据集构建
167
 
168
  - **收集:** 行从每个来源的公开列表、API 和存档中收集,加入时即校验:
169
  地址必须能访问,大小和哈希值会重新取回并比对。
170
- 材料集来自为每个合集对候选所做的搜索和收录,不取自语料库。
 
171
  每个材料集和每个语料库及其计数:[docs/sources.zh.md](docs/sources.zh.md)。
172
  - **发布前的检查:** 每个来源都记录了许可证并归入一个等级,个人信息从样本文件中判读,各行与公开基准测试比对;
173
  复制了公开基准测试的行会被移除,不满足某项要求的单元会连同原因一起列出:[docs/access.zh.md](docs/access.zh.md)。
@@ -183,6 +191,8 @@ ORDER BY size_bytes LIMIT 20;
183
  - 每个文件保留其来源的许可证:见每一行的许可证列(没有自己许可证的行沿用所在批次的许可证,即 `meta/sources.json` 中的 `batches[].license`)。
184
  - 许可证等级按许可内容对每个来源的许可证分组;
185
  每个来源的等级以及每个等级的含义:[docs/sources.zh.md](docs/sources.zh.md)。
 
 
186
 
187
  ## 个人与敏感信息
188
 
@@ -190,19 +200,19 @@ ORDER BY size_bytes LIMIT 20;
190
  它本身不含任何个人记录。
191
  - 有些来源指向署有作者姓名的材料,例如邮件列表、缺陷跟踪系统和论坛。
192
  - 指向健康或未成年人相关数据的行,在行列 `sensitive` 中标有 `health` 或 `minors`,可以据此过滤掉。
193
- - 从样本文件读出的个人信息:目前 1,932 个来源中已判读 18 个,每个来源读了 9 到 30 个样本文件,读取日期为 2026-10-04 和 2026-10-05;按等级:`none_known` 3, `present` 14, `suspected` 1。其余 1,914 个来源显示 `not judged yet`(尚未判读)。每张来源卡片给出它的等级和所依据的样本文件(YAML 卡片里的 `risk.pii` 和 `risk.pii_evidence`)。
194
 
195
  ## 偏差、风险与局限
196
 
197
- - 大小:1,932 个来源中有 1,886 个每一行都有大小。其余 46 个来源里,来源没有公布大小的行显示为未知,不计入大小合计,也从不按 0 计。
198
- - 判读字段:1,932 个来源中有 0 个已从样本文件读出摘要和任务适配度,1,932 个中有 18 个已从样本文件读出个人信息(与上一节的计数相同);尚未判读的字段会注明。每个来源的计数、类型、大小和许可证都直接来自索引本身。
199
- - 仍在增长的来源:1,932 个来源中有 23 个仍在收集中,它们的行数随每次发布增加。
200
 
201
  ## 更新与版本
202
 
203
- - 最新发布:`release-2026-11-05.3`(其最新的 CHANGELOG 条目日期为 2026-11-05)。 从 2026-10-07 到 2026-11-05 的 30 天里,有 17 天发布过新版本。
204
  - 各次发布列在 [CHANGELOG.md](CHANGELOG.md) 中;
205
- 每次发布是一个带有自己发布标签的提交(本次为 [`release-2026-11-05.3`](https://huggingface.co/datasets/Battam/agentic-env-index/tree/release-2026-11-05.3)),
206
  因此任何较早的状态都可以在它的标签处读到。
207
  - `meta/projection.json` 记录发布标签,以及生成本卡片的程序和模板的版本。
208
  - Croissant 元数据:[由 Hub 生成](https://huggingface.co/api/datasets/Battam/agentic-env-index/croissant);
@@ -232,7 +242,7 @@ ORDER BY size_bytes LIMIT 20;
232
  @misc{agentic_env_index,
233
  title = {agentic-env-index: an index of public materials for building agent environments},
234
  howpublished = {https://huggingface.co/datasets/Battam/agentic-env-index},
235
- note = {Release release-2026-11-05.3},
236
  year = {2026}
237
  }
238
  ```
 
11
  ---
12
  <!-- Generated by artpool.index_projection from meta/, cards/*.yaml, the field registry and its templates. Every release regenerates it, so edits to this file do not last. -->
13
 
14
+ ![公开网络,分两层(按合集的任务材料;语料库),每个文件一行(地址、大小、SHA-256、许可证),你的选择,下载并校验](https://huggingface.co/datasets/Battam/agentic-env-index/resolve/release-2026-11-09.1/docs/banner.zh.svg)
15
 
16
  # Agentic Env Index
17
 
 
19
 
20
  [English](README.md) | 简体中文
21
 
22
+ [![许可证: CC0-1.0](https://img.shields.io/badge/%E8%AE%B8%E5%8F%AF%E8%AF%81-CC0--1.0-4F6D80)](https://creativecommons.org/publicdomain/zero/1.0/) [![材料集: 1,799](https://img.shields.io/badge/%E6%9D%90%E6%96%99%E9%9B%86-1%2C799-4F6D80)](docs/sources.zh.md) [![语料库: 138](https://img.shields.io/badge/%E8%AF%AD%E6%96%99%E5%BA%93-138-4F6D80)](docs/sources.zh.md) [![版本: release-2026-11-09.1](https://img.shields.io/badge/%E7%89%88%E6%9C%AC-release--2026--11--09.1-4F6D80)](https://huggingface.co/datasets/Battam/agentic-env-index/tree/release-2026-11-09.1)
23
 
24
  ## 数据集详情
25
 
26
  索引在同一片公开网络上分两层,深度不同、用途不同;
27
  两层的计数从不相加,大小可以相加。
28
+ 材料集是从同一个来源挑出来的一批文件(一个网站、一个仓库或一个数据集),用来构建任务,按任务类别分到各个合集;语料库是把整个上游全部读下来;一组语料库行只有通过准入才成为材料集。
29
 
30
+ - **任务材料:** 13 个合集中的 1,799 个材料集,共 4,183 个单元。
31
  材料集(任务就绪)是从同一个来源挑出来的一批文件(一个网站、一个仓库或一个数据集),用来构建任务;
32
+ 它的文件分组为单元。
33
+ - **语料库:** 138 个语料库,共 1,235,871,975 行,在没有合适的材料集时用来搜索。
34
  语料库(原始)是采集器持续读取的一个完整上游,例如软件包仓库、缺陷跟踪系统、邮件列表或网页存档;
35
+ 它的行指向文件或记录,每行是一个单元。
36
+ - **单元:** 单元是上游作为一件东西发布的一组文件,整体交付。它的键 `unit_id` 是它的各行在 `source_meta` 中共用的 `unit_group`(一个测试用例、一个工作簿、一个软件包),否则是该行的自然键,写作它的 `uri`,字节范围再接 `@<range_offset>`,存档成员再接 `!/<member_path>`;它在一个来源内唯一。索引为集合来源写的 README.md 和 LICENSE 属于来源,不属于任何单元。`tools/fetch.py` 把每个单元放进以它的 `unit_id` 命名的单独文件夹,并在它的清单的每一行写上 `unit_id`。单元是材料,不是题:一个单元可以出几道题,一道题也可以用几个单元。
37
  - **是什么:** 每个文件一行,写明下载地址、大小、哈希值和许可证。
38
  数据集查看器有两个子集:`catalog`,每个材料集或语料库一行,带有它的 `kind` 和 `collection`;`preview`,每个的前 10 行。
39
+ 单个材料集或语料库从它在目录中那一行的 `data_files` 列出的数据文件读取,见下面的快速上手和 `tools/fetch.py`。
40
+ - **整个索引的大小:** 记录的大小 4.50 PB;去重后大小 4.09 PB;大小核对:99.7% 一致(读过 44,629 行,比对了 43,648 行);未比对:地址无法访问 466 行、文件已不存在 119 行、需要登录 331 行、服务器未给出大小 61 行、其他错误 4 行;没有记录大小的 311,970,416 行中 279,909,103 行的估计大小:583 TB(这是估计值;95% 范围 497 TB 到 701 TB,为其各来源范围之和,比联合区间更宽;依据读过的 15,964 行),其余 32,061,313 行读过的太少,无法估计。
41
  记录的大小是有大小的各行 `size_bytes` 之和(目录列 `bytes`)。核对读取每个数据文件中抽取的一组行(`meta/sizes.parquet`):有记录大小的行抽 20 行;没有记录大小的行按其所在来源的样本分摊,样本量按目前读到的大小的离散程度定,使其平均大小在 95% 置信度下误差不超过 25%(每个来源每轮最多 400 行,还没有读数的来源取 20 行);其比例是比对过的行(状态 `match`、`size_mismatch` 或 `hash_mismatch`)中与记录的大小一致的行所占的比例,没能比对的行(没有取到文件)按状态另行计数(`unreachable`、`gone`、`needs_auth`、`no_length`、`error`;目录列 `size_check`)。一个来源没有记录大小的行的估计(目录列 `unsized_rows` 和 `unsized_estimate`)是这些行的行数乘以读过的这类行的平均实测大小,并给出 95% BCa 自助法范围(`scipy.stats.bootstrap`,用其默认重抽样次数);读过的这类行少于 10 行的来源没有估计。在来源之上的层级,估计是其各来源估计之和,范围是各来源范围之和,比联合区间更宽。每个层级的数字都在 `meta/size_figures.json` 中。
42
  - **给谁用:** 为终端任务挑选材料、然后自己下载文件的人和智能体。
43
  - **不是什么:** 不是现成的任务或环境列表,也不是文件的副本。
44
 
45
  ## 快速上手
46
 
47
+ 下面的一切都固定在发布标签 `release-2026-11-09.1` 上。
48
 
49
  1. 用一条命令下载文件([uv](https://docs.astral.sh/uv/) 会读取脚本声明的依赖):
50
 
51
  ```sh
52
+ uv run https://huggingface.co/datasets/Battam/agentic-env-index/resolve/release-2026-11-09.1/tools/fetch.py --form spreadsheets --sample 3 --out ./data
53
  ```
54
 
55
+ 这一步返回单元:形式为 `spreadsheets` 的每个材料集各 3 个单元,每个单元放在以它的 `unit_id` 命名的单独文件夹里。
56
 
57
  2. 用 [DuckDB](https://duckdb.org/) 从目录中挑选材料集:
58
 
59
  ```sql
60
+ SELECT source, collection, title, units, files, top_types, license.tier AS license_tier, needs_auth
61
+ FROM 'hf://datasets/Battam/agentic-env-index@release-2026-11-09.1/meta/catalog.parquet'
62
  WHERE kind = 'material_set' AND needs_auth = 'no' AND license.tier IN ('permissive', 'share-alike')
63
  ORDER BY collection, units DESC;
64
  ```
 
69
  3. 在一个语料库里挑选文件:
70
 
71
  ```sql
72
+ SET VARIABLE files = (SELECT list('hf://datasets/Battam/agentic-env-index@release-2026-11-09.1/' || f) FROM (SELECT unnest(data_files) AS f FROM 'hf://datasets/Battam/agentic-env-index@release-2026-11-09.1/meta/catalog.parquet' WHERE source = 'if-archive'));
73
  SELECT uri, size_bytes, lower(hex(sha256)) AS sha256
74
+ FROM read_parquet(getvariable('files'))
75
+ WHERE source = 'if-archive' AND ext = 'pdf' AND size_bytes BETWEEN 100000 AND 5000000
76
  ORDER BY size_bytes LIMIT 20;
77
  ```
78
 
79
  这一步返回语料库 `if-archive` 的行,每个文件一行。
80
+ 第一条语句从目录的 `data_files` 列取得数据文件,所以查询只读这些文件,不列出 `data/` 下的任何文件夹。
81
+ 一个数据文件可能存放几个来源的行,所以查询按 `source` 列只保留这个来源的行。
82
 
83
  使用示例视频有中文版和英文版,各带字幕,展示了本索引的用法:[在数据页播放](https://huggingface.co/spaces/Battam/agentic-env-catalog#usage-video)。
84
  数据页在“获取数据”部分播放��页面语言相同的视频和字幕。
 
91
 
92
  ### 任务材料,按合集
93
 
94
+ 13 个合集中的 1,799 个材料集,共 4,183 个单元。
95
+ 记录的大小 2.90 GB;去重后大小 2.82 GB;大小核对:99.7% 一致(读过 8,144 行,比对了 8,122 行);未比对:地址无法访问 10 行、文件已不存在 2 行、需要登录 8 行、其他错误 2 行。
96
  上游是一个合集的材料集所来自的上游作品。
97
 
98
+ | 合集 | 材料集数 | 单元数 | 文件数 | 上游数 |
99
  |---|---:|---:|---:|---:|
100
+ | 电子表格 | 67 | 536 | 1,200 | 55 |
101
+ | 磁盘镜像 | 52 | 179 | 470 | 40 |
102
+ | 复现包与运行记录 | 88 | 105 | 965 | 88 |
103
  | SEC 申报审阅 | 0 | 未知 | 未知 | 0 |
104
+ | 求解器实例 | 63 | 187 | 325 | 60 |
105
+ | 浏览器测试 | 123 | 439 | 852 | 88 |
106
+ | 硬件设计(EDA)测试 | 81 | 258 | 651 | 75 |
107
+ | 问题跟踪系统 | 212 | 258 | 682 | 212 |
108
+ | 媒体与文档 | 276 | 391 | 1,088 | 269 |
109
+ | 模型服务配置 | 102 | 397 | 1,169 | 82 |
110
+ | 规则计算 | 177 | 329 | 727 | 154 |
111
+ | 安全检测 | 110 | 328 | 983 | 77 |
112
+ | 科学参考数据 | 161 | 342 | 671 | 157 |
113
+ | 测试与练习 | 286 | 425 | 2,544 | 265 |
114
+ | 不属于任何合集 | 1 | 9 | 19 | 1 |
115
 
116
  ### 语料库
117
 
118
+ 138 个语料库,共 1,235,871,975 行。
119
+ 记录的大小 4.50 PB;去重后大小 4.09 PB;大小核对:99.7% 一致(读过 36,485 行,比对了 35,526 行);未比对:地址无法访问 456 行、文件已不存在 117 行、需要登录 323 行、服务器未给出大小 61 行、其他错误 2 行;没有记录大小的 311,970,416 行中 279,909,103 行的估计大小:583 TB(这是估计值;95% 范围 497 TB 到 701 TB,为其各来源范围之和,比联合区间更宽;依据读过的 15,964 行),其余 32,061,313 行读过的太少,无法估计。
120
  每个语料库计入它的主要领域。
121
 
122
  | 领域 | 语料库数 | 行数 | 去重后大小 |
123
  |---|---:|---:|---:|
124
+ | 数据科学 / 数据处理 / 查询 | 23 | 750,815,142 | 600 TB |
125
+ | 系统管理 / DevOps / 环境构建 | 29 | 185,547,838 | 182 TB |
126
+ | 多领域 | 16 | 109,251,421 | 224 TB |
127
+ | 软件工程 | 12 | 65,693,542 | 86.5 TB |
128
+ | 科学 | 15 | 44,541,599 | 251 TB |
129
+ | 业务运营流程 | 6 | 39,404,642 | 1.78 TB |
130
+ | 媒体 | 9 | 12,751,384 | 874 TB |
131
+ | 机器学习 / 模型运维 | 6 | 9,520,852 | 1.86 PB |
132
+ | 硬件 | 5 | 8,715,949 | 1.66 TB |
133
+ | 未分类 | 4 | 4,869,099 | 291 GB |
134
+ | 安全 | 10 | 4,519,397 | 11.9 TB |
135
  | 杂项 / 游戏 / 策略 / 约束调度 | 3 | 241,110 | 2.87 TB |
136
 
137
  - [数据页面](https://huggingface.co/spaces/Battam/agentic-env-catalog):按合集浏览任务材料,或浏览语料库,每次一层,并生成下载命令。
138
  - [数据集查看器中的 `catalog` 子集](https://huggingface.co/datasets/Battam/agentic-env-index/viewer/catalog):每个材料集或语料库一行,带有它的 `kind` 和 `collection`。
139
+ - [SQL 控制台](https://huggingface.co/datasets/Battam/agentic-env-index?sql_console=true&sql=--+every+source+of+the+catalog+as+published%3A+the+default+profile+of+tools%2Ffetch.py+is+not+applied+here%0ASELECT+source%2C+kind%2C+collection%2C+title%2C+units%2C+files%0AFROM+train%0AORDER+BY+kind%2C+collection%2C+files+DESC%0ALIMIT+50):在浏览器里查询目录;
140
+ 它按发布原样列出每个来源,不应用 `tools/fetch.py` 的默认配置,命令才应用它。
141
  - [全部材料集(按合集)和全部语料库](docs/sources.zh.md):每个是什么,它的计数、大小、许可证等级和登录要求。
142
 
143
  ## 数据集结构
144
 
145
+ `data/` 下数据文件的每一行指向一个文件或一条记录;
146
+ 目录的 `data_files` 列列出每个来源的文件。
147
  最常用的字段:
148
 
149
  | 字段 | 类型 | 含义 |
 
157
  | `license_spdx` | 文本 | 该行自己的许可证(SPDX);沿用所在批次的许可证时为空。 |
158
  | `batch` | 文本 | 该行发布时所在的批次;批次的许可证见 meta/sources.json。 |
159
  | `sensitive` | 文本 | 条目涉及健康或未成年人时为 health 或 minors。 |
160
+ | `extra` | 文本 | 该行其余字段的 JSON,例如它在集合中的单元分组。 |
161
 
162
  全部行字段和目录的全部列:[docs/fields.zh.md](docs/fields.zh.md)。
163
 
 
166
  - **直接用途:** 为构建智能体任务和环境挑选来源和文件,再从各自的来源下载每个文件,并按它所在行记录的大小和哈希值校验。
167
  - **不适用的用途:** 本数据集不是文件的副本,本身也不是训练语料:
168
  文件留在各自的来源,遵循各自的许可证。
169
+ 一道题检查什么,在用材料构建这道题时才决定;
170
+ 索引只存材料。
171
  - 如何用这些文件构建任务、`tools/fetch.py` 的默认配置会去掉什么、哪些内容不能放进智能体的环境:[docs/for-agents.zh.md](docs/for-agents.zh.md)。
172
 
173
  ## 数据集构建
174
 
175
  - **收集:** 行从每个来源的公开列表、API 和存档中收集,加入时即校验:
176
  地址必须能访问,大小和哈希值会重新取回并比对。
177
+ 材料集来自为每个合集对候选所做的搜索和收录;
178
+ 一组语料库行只有通过同样的收录才成为材料集。
179
  每个材料集和每个语料库及其计数:[docs/sources.zh.md](docs/sources.zh.md)。
180
  - **发布前的检查:** 每个来源都记录了许可证并归入一个等级,个人信息从样本文件中判读,各行与公开基准测试比对;
181
  复制了公开基准测试的行会被移除,不满足某项要求的单元会连同原因一起列出:[docs/access.zh.md](docs/access.zh.md)。
 
191
  - 每个文件保留其来源的许可证:见每一行的许可证列(没有自己许可证的行沿用所在批次的许可证,即 `meta/sources.json` 中的 `batches[].license`)。
192
  - 许可证等级按许可内容对每个来源的许可证分组;
193
  每个来源的等级以及每个等级的含义:[docs/sources.zh.md](docs/sources.zh.md)。
194
+ - 许可证等级和原因码由 ScanCode LicenseDB 的许可证类别算出,它以 CC-BY-4.0 许可;
195
+ 它的署名:[`NOTICE`](NOTICE)。
196
 
197
  ## 个人与敏感信息
198
 
 
200
  它本身不含任何个人记录。
201
  - 有些来源指向署有作者姓名的材料,例如邮件列表、缺陷跟踪系统和论坛。
202
  - 指向健康或未成年人相关数据的行,在行列 `sensitive` 中标有 `health` 或 `minors`,可以据此过滤掉。
203
+ - 从样本文件读出的个人信息:目前 1,937 个来源中已判读 18 个,每个来源读了 9 到 30 个样本文件,读取日期为 2026-10-04 和 2026-10-05;按等级:`none_known` 3, `present` 14, `suspected` 1。其余 1,919 个来源显示 `not judged yet`(尚未判读)。每个来源的卡片给出它的等级和所依据的��本文件(`cards/<source>.yaml` 里的 `risk.pii` 和 `risk.pii_evidence`)。
204
 
205
  ## 偏差、风险与局限
206
 
207
+ - 大小:1,937 个来源中有 1,890 个每一行都有大小。其余 47 个来源里,来源没有公布大小的行显示为未知,不计入大小合计,也从不按 0 计。
208
+ - 判读字段:1,937 个来源中有 0 个已从样本文件读出摘要和任务适配度,1,937 个中有 18 个已从样本文件读出个人信息(与上一节的计数相同);尚未判读的字段会注明。每个来源的计数、类型、大小和许可证都直接来自索引本身。
209
+ - 仍在增长的来源:1,937 个来源中有 25 个仍在收集中,它们的行数随每次发布增加。
210
 
211
  ## 更新与版本
212
 
213
+ - 最新发布:`release-2026-11-09.1`(其最新的 CHANGELOG 条目日期为 2026-11-09)。 从 2026-10-11 到 2026-11-09 的 30 天里,有 14 天发布过新版本。
214
  - 各次发布列在 [CHANGELOG.md](CHANGELOG.md) 中;
215
+ 每次发布是一个带有自己发布标签的提交(本次为 [`release-2026-11-09.1`](https://huggingface.co/datasets/Battam/agentic-env-index/tree/release-2026-11-09.1)),
216
  因此任何较早的状态都可以在它的标签处读到。
217
  - `meta/projection.json` 记录发布标签,以及生成本卡片的程序和模板的版本。
218
  - Croissant 元数据:[由 Hub 生成](https://huggingface.co/api/datasets/Battam/agentic-env-index/croissant);
 
242
  @misc{agentic_env_index,
243
  title = {agentic-env-index: an index of public materials for building agent environments},
244
  howpublished = {https://huggingface.co/datasets/Battam/agentic-env-index},
245
+ note = {Release release-2026-11-09.1},
246
  year = {2026}
247
  }
248
  ```
campaign_files/2026-10-b-disk-images-3/btrfs-progs-convert-needs-recovery/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether the unit needs other upstream files is not stated. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: corruption_desc.
12
-
13
  **Tools to open and check:** Open with tools for these file types: .raw.
14
 
15
  **Licence (SPDX):** GPL-2.0-only
@@ -20,6 +18,5 @@ Units:
20
 
21
  - `tests/convert-tests/020-refuse-needs-recovery/ext4.raw`: Btrfs-progs convert test image: ext4 with needs_recovery set (020-refuse-needs-recovery).
22
  Decompressed from: https://raw.githubusercontent.com/kdave/btrfs-progs/99e492926eb8a7eadee897677d987d628fe9d218/tests/convert-tests/020-refuse-needs-recovery/ext4.raw.xz
23
- Reference result: corruption_desc
24
  Companions (not included): tests/convert-tests/020-refuse-needs-recovery/test.sh
25
  Dependencies: btrfs-progs btrfs-convert and e2fsprogs dumpe2fs, only to rerun the upstream test
 
8
 
9
  **Self-contained:** Whether the unit needs other upstream files is not stated. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Open with tools for these file types: .raw.
12
 
13
  **Licence (SPDX):** GPL-2.0-only
 
18
 
19
  - `tests/convert-tests/020-refuse-needs-recovery/ext4.raw`: Btrfs-progs convert test image: ext4 with needs_recovery set (020-refuse-needs-recovery).
20
  Decompressed from: https://raw.githubusercontent.com/kdave/btrfs-progs/99e492926eb8a7eadee897677d987d628fe9d218/tests/convert-tests/020-refuse-needs-recovery/ext4.raw.xz
 
21
  Companions (not included): tests/convert-tests/020-refuse-needs-recovery/test.sh
22
  Dependencies: btrfs-progs btrfs-convert and e2fsprogs dumpe2fs, only to rerun the upstream test
campaign_files/2026-10-g-solver-instances-2/2dpacklib/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Known optimal strip height (perfect packing exists by construction); known optimal strip height (perfect packing). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Program: read the agent placement (x_i, y_i lower-left corners, items not rotated: 2D-SPP oriented, 2SP|O|F on the 2DPackLib page); feasible when every item i (demand 1 each) lies inside 0 <= x_i, x_i + w_i <= 200 and 0 <= y_i, and no two items overlap in an area of positive size; height = max over items of y_i + h_i; pass when feasible and height <= 200 (the known optimum; the area bound makes a lower height impossible, so lower means a checker error); no guillotine condition is required (set N is non-guillotineable); Program: read the agent placement (x_i, y_i lower-left corners, items not rotated: 2D-SPP oriented as on the 2DPackLib page); feasible when every item lies inside 0 <= x_i, x_i + w_i <= 30 and 0 <= y_i, and no two items overlap in an area of positive size; height = max over items of y_i + h_i; pass when feasible and height <= 50 (the optimum; the area bound makes a lower height impossible, so lower means a checker error).
14
 
15
  **Licence (SPDX):** NOASSERTION AND CC0-1.0 AND NOASSERTION
@@ -21,16 +19,12 @@ Files:
21
  - `N_T/N_T/N4a.ins2D`: 2D strip packing instance N4a (Hopper 2000 set N, 49 rectangles, strip width 200, known optimal height 200).
22
  Format: 2DPackLib ins2D plain text: m, then W H, then per item: id w h demand max_copies profit (N_T/format_description.txt); one case = one instance; the answer is a placement (x, y) of every item
23
  Origin: published benchmark instance generated by Hopper (2000) and converted to the 2DPackLib format; recorded tool_fixture
24
- Reference result: known optimal strip height (perfect packing exists by construction)
25
- Reference source: ESICUP t_n/readme.txt line 3: "This set of guillotineable and non-guillotineable test problems with known optima was constructed with the problem generators described in Hopper (2000)"; 2DPackLib sets H=200 in the file (2D orthogonal packing use, Cote and Iori 2018 on the 2DPackLib page).
26
  Check: Program: read the agent placement (x_i, y_i lower-left corners, items not rotated: 2D-SPP oriented, 2SP|O|F on the 2DPackLib page); feasible when every item i (demand 1 each) lies inside 0 <= x_i, x_i + w_i <= 200 and 0 <= y_i, and no two items overlap in an area of positive size; height = max over items of y_i + h_i; pass when feasible and height <= 200 (the known optimum; the area bound makes a lower height impossible, so lower means a checker error); no guillotine condition is required (set N is non-guillotineable)
27
  Companions (not included): N_T/format_description.txt; esicup/2d_rectangular/t_n/readme.txt
28
  Dependencies: no upstream code used
29
  - `BKW/BKW/BKW02.ins2D`: 2D strip packing instance BKW02 (Burke, Kendall and Whitwell 2004, 20 rectangles, strip width 30, optimal height 50).
30
  Format: 2DPackLib ins2D plain text: m, then W H, then per item: id w h demand max_copies profit (BKW/format_description.txt); one case = one instance; the answer is a placement (x, y) of every item
31
  Origin: published benchmark instance generated by Burke, Kendall and Whitwell (2004) and converted to the 2DPackLib format; recorded tool_fixture
32
- Reference result: known optimal strip height (perfect packing)
33
- Reference source: 2DPackLib page: "BKW These are the instances originally proposed and solved as a 2D-SPP by Burke, Kendall, and Whitwell (2004)"; BKW02.ins2D line 2 gives H = 50.
34
  Check: Program: read the agent placement (x_i, y_i lower-left corners, items not rotated: 2D-SPP oriented as on the 2DPackLib page); feasible when every item lies inside 0 <= x_i, x_i + w_i <= 30 and 0 <= y_i, and no two items overlap in an area of positive size; height = max over items of y_i + h_i; pass when feasible and height <= 50 (the optimum; the area bound makes a lower height impossible, so lower means a checker error)
35
  Companions (not included): BKW/format_description.txt
36
  Dependencies: no upstream code used
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Program: read the agent placement (x_i, y_i lower-left corners, items not rotated: 2D-SPP oriented, 2SP|O|F on the 2DPackLib page); feasible when every item i (demand 1 each) lies inside 0 <= x_i, x_i + w_i <= 200 and 0 <= y_i, and no two items overlap in an area of positive size; height = max over items of y_i + h_i; pass when feasible and height <= 200 (the known optimum; the area bound makes a lower height impossible, so lower means a checker error); no guillotine condition is required (set N is non-guillotineable); Program: read the agent placement (x_i, y_i lower-left corners, items not rotated: 2D-SPP oriented as on the 2DPackLib page); feasible when every item lies inside 0 <= x_i, x_i + w_i <= 30 and 0 <= y_i, and no two items overlap in an area of positive size; height = max over items of y_i + h_i; pass when feasible and height <= 50 (the optimum; the area bound makes a lower height impossible, so lower means a checker error).
12
 
13
  **Licence (SPDX):** NOASSERTION AND CC0-1.0 AND NOASSERTION
 
19
  - `N_T/N_T/N4a.ins2D`: 2D strip packing instance N4a (Hopper 2000 set N, 49 rectangles, strip width 200, known optimal height 200).
20
  Format: 2DPackLib ins2D plain text: m, then W H, then per item: id w h demand max_copies profit (N_T/format_description.txt); one case = one instance; the answer is a placement (x, y) of every item
21
  Origin: published benchmark instance generated by Hopper (2000) and converted to the 2DPackLib format; recorded tool_fixture
 
 
22
  Check: Program: read the agent placement (x_i, y_i lower-left corners, items not rotated: 2D-SPP oriented, 2SP|O|F on the 2DPackLib page); feasible when every item i (demand 1 each) lies inside 0 <= x_i, x_i + w_i <= 200 and 0 <= y_i, and no two items overlap in an area of positive size; height = max over items of y_i + h_i; pass when feasible and height <= 200 (the known optimum; the area bound makes a lower height impossible, so lower means a checker error); no guillotine condition is required (set N is non-guillotineable)
23
  Companions (not included): N_T/format_description.txt; esicup/2d_rectangular/t_n/readme.txt
24
  Dependencies: no upstream code used
25
  - `BKW/BKW/BKW02.ins2D`: 2D strip packing instance BKW02 (Burke, Kendall and Whitwell 2004, 20 rectangles, strip width 30, optimal height 50).
26
  Format: 2DPackLib ins2D plain text: m, then W H, then per item: id w h demand max_copies profit (BKW/format_description.txt); one case = one instance; the answer is a placement (x, y) of every item
27
  Origin: published benchmark instance generated by Burke, Kendall and Whitwell (2004) and converted to the 2DPackLib format; recorded tool_fixture
 
 
28
  Check: Program: read the agent placement (x_i, y_i lower-left corners, items not rotated: 2D-SPP oriented as on the 2DPackLib page); feasible when every item lies inside 0 <= x_i, x_i + w_i <= 30 and 0 <= y_i, and no two items overlap in an area of positive size; height = max over items of y_i + h_i; pass when feasible and height <= 50 (the optimum; the area bound makes a lower height impossible, so lower means a checker error)
29
  Companions (not included): BKW/format_description.txt
30
  Dependencies: no upstream code used
campaign_files/2026-10-g-solver-instances-2/asiekierka-free-game-levels/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 19 of 19 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Solvability (no solution shipped); success = every box on a goal. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
14
 
15
  **Licence (SPDX):** LicenseRef-DavidSkinner-Sokoban-Sets AND LicenseRef-DavidHolland-Sokoban AND LicenseRef-EvgenyGrigoriev-Sokoban AND LicenseRef-RickSladkey-Sokoban
@@ -21,152 +19,114 @@ Files:
21
  - `free/sokoban/davidskinner/microban/m1.txt`: Microban I: 155 Sokoban levels by David W. Skinner.
22
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 155 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
23
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Microban I), mirrored as-is
24
- Reference result: solvability (no solution shipped); success = every box on a goal
25
- Reference source: No solutions ship. The author's README line 7: "Be assured that (except for unfinished sets) ALL are solvable." This set is a finished set.
26
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
27
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
28
  Dependencies: not packed
29
  - `free/sokoban/davidskinner/microban/m2.txt`: Microban II: 135 Sokoban levels by David W. Skinner.
30
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 135 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
31
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Microban II), mirrored as-is
32
- Reference result: solvability (no solution shipped); success = every box on a goal
33
- Reference source: No solutions ship. The author's README line 7: "Be assured that (except for unfinished sets) ALL are solvable." This set is a finished set.
34
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
35
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
36
  Dependencies: not packed
37
  - `free/sokoban/davidskinner/microban/m3.txt`: Microban III: 101 Sokoban levels by David W. Skinner.
38
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 101 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
39
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Microban III), mirrored as-is
40
- Reference result: solvability (no solution shipped); success = every box on a goal
41
- Reference source: No solutions ship. The author's README line 7: "Be assured that (except for unfinished sets) ALL are solvable." This set is a finished set.
42
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
43
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
44
  Dependencies: not packed
45
  - `free/sokoban/davidskinner/microban/m4.txt`: Microban IV: 102 Sokoban levels by David W. Skinner.
46
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 102 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
47
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Microban IV), mirrored as-is
48
- Reference result: solvability (no solution shipped); success = every box on a goal
49
- Reference source: No solutions ship. The author's README line 7: "Be assured that (except for unfinished sets) ALL are solvable." This set is a finished set.
50
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
51
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
52
  Dependencies: not packed
53
  - `free/sokoban/davidskinner/sasquatch/s1.txt`: Sasquatch I: 50 Sokoban levels by David W. Skinner.
54
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
55
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Sasquatch), mirrored as-is
56
- Reference result: solvability (no solution shipped); success = every box on a goal
57
- Reference source: No solutions ship. The author's README line 7: "Be assured that (except for unfinished sets) ALL are solvable." This set is a finished set.
58
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
59
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
60
  Dependencies: not packed
61
  - `free/sokoban/davidskinner/sasquatch/s2.txt`: Sasquatch II: 50 Sokoban levels by David W. Skinner.
62
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
63
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Sasquatch II), mirrored as-is
64
- Reference result: solvability (no solution shipped); success = every box on a goal
65
- Reference source: No solutions ship. The author's README line 7: "Be assured that (except for unfinished sets) ALL are solvable." This set is a finished set.
66
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
67
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
68
  Dependencies: not packed
69
  - `free/sokoban/davidskinner/sasquatch/s3.txt`: Sasquatch III: 50 Sokoban levels by David W. Skinner.
70
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
71
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Sasquatch III), mirrored as-is
72
- Reference result: solvability (no solution shipped); success = every box on a goal
73
- Reference source: No solutions ship. The author's README line 7: "Be assured that (except for unfinished sets) ALL are solvable." This set is a finished set.
74
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
75
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
76
  Dependencies: not packed
77
  - `free/sokoban/davidskinner/sasquatch/s4.txt`: Sasquatch IV: 50 Sokoban levels by David W. Skinner.
78
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
79
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Sasquatch IV), mirrored as-is
80
- Reference result: solvability (no solution shipped); success = every box on a goal
81
- Reference source: No solutions ship. The author's README line 7: "Be assured that (except for unfinished sets) ALL are solvable." This set is a finished set.
82
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
83
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
84
  Dependencies: not packed
85
  - `free/sokoban/davidholland/bagatelle/bagatelle.txt`: Bagatelle: 20 Sokoban levels by David Holland (2001).
86
  Format: plain text, XSB level notation, levels introduced by ';N' lines, Latin-1 bytes for the © sign; 20 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
87
  Origin: Sokoban levels published by their author David Holland: hand-designed levels; mirrored as-is
88
- Reference result: solvability (no solution shipped); success = every box on a goal
89
- Reference source: No solutions ship and the author does not state solvability for every level.
90
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
91
  Companions (not included): free/sokoban/davidholland/LICENSE.txt; free/sokoban/davidholland/bagatelle/README.md
92
  Dependencies: not packed
93
  - `free/sokoban/davidholland/cantrip/cantrip.txt`: Cantrip: 20 Sokoban levels by David Holland (2001).
94
  Format: plain text, XSB level notation, levels introduced by ';N' lines, Latin-1 bytes for the © sign; 20 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
95
  Origin: Sokoban levels published by their author David Holland: maze and goals hand-designed, start positions computer-generated by the author's own program (author's README); mirrored as-is
96
- Reference result: solvability (no solution shipped); success = every box on a goal
97
- Reference source: No solutions ship and the author does not state solvability for every level.
98
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
99
  Companions (not included): free/sokoban/davidholland/LICENSE.txt; free/sokoban/davidholland/cantrip/README.md
100
  Dependencies: not packed
101
  - `free/sokoban/davidholland/dh1/dh1.txt`: dh1: 10 Sokoban levels by David Holland (2000).
102
  Format: plain text, XSB level notation, levels introduced by ';N' lines, Latin-1 bytes for the © sign; 10 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
103
  Origin: Sokoban levels published by their author David Holland: partly co-authored with the author's generator sokgen, which searched backwards from the solved position (author's README), so those levels are solvable by construction; mirrored as-is
104
- Reference result: solvability (no solution shipped); success = every box on a goal
105
- Reference source: No solutions ship and the author does not state solvability for every level (the README says sokgen produced push-minimal solutions for the co-authored levels).
106
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
107
  Companions (not included): free/sokoban/davidholland/LICENSE.txt; free/sokoban/davidholland/dh1/README.md
108
  Dependencies: not packed
109
  - `free/sokoban/davidholland/maelstrom/maelstrm.txt`: Maelstrom: 21 Sokoban levels by David Holland.
110
  Format: plain text, XSB level notation, levels introduced by ';N' lines, Latin-1 bytes for the © sign; 21 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
111
  Origin: Sokoban levels published by their author David Holland: maze and goals hand-designed, many small levels with computer-generated start positions (author's README); mirrored as-is
112
- Reference result: solvability (no solution shipped); success = every box on a goal
113
- Reference source: No solutions ship and the author does not state solvability for every level.
114
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
115
  Companions (not included): free/sokoban/davidholland/LICENSE.txt; free/sokoban/davidholland/maelstrom/README.md
116
  Dependencies: not packed
117
  - `free/sokoban/grigr/Grigr2001/grigr2001.txt`: Grigr2001: 100 Sokoban levels by Evgeny Grigoriev.
118
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 100 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
119
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
120
- Reference result: solvability (no solution shipped); success = every box on a goal
121
- Reference source: No solutions ship.
122
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
123
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md; free/sokoban/grigr/Grigr2001/readme.txt
124
  Dependencies: not packed
125
  - `free/sokoban/grigr/Grigr2002/grigr2002.txt`: Grigr2002: 40 Sokoban levels by Evgeny Grigoriev.
126
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 40 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
127
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
128
- Reference result: solvability (no solution shipped); success = every box on a goal
129
- Reference source: No solutions ship.
130
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
131
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md; free/sokoban/grigr/Grigr2002/README.TXT
132
  Dependencies: not packed
133
  - `free/sokoban/grigr/GrigrComet/GrigrComet.txt`: GrigrComet: 30 Sokoban levels by Evgeny Grigoriev.
134
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 30 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
135
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
136
- Reference result: solvability (no solution shipped); success = every box on a goal
137
- Reference source: No solutions ship.
138
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
139
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md
140
  Dependencies: not packed
141
  - `free/sokoban/grigr/GrigrSpecial/GrigrSpecial.txt`: GrigrSpecial: 40 Sokoban levels by Evgeny Grigoriev.
142
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 40 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
143
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
144
- Reference result: solvability (no solution shipped); success = every box on a goal
145
- Reference source: No solutions ship.
146
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
147
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md
148
  Dependencies: not packed
149
  - `free/sokoban/grigr/GrigrStar/GrigrStar.txt`: GrigrStar: 30 Sokoban levels by Evgeny Grigoriev.
150
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 30 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
151
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
152
- Reference result: solvability (no solution shipped); success = every box on a goal
153
- Reference source: No solutions ship.
154
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
155
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md
156
  Dependencies: not packed
157
  - `free/sokoban/grigr/GrigrSun/GrigrSun.txt`: GrigrSun: 10 Sokoban levels by Evgeny Grigoriev.
158
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 10 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
159
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
160
- Reference result: solvability (no solution shipped); success = every box on a goal
161
- Reference source: No solutions ship.
162
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
163
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md
164
  Dependencies: not packed
165
  - `free/sokoban/ricksladkey/Warehouse-Volume-I/Warehouse-Volume-I.txt`: The Warehouse Volume I, Vertical Mobility: 50 Sokoban levels by Rick Sladkey (2009).
166
  Format: plain text, CRLF, XSB level notation with a header block, levels introduced by quoted titles ('Level N'); 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
167
  Origin: hand-made Sokoban levels published by their author Rick Sladkey in 2009, mirrored as-is; the README says ten of the fifty are his own versions of the LOMA designs by Aymeric du Peloux
168
- Reference result: solvability (no solution shipped); success = every box on a goal
169
- Reference source: No solutions ship and the author does not state solvability.
170
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
171
  Companions (not included): free/sokoban/ricksladkey/LICENSE.txt; free/sokoban/ricksladkey/Warehouse-Volume-I/README.md
172
  Dependencies: not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 19 of 19 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
12
 
13
  **Licence (SPDX):** LicenseRef-DavidSkinner-Sokoban-Sets AND LicenseRef-DavidHolland-Sokoban AND LicenseRef-EvgenyGrigoriev-Sokoban AND LicenseRef-RickSladkey-Sokoban
 
19
  - `free/sokoban/davidskinner/microban/m1.txt`: Microban I: 155 Sokoban levels by David W. Skinner.
20
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 155 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
21
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Microban I), mirrored as-is
 
 
22
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
23
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
24
  Dependencies: not packed
25
  - `free/sokoban/davidskinner/microban/m2.txt`: Microban II: 135 Sokoban levels by David W. Skinner.
26
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 135 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
27
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Microban II), mirrored as-is
 
 
28
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
29
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
30
  Dependencies: not packed
31
  - `free/sokoban/davidskinner/microban/m3.txt`: Microban III: 101 Sokoban levels by David W. Skinner.
32
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 101 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
33
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Microban III), mirrored as-is
 
 
34
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
35
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
36
  Dependencies: not packed
37
  - `free/sokoban/davidskinner/microban/m4.txt`: Microban IV: 102 Sokoban levels by David W. Skinner.
38
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 102 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
39
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Microban IV), mirrored as-is
 
 
40
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
41
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
42
  Dependencies: not packed
43
  - `free/sokoban/davidskinner/sasquatch/s1.txt`: Sasquatch I: 50 Sokoban levels by David W. Skinner.
44
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
45
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Sasquatch), mirrored as-is
 
 
46
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
47
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
48
  Dependencies: not packed
49
  - `free/sokoban/davidskinner/sasquatch/s2.txt`: Sasquatch II: 50 Sokoban levels by David W. Skinner.
50
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
51
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Sasquatch II), mirrored as-is
 
 
52
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
53
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
54
  Dependencies: not packed
55
  - `free/sokoban/davidskinner/sasquatch/s3.txt`: Sasquatch III: 50 Sokoban levels by David W. Skinner.
56
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
57
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Sasquatch III), mirrored as-is
 
 
58
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
59
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
60
  Dependencies: not packed
61
  - `free/sokoban/davidskinner/sasquatch/s4.txt`: Sasquatch IV: 50 Sokoban levels by David W. Skinner.
62
  Format: plain text, XSB level notation, levels introduced by '; N' lines; 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
63
  Origin: hand-made Sokoban levels published by their author David W. Skinner (Sasquatch IV), mirrored as-is
 
 
64
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
65
  Companions (not included): free/sokoban/davidskinner/LICENSE; free/sokoban/davidskinner/README
66
  Dependencies: not packed
67
  - `free/sokoban/davidholland/bagatelle/bagatelle.txt`: Bagatelle: 20 Sokoban levels by David Holland (2001).
68
  Format: plain text, XSB level notation, levels introduced by ';N' lines, Latin-1 bytes for the © sign; 20 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
69
  Origin: Sokoban levels published by their author David Holland: hand-designed levels; mirrored as-is
 
 
70
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
71
  Companions (not included): free/sokoban/davidholland/LICENSE.txt; free/sokoban/davidholland/bagatelle/README.md
72
  Dependencies: not packed
73
  - `free/sokoban/davidholland/cantrip/cantrip.txt`: Cantrip: 20 Sokoban levels by David Holland (2001).
74
  Format: plain text, XSB level notation, levels introduced by ';N' lines, Latin-1 bytes for the © sign; 20 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
75
  Origin: Sokoban levels published by their author David Holland: maze and goals hand-designed, start positions computer-generated by the author's own program (author's README); mirrored as-is
 
 
76
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
77
  Companions (not included): free/sokoban/davidholland/LICENSE.txt; free/sokoban/davidholland/cantrip/README.md
78
  Dependencies: not packed
79
  - `free/sokoban/davidholland/dh1/dh1.txt`: dh1: 10 Sokoban levels by David Holland (2000).
80
  Format: plain text, XSB level notation, levels introduced by ';N' lines, Latin-1 bytes for the © sign; 10 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
81
  Origin: Sokoban levels published by their author David Holland: partly co-authored with the author's generator sokgen, which searched backwards from the solved position (author's README), so those levels are solvable by construction; mirrored as-is
 
 
82
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
83
  Companions (not included): free/sokoban/davidholland/LICENSE.txt; free/sokoban/davidholland/dh1/README.md
84
  Dependencies: not packed
85
  - `free/sokoban/davidholland/maelstrom/maelstrm.txt`: Maelstrom: 21 Sokoban levels by David Holland.
86
  Format: plain text, XSB level notation, levels introduced by ';N' lines, Latin-1 bytes for the © sign; 21 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
87
  Origin: Sokoban levels published by their author David Holland: maze and goals hand-designed, many small levels with computer-generated start positions (author's README); mirrored as-is
 
 
88
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
89
  Companions (not included): free/sokoban/davidholland/LICENSE.txt; free/sokoban/davidholland/maelstrom/README.md
90
  Dependencies: not packed
91
  - `free/sokoban/grigr/Grigr2001/grigr2001.txt`: Grigr2001: 100 Sokoban levels by Evgeny Grigoriev.
92
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 100 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
93
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
 
 
94
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
95
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md; free/sokoban/grigr/Grigr2001/readme.txt
96
  Dependencies: not packed
97
  - `free/sokoban/grigr/Grigr2002/grigr2002.txt`: Grigr2002: 40 Sokoban levels by Evgeny Grigoriev.
98
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 40 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
99
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
 
 
100
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
101
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md; free/sokoban/grigr/Grigr2002/README.TXT
102
  Dependencies: not packed
103
  - `free/sokoban/grigr/GrigrComet/GrigrComet.txt`: GrigrComet: 30 Sokoban levels by Evgeny Grigoriev.
104
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 30 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
105
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
 
 
106
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
107
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md
108
  Dependencies: not packed
109
  - `free/sokoban/grigr/GrigrSpecial/GrigrSpecial.txt`: GrigrSpecial: 40 Sokoban levels by Evgeny Grigoriev.
110
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 40 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
111
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
 
 
112
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
113
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md
114
  Dependencies: not packed
115
  - `free/sokoban/grigr/GrigrStar/GrigrStar.txt`: GrigrStar: 30 Sokoban levels by Evgeny Grigoriev.
116
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 30 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
117
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
 
 
118
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
119
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md
120
  Dependencies: not packed
121
  - `free/sokoban/grigr/GrigrSun/GrigrSun.txt`: GrigrSun: 10 Sokoban levels by Evgeny Grigoriev.
122
  Format: plain text, CRLF, XSB level notation, levels introduced by ';NN' comment lines; 10 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
123
  Origin: hand-made Sokoban levels published by their author Evgeny Grigoriev on grigr.narod.ru, mirrored as-is
 
 
124
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
125
  Companions (not included): free/sokoban/grigr/LICENSE.txt; free/sokoban/grigr/README.md
126
  Dependencies: not packed
127
  - `free/sokoban/ricksladkey/Warehouse-Volume-I/Warehouse-Volume-I.txt`: The Warehouse Volume I, Vertical Mobility: 50 Sokoban levels by Rick Sladkey (2009).
128
  Format: plain text, CRLF, XSB level notation with a header block, levels introduced by quoted titles ('Level N'); 50 levels; a case is one level; delivered whole as the upstream file; the judge splits levels at blank, comment or title lines (lines that are not made only of the XSB cell characters # @ + $ * . space)
129
  Origin: hand-made Sokoban levels published by their author Rick Sladkey in 2009, mirrored as-is; the README says ten of the fifty are his own versions of the LOMA designs by Aymeric du Peloux
 
 
130
  Check: correct when every move is legal and every box stands on a goal. Any legal sequence counts (rule check by simulation, uniqueness not required, rule G of the spec), so no reference solution and no solver are needed at judge time.
131
  Companions (not included): free/sokoban/ricksladkey/LICENSE.txt; free/sokoban/ricksladkey/Warehouse-Volume-I/README.md
132
  Dependencies: not packed
campaign_files/2026-10-g-solver-instances-2/berthier-sudoku-classif-dob39c/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Unique solution grid. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: No solver is needed at judge time.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `dob-2650-39c/puzzles.txt`: All 2,650 known minimal 39-clue sudoku puzzles (Mladen Dobrichev), as republished in Sudoku-classif.
22
  Format: plain text, 2,650 lines of 81 puzzle characters, LF; a case is one line.
23
  Origin: Mladen Dobrichev's collection of the currently known minimal 39-clue puzzles ([c39] in README section 7), republished as-is by Denis Berthier
24
- Reference result: unique solution grid
25
- Reference source: No solution file ships in this folder.
26
  Check: No solver is needed at judge time.
27
  Dependencies: not packed; path and commit filled when the script is committed, before the run; not packed
 
8
 
9
  **Self-contained:** The file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: No solver is needed at judge time.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `dob-2650-39c/puzzles.txt`: All 2,650 known minimal 39-clue sudoku puzzles (Mladen Dobrichev), as republished in Sudoku-classif.
20
  Format: plain text, 2,650 lines of 81 puzzle characters, LF; a case is one line.
21
  Origin: Mladen Dobrichev's collection of the currently known minimal 39-clue puzzles ([c39] in README section 7), republished as-is by Denis Berthier
 
 
22
  Check: No solver is needed at judge time.
23
  Dependencies: not packed; path and commit filled when the script is committed, before the run; not packed
campaign_files/2026-10-g-solver-instances-2/berthier-sudoku-classif-eleven-te2/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Unique solution grid. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Check grid validity and agreement with every given (or exact match with solutions.txt).
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,8 +19,6 @@ Files:
21
  - `eleven-26370-TE2/puzzles.txt`: 26,370 minimal T&E(2) (very hard) sudoku puzzles by "eleven", with full solution grids and Sudoku Explainer ratings.
22
  Format: plain text, 26,370 lines, one 81-character puzzle per line; delivered whole as the upstream file; a case is one line, split on LF; its answer is the same line of the companion solutions.txt (one 81-digit grid per line)
23
  Origin: 26,370 minimal puzzles of T&E depth 2 from the collection by the forum user eleven, republished as-is with solutions by Denis Berthier in Sudoku-classif (was curated_collection)
24
- Reference result: unique solution grid
25
- Reference source: solutions.txt (Berthier); README 3.1: minimal puzzles in T&E(2).
26
  Check: Check grid validity and agreement with every given (or exact match with solutions.txt).
27
  Companions (not included): eleven-26370-TE2/solutions.txt
28
  Dependencies: not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Check grid validity and agreement with every given (or exact match with solutions.txt).
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `eleven-26370-TE2/puzzles.txt`: 26,370 minimal T&E(2) (very hard) sudoku puzzles by "eleven", with full solution grids and Sudoku Explainer ratings.
20
  Format: plain text, 26,370 lines, one 81-character puzzle per line; delivered whole as the upstream file; a case is one line, split on LF; its answer is the same line of the companion solutions.txt (one 81-digit grid per line)
21
  Origin: 26,370 minimal puzzles of T&E depth 2 from the collection by the forum user eleven, republished as-is with solutions by Denis Berthier in Sudoku-classif (was curated_collection)
 
 
22
  Check: Check grid validity and agreement with every given (or exact match with solutions.txt).
23
  Companions (not included): eleven-26370-TE2/solutions.txt
24
  Dependencies: not packed
campaign_files/2026-10-g-solver-instances-2/berthier-sudoku-classif-royle-17c/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Unique solution grid (plus difficulty ratings). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: No solver is needed at judge time.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `Royle-49158-17c/puzzles.txt`: All 49,158 minimal 17-clue sudoku puzzles (Royle collection) with Sudoku Explainer and W ratings.
22
  Format: plain text, 49,158 lines of 97 bytes: 81 puzzle characters, a space, ED=x/y/z (three difficulty ratings), LF; delivered whole as the upstream file; a case is one line, split on LF, the puzzle is the first 81 characters.
23
  Origin: the known collection of 17-clue sudoku puzzles (Gordon Royle's list, collected by search over many years by many people), republished as-is by Denis Berthier in Sudoku-classif (was curated_collection)
24
- Reference result: unique solution grid (plus difficulty ratings)
25
- Reference source: No solution file ships in this folder. Each 17-clue puzzle has a unique solution (README 5.1).
26
  Check: No solver is needed at judge time.
27
  Dependencies: not packed; path and commit filled when the script is committed, before the run; not packed
 
8
 
9
  **Self-contained:** The file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: No solver is needed at judge time.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `Royle-49158-17c/puzzles.txt`: All 49,158 minimal 17-clue sudoku puzzles (Royle collection) with Sudoku Explainer and W ratings.
20
  Format: plain text, 49,158 lines of 97 bytes: 81 puzzle characters, a space, ED=x/y/z (three difficulty ratings), LF; delivered whole as the upstream file; a case is one line, split on LF, the puzzle is the first 81 characters.
21
  Origin: the known collection of 17-clue sudoku puzzles (Gordon Royle's list, collected by search over many years by many people), republished as-is by Denis Berthier in Sudoku-classif (was curated_collection)
 
 
22
  Check: No solver is needed at judge time.
23
  Dependencies: not packed; path and commit filled when the script is committed, before the run; not packed
campaign_files/2026-10-g-solver-instances-2/biq-mac-library/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Proven optimal objective value. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Open with tools for these file types: .sparse.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `library/biq/be/be100.1.sparse`: Biq Mac Library unconstrained binary quadratic instance be100.1 (n=100, density 1.0) with proven optimum -19412.
22
  Format: sparse text: first line n and nonzero count, then i j q_ij lines (minimise x'Qx over x in {0,1}^n); answer in companion biqmaclib.tex table
23
  Origin: instance produced upstream by a generator: Billionnet and Elloumi (2007) with the Pardalos-Rodgers generator, seed 1 (biqmaclib.tex lines 300 to 309); collected with proven optima by Biq Mac. was generated until then.
24
- Reference result: proven optimal objective value
25
- Reference source: biqmaclib.tex line 325: "be100.1 & -19412 \\"; lines 311 to 313 say the optimal solution values are in these tables.
26
  Companions (not included): library/biqmaclib.tex
27
  Dependencies: not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Open with tools for these file types: .sparse.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `library/biq/be/be100.1.sparse`: Biq Mac Library unconstrained binary quadratic instance be100.1 (n=100, density 1.0) with proven optimum -19412.
20
  Format: sparse text: first line n and nonzero count, then i j q_ij lines (minimise x'Qx over x in {0,1}^n); answer in companion biqmaclib.tex table
21
  Origin: instance produced upstream by a generator: Billionnet and Elloumi (2007) with the Pardalos-Rodgers generator, seed 1 (biqmaclib.tex lines 300 to 309); collected with proven optima by Biq Mac. was generated until then.
 
 
22
  Companions (not included): library/biqmaclib.tex
23
  Dependencies: not packed
campaign_files/2026-10-g-solver-instances-2/biqmac-library/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 7 of 7 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Published optimal objective value. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Tools are named in the check written per file in the notes.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,56 +19,42 @@ Files:
21
  - `library/mac/rudy/g05_60.0`: Biq Mac Library max-cut instance g05_60.0: unweighted G(n,1/2) random graph, 60 vertices, 885 edges, generated with rudy (command at biqmaclib.tex line 1098), with the proven optimal cut value.
22
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
23
  Origin: benchmark instance generated with the rudy graph generator (or taken from a physics application) and published with its optimum by the library authors for testing exact max-cut and BQP solvers
24
- Reference result: published optimal objective value
25
- Reference source: biqmaclib.tex line 519: "g05\_60.0 & 536" in the max-cut tables (table caption line 557: unweighted graphs with edge probability 1/2); lines 33 to 37 say the tables list "the optimal solution values" and give lower/upper bounds where the optimum is not known, which the tables do in a separate "lower bound" column (lines 140 and 285); this table has only the "solution" column; the problem is stated at lines 475 to 481 as max x^T L x over x in {-1,1}^n with L the Laplace matrix. Values published by the library authors, proved optimal with the Biq Mac solver.
26
  Check: Exact: the agent returns a side (0 or 1) for every vertex; the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file and must equal 536; since 536 is a proved optimum, a larger value signals a reading error
27
  Companions (not included): biqmaclib.tex
28
  Dependencies: not packed
29
  - `library/mac/rudy/pm1s_80.0`: Biq Mac Library max-cut instance pm1s_80.0: sparse graph with edge weights in {-1,0,1}, density 10%, 80 vertices, 316 edges, generated with rudy (command at biqmaclib.tex line 1131), with the proven optimal cut value.
30
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
31
  Origin: benchmark instance generated with the rudy graph generator (or taken from a physics application) and published with its optimum by the library authors for testing exact max-cut and BQP solvers
32
- Reference result: published optimal objective value
33
- Reference source: biqmaclib.tex line 573: "pm1s\_80.0 & 79" in the max-cut tables (table caption line 633: G_{-1/0/1}, density 10% and 99%); lines 33 to 37 say the tables list "the optimal solution values" and give lower/upper bounds where the optimum is not known, which the tables do in a separate "lower bound" column (lines 140 and 285); this table has only the "solution" column; the problem is stated at lines 475 to 481 as max x^T L x over x in {-1,1}^n with L the Laplace matrix. Values published by the library authors, proved optimal with the Biq Mac solver.
34
  Check: Exact: the agent returns a side (0 or 1) for every vertex; the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file and must equal 79; since 79 is a proved optimum, a larger value signals a reading error
35
  Companions (not included): biqmaclib.tex
36
  Dependencies: not packed
37
  - `library/mac/rudy/w01_100.0`: Biq Mac Library max-cut instance w01_100.0: graph with integer edge weights in [-10,10], density 10%, 100 vertices, 495 edges, generated with rudy (command at biqmaclib.tex line 1175), with the proven optimal cut value.
38
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
39
  Origin: benchmark instance generated with the rudy graph generator (or taken from a physics application) and published with its optimum by the library authors for testing exact max-cut and BQP solvers
40
- Reference result: published optimal objective value
41
- Reference source: biqmaclib.tex line 647: "w01\_100.0 & 651" in the max-cut tables (table caption line 733: G_{[-10,10]} (w) and G_{[1,10]} (pw)); lines 33 to 37 say the tables list "the optimal solution values" and give lower/upper bounds where the optimum is not known, which the tables do in a separate "lower bound" column (lines 140 and 285); this table has only the "solution" column; the problem is stated at lines 475 to 481 as max x^T L x over x in {-1,1}^n with L the Laplace matrix. Values published by the library authors, proved optimal with the Biq Mac solver.
42
  Check: Exact: the agent returns a side (0 or 1) for every vertex; the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file and must equal 651; since 651 is a proved optimum, a larger value signals a reading error
43
  Companions (not included): biqmaclib.tex
44
  Dependencies: not packed
45
  - `library/mac/ising/ising2.5-100_5555`: Biq Mac Library max-cut instance ising2.5-100_5555: one-dimensional Ising chain instance from Frauke Liers, 100 spins, 4950 integer-weighted edges, from the statistical-physics section of the library (lines 737 to 747), with the proven optimal cut value.
46
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
47
  Origin: benchmark instance generated with the rudy graph generator (or taken from a physics application) and published with its optimum by the library authors for testing exact max-cut and BQP solvers
48
- Reference result: published optimal objective value
49
- Reference source: biqmaclib.tex line 757: "ising2.5-100\_5555 & 2460049" in the max-cut tables (table caption line 801: Test runs on one-dimensional Ising chain instances); lines 33 to 37 say the tables list "the optimal solution values" and give lower/upper bounds where the optimum is not known, which the tables do in a separate "lower bound" column (lines 140 and 285); this table has only the "solution" column; the problem is stated at lines 475 to 481 as max x^T L x over x in {-1,1}^n with L the Laplace matrix. Values published by the library authors, proved optimal with the Biq Mac solver.
50
  Check: Exact: the agent returns a side (0 or 1) for every vertex; the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file and must equal 2460049; since 2460049 is a proved optimum, a larger value signals a reading error
51
  Companions (not included): biqmaclib.tex
52
  Dependencies: not packed
53
  - `library/biq/be/be120.3.1.sparse`: Biq Mac Library unconstrained binary quadratic instance be120.3.1 (Billionnet-Elloumi, n=120, density 0.3), with the proven optimum.
54
  Format: sparse matrix: first line "n m", then m lines "i j q" (1-based, i <= j, upper triangle and diagonal of the symmetric Q); one file is one case
55
  Origin: benchmark instance from the Billionnet-Elloumi generator published with its optimum by the library authors for testing exact BQP solvers
56
- Reference result: published optimal objective value
57
- Reference source: biqmaclib.tex line 338: "be120.3.1 & -13067" in the table of instances from Billionnet and Elloumi (caption line 398); the problem is stated at lines 62 to 66 as min y^T Q y over y in {0,1}^n with Q symmetric. The file lists the upper triangle, so each off-diagonal entry counts twice in y^T Q y.
58
  Check: Exact: the agent returns y in {0,1}^120; y^T Q y is recomputed from the file with Q symmetric (diagonal entries once, each listed off-diagonal entry twice) and must equal -13067; a smaller value signals a reading error since -13067 is a proved optimum
59
  Companions (not included): biqmaclib.tex
60
  Dependencies: not packed
61
  - `library/mac/rudy/pw01_100.0`: Biq Mac Library max-cut instance pw01_100.0: random graph G(100, 0.1) with integer edge weights uniform in [1,10], generated with rudy (command at biqmaclib.tex line 1208), with the proven optimal cut value.
62
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
63
  Origin: benchmark instance published with its optimum by the library authors for testing exact max-cut and BQP solvers
64
- Reference result: published optimal objective value
65
- Reference source: biqmaclib.tex line 694: "pw01\_100.0 & 2019 \\"; table caption line 733: "$G_{[-10,10]}$ (w) and $G_{[1,10]}$ (pw)"; lines 34 to 37 say the tables list "the optimal solution values" and give lower/upper bounds where the optimum is not known; this table has only the solution column. Values published by the library authors, proved optimal with the Biq Mac solver.
66
  Check: the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file; it passes when the weight is not less than 2019; since 2019 is the published optimum, a pass means the weight equals it and a larger value signals a reading error
67
  Companions (not included): biqmaclib.tex
68
  Dependencies: not packed
69
  - `library/mac/ising/t2g10_5555`: Biq Mac Library max-cut instance t2g10_5555: 10x10 two-dimensional toroidal grid with Gaussian distributed integer weights (Ising spin glass instance from Frauke Liers), with the proven optimal cut value.
70
  Format: edge list: first line "n m", then m lines "i j w" (1-based vertices, signed integer weight); one file is one case
71
  Origin: benchmark instance published with its optimum by the library authors for testing exact max-cut and BQP solvers
72
- Reference result: published optimal objective value
73
- Reference source: biqmaclib.tex line 818: "t2g10\_5555 & 6049461 \\"; table caption line 855: "Torus graphs with Gaussian distributed weights from Frauke Liers"; section text at lines 740 to 745 says the dimensions and optimal values are given in this table; lines 34 to 37 say the tables list "the optimal solution values" and give lower/upper bounds where the optimum is not known; this table has only the solution column. Values published by the library authors, proved optimal with the Biq Mac solver.
74
  Check: the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file; it passes when the weight is not less than 6049461; since 6049461 is the published optimum, a pass means the weight equals it and a larger value signals a reading error
75
  Companions (not included): biqmaclib.tex
76
  Dependencies: not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 7 of 7 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Tools are named in the check written per file in the notes.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `library/mac/rudy/g05_60.0`: Biq Mac Library max-cut instance g05_60.0: unweighted G(n,1/2) random graph, 60 vertices, 885 edges, generated with rudy (command at biqmaclib.tex line 1098), with the proven optimal cut value.
20
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
21
  Origin: benchmark instance generated with the rudy graph generator (or taken from a physics application) and published with its optimum by the library authors for testing exact max-cut and BQP solvers
 
 
22
  Check: Exact: the agent returns a side (0 or 1) for every vertex; the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file and must equal 536; since 536 is a proved optimum, a larger value signals a reading error
23
  Companions (not included): biqmaclib.tex
24
  Dependencies: not packed
25
  - `library/mac/rudy/pm1s_80.0`: Biq Mac Library max-cut instance pm1s_80.0: sparse graph with edge weights in {-1,0,1}, density 10%, 80 vertices, 316 edges, generated with rudy (command at biqmaclib.tex line 1131), with the proven optimal cut value.
26
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
27
  Origin: benchmark instance generated with the rudy graph generator (or taken from a physics application) and published with its optimum by the library authors for testing exact max-cut and BQP solvers
 
 
28
  Check: Exact: the agent returns a side (0 or 1) for every vertex; the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file and must equal 79; since 79 is a proved optimum, a larger value signals a reading error
29
  Companions (not included): biqmaclib.tex
30
  Dependencies: not packed
31
  - `library/mac/rudy/w01_100.0`: Biq Mac Library max-cut instance w01_100.0: graph with integer edge weights in [-10,10], density 10%, 100 vertices, 495 edges, generated with rudy (command at biqmaclib.tex line 1175), with the proven optimal cut value.
32
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
33
  Origin: benchmark instance generated with the rudy graph generator (or taken from a physics application) and published with its optimum by the library authors for testing exact max-cut and BQP solvers
 
 
34
  Check: Exact: the agent returns a side (0 or 1) for every vertex; the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file and must equal 651; since 651 is a proved optimum, a larger value signals a reading error
35
  Companions (not included): biqmaclib.tex
36
  Dependencies: not packed
37
  - `library/mac/ising/ising2.5-100_5555`: Biq Mac Library max-cut instance ising2.5-100_5555: one-dimensional Ising chain instance from Frauke Liers, 100 spins, 4950 integer-weighted edges, from the statistical-physics section of the library (lines 737 to 747), with the proven optimal cut value.
38
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
39
  Origin: benchmark instance generated with the rudy graph generator (or taken from a physics application) and published with its optimum by the library authors for testing exact max-cut and BQP solvers
 
 
40
  Check: Exact: the agent returns a side (0 or 1) for every vertex; the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file and must equal 2460049; since 2460049 is a proved optimum, a larger value signals a reading error
41
  Companions (not included): biqmaclib.tex
42
  Dependencies: not packed
43
  - `library/biq/be/be120.3.1.sparse`: Biq Mac Library unconstrained binary quadratic instance be120.3.1 (Billionnet-Elloumi, n=120, density 0.3), with the proven optimum.
44
  Format: sparse matrix: first line "n m", then m lines "i j q" (1-based, i <= j, upper triangle and diagonal of the symmetric Q); one file is one case
45
  Origin: benchmark instance from the Billionnet-Elloumi generator published with its optimum by the library authors for testing exact BQP solvers
 
 
46
  Check: Exact: the agent returns y in {0,1}^120; y^T Q y is recomputed from the file with Q symmetric (diagonal entries once, each listed off-diagonal entry twice) and must equal -13067; a smaller value signals a reading error since -13067 is a proved optimum
47
  Companions (not included): biqmaclib.tex
48
  Dependencies: not packed
49
  - `library/mac/rudy/pw01_100.0`: Biq Mac Library max-cut instance pw01_100.0: random graph G(100, 0.1) with integer edge weights uniform in [1,10], generated with rudy (command at biqmaclib.tex line 1208), with the proven optimal cut value.
50
  Format: rudy output: first line "n m", then m lines "i j w" (1-based vertices, integer weight); one file is one case
51
  Origin: benchmark instance published with its optimum by the library authors for testing exact max-cut and BQP solvers
 
 
52
  Check: the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file; it passes when the weight is not less than 2019; since 2019 is the published optimum, a pass means the weight equals it and a larger value signals a reading error
53
  Companions (not included): biqmaclib.tex
54
  Dependencies: not packed
55
  - `library/mac/ising/t2g10_5555`: Biq Mac Library max-cut instance t2g10_5555: 10x10 two-dimensional toroidal grid with Gaussian distributed integer weights (Ising spin glass instance from Frauke Liers), with the proven optimal cut value.
56
  Format: edge list: first line "n m", then m lines "i j w" (1-based vertices, signed integer weight); one file is one case
57
  Origin: benchmark instance published with its optimum by the library authors for testing exact max-cut and BQP solvers
 
 
58
  Check: the cut weight (sum of w over edges whose ends are on different sides) is recomputed from the file; it passes when the weight is not less than 6049461; since 6049461 is the published optimum, a pass means the weight equals it and a larger value signals a reading error
59
  Companions (not included): biqmaclib.tex
60
  Dependencies: not packed
campaign_files/2026-10-g-solver-instances-2/borroot-akari/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each unit is self-contained. 5 of 5 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Any valid bulb placement. Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
14
 
15
  **Licence (SPDX):** NOASSERTION AND GPL-3.0-only
@@ -21,40 +19,30 @@ Units:
21
  - `misc/internet/` (3 files): Akari (Light Up) 7x7 puzzles, easy, normal and hard (2,997 puzzles).
22
  Format: plain text, LF; three files (7x7_easy, 7x7_normal, 7x7_hard), one puzzle per line in the puzzle-light-up.com task encoding (7x7 grid); delivered whole as the upstream files plus LICENSE
23
  Origin: puzzles of the public site www.puzzle-light-up.com, saved by the repository author for his thesis experiments
24
- Reference result: any valid bulb placement
25
- Reference source: No solutions ship with the files. more in REPORT.md. judging does not need it.
26
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
27
  Companions (not included): LICENSE
28
  Dependencies: not packed; not packed
29
  - `misc/internet/` (3 files): Akari (Light Up) 10x10 puzzles, easy, normal and hard (2,997 puzzles).
30
  Format: plain text, LF; three files (10x10_easy, 10x10_normal, 10x10_hard), one puzzle per line in the puzzle-light-up.com task encoding (10x10 grid); delivered whole as the upstream files plus LICENSE
31
  Origin: puzzles of the public site www.puzzle-light-up.com, saved by the repository author for his thesis experiments
32
- Reference result: any valid bulb placement
33
- Reference source: No solutions ship with the files. more in REPORT.md. judging does not need it.
34
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
35
  Companions (not included): LICENSE
36
  Dependencies: not packed; not packed
37
  - `misc/internet/` (3 files): Akari (Light Up) 14x14 puzzles, easy, normal and hard (2,997 puzzles).
38
  Format: plain text, LF; three files (14x14_easy, 14x14_normal, 14x14_hard), one puzzle per line in the puzzle-light-up.com task encoding (14x14 grid); delivered whole as the upstream files plus LICENSE
39
  Origin: puzzles of the public site www.puzzle-light-up.com, saved by the repository author for his thesis experiments
40
- Reference result: any valid bulb placement
41
- Reference source: No solutions ship with the files. more in REPORT.md. judging does not need it.
42
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
43
  Companions (not included): LICENSE
44
  Dependencies: not packed; not packed
45
  - `misc/internet/` (3 files): Akari (Light Up) 25x25 puzzles, easy, normal and hard (2,998 puzzles).
46
  Format: plain text, LF; three files (25x25_easy, 25x25_normal, 25x25_hard), one puzzle per line in the puzzle-light-up.com task encoding (25x25 grid); delivered whole as the upstream files plus LICENSE
47
  Origin: puzzles of the public site www.puzzle-light-up.com, saved by the repository author for his thesis experiments
48
- Reference result: any valid bulb placement
49
- Reference source: No solutions ship with the files. more in REPORT.md. judging does not need it.
50
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
51
  Companions (not included): LICENSE
52
  Dependencies: not packed; not packed
53
  - `misc/created/` (2 files): Akari (Light Up) 30x30 and 40x40 puzzles made by the repository author (39 puzzles).
54
  Format: plain text, LF; two files (30x30: 15 puzzles, 40x40: 24 puzzles), one puzzle per line in the same task encoding; delivered whole as the upstream files plus LICENSE
55
  Origin: larger puzzles the repository author made for his solver experiments (directory misc/created)
56
- Reference result: any valid bulb placement
57
- Reference source: No solutions ship with the files. more in REPORT.md. judging does not need it.
58
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
59
  Companions (not included): LICENSE
60
  Dependencies: not packed; not packed
 
8
 
9
  **Self-contained:** Each unit is self-contained. 5 of 5 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
12
 
13
  **Licence (SPDX):** NOASSERTION AND GPL-3.0-only
 
19
  - `misc/internet/` (3 files): Akari (Light Up) 7x7 puzzles, easy, normal and hard (2,997 puzzles).
20
  Format: plain text, LF; three files (7x7_easy, 7x7_normal, 7x7_hard), one puzzle per line in the puzzle-light-up.com task encoding (7x7 grid); delivered whole as the upstream files plus LICENSE
21
  Origin: puzzles of the public site www.puzzle-light-up.com, saved by the repository author for his thesis experiments
 
 
22
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
23
  Companions (not included): LICENSE
24
  Dependencies: not packed; not packed
25
  - `misc/internet/` (3 files): Akari (Light Up) 10x10 puzzles, easy, normal and hard (2,997 puzzles).
26
  Format: plain text, LF; three files (10x10_easy, 10x10_normal, 10x10_hard), one puzzle per line in the puzzle-light-up.com task encoding (10x10 grid); delivered whole as the upstream files plus LICENSE
27
  Origin: puzzles of the public site www.puzzle-light-up.com, saved by the repository author for his thesis experiments
 
 
28
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
29
  Companions (not included): LICENSE
30
  Dependencies: not packed; not packed
31
  - `misc/internet/` (3 files): Akari (Light Up) 14x14 puzzles, easy, normal and hard (2,997 puzzles).
32
  Format: plain text, LF; three files (14x14_easy, 14x14_normal, 14x14_hard), one puzzle per line in the puzzle-light-up.com task encoding (14x14 grid); delivered whole as the upstream files plus LICENSE
33
  Origin: puzzles of the public site www.puzzle-light-up.com, saved by the repository author for his thesis experiments
 
 
34
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
35
  Companions (not included): LICENSE
36
  Dependencies: not packed; not packed
37
  - `misc/internet/` (3 files): Akari (Light Up) 25x25 puzzles, easy, normal and hard (2,998 puzzles).
38
  Format: plain text, LF; three files (25x25_easy, 25x25_normal, 25x25_hard), one puzzle per line in the puzzle-light-up.com task encoding (25x25 grid); delivered whole as the upstream files plus LICENSE
39
  Origin: puzzles of the public site www.puzzle-light-up.com, saved by the repository author for his thesis experiments
 
 
40
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
41
  Companions (not included): LICENSE
42
  Dependencies: not packed; not packed
43
  - `misc/created/` (2 files): Akari (Light Up) 30x30 and 40x40 puzzles made by the repository author (39 puzzles).
44
  Format: plain text, LF; two files (30x30: 15 puzzles, 40x40: 24 puzzles), one puzzle per line in the same task encoding; delivered whole as the upstream files plus LICENSE
45
  Origin: larger puzzles the repository author made for his solver experiments (directory misc/created)
 
 
46
  Check: Decode the puzzle line (a lowercase letter a to z stands for 1 to 26 empty cells in a row, B a wall without a number, 0 to 4 a wall with that number; row-major). Check the agent's bulb placement by the rules: bulbs only on empty cells, every empty cell lit along its row or column up to the next wall, no bulb lights another bulb, and every numbered wall has exactly that many bulbs on its four sides. Any placement that passes is correct (rule check, SPEC line 283), so uniqueness is not required.
47
  Companions (not included): LICENSE
48
  Dependencies: not packed; not packed
campaign_files/2026-10-g-solver-instances/cblib/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Optimal objective bracketed by certified primal and dual objectives. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Check: Accept the agent's objective when it lies in the interval between the published dual objective and primal objective, widened by the published certificate errors of the stat table.
14
 
15
  **Licence (SPDX):** LicenseRef-ZIB-CBLIB-data AND LicenseRef-ZIB-DTU-cblib-base
@@ -21,8 +19,6 @@ Files:
21
  - `download/all/chainsing-1000-1.cbf.gz`: CBLIB conic instance chainsing-1000-1 (second-order cone, CBLIB 2014 core set) with certified primal and dual objectives.
22
  Format: gzip of Conic Benchmark Format (.cbf) text; companion stat.set-cblib2014.csv (semicolon separated: PACK;NAME;PRIMAL SENSE;PRIMAL CERTIFICATE;ERROR;OBJECTIVE;DUAL CERTIFICATE;ERROR;OBJECTIVE;CLAIM)
23
  Origin: chained singular function problem from the CBLIB collection (contributed test problem family)
24
- Reference result: optimal objective bracketed by certified primal and dual objectives
25
- Reference source: stat.set-cblib2014.csv line 2: "chainsing;chainsing-1000-1;MIN;FEASIBILITY;2.052778132794635e-12;30.18015748754364;FEASIBILITY;5.385469847851709e-12;30.18015746559776;" (header line 1). Published by the library.
26
  Check: Accept the agent's objective when it lies in the interval between the published dual objective and primal objective, widened by the published certificate errors of the stat table.
27
  Companions (not included): cblib/instances/stat.set-cblib2014.csv
28
  Dependencies: not packed
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Accept the agent's objective when it lies in the interval between the published dual objective and primal objective, widened by the published certificate errors of the stat table.
12
 
13
  **Licence (SPDX):** LicenseRef-ZIB-CBLIB-data AND LicenseRef-ZIB-DTU-cblib-base
 
19
  - `download/all/chainsing-1000-1.cbf.gz`: CBLIB conic instance chainsing-1000-1 (second-order cone, CBLIB 2014 core set) with certified primal and dual objectives.
20
  Format: gzip of Conic Benchmark Format (.cbf) text; companion stat.set-cblib2014.csv (semicolon separated: PACK;NAME;PRIMAL SENSE;PRIMAL CERTIFICATE;ERROR;OBJECTIVE;DUAL CERTIFICATE;ERROR;OBJECTIVE;CLAIM)
21
  Origin: chained singular function problem from the CBLIB collection (contributed test problem family)
 
 
22
  Check: Accept the agent's objective when it lies in the interval between the published dual objective and primal objective, widened by the published certificate errors of the stat table.
23
  Companions (not included): cblib/instances/stat.set-cblib2014.csv
24
  Dependencies: not packed
campaign_files/2026-10-gt-browser-tests-2/allenporter-ical-recurring-single-change/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_values_in_test_code. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Python ical (IcsCalendarStream.calendar_from_ics, calendar.timeline) or icalendar plus recurring-ical-events; a script expands the series and writes the first occurrences (summary, recurrence id, start) to a CSV or JSON file
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `tests/examples/testdata/recurring_with_single_change.ics`: Expand a Google Calendar daily series with one renamed occurrence.
22
  Format: iCalendar (.ics) file; expected values in a pytest file
23
  Origin: allenporter/ical example test fixture (tool_fixture)
24
- Reference result: expected_values_in_test_code
25
- Reference source: tests/examples/test_ics_examples.py lines 36-58 (test_recurring_with_single_change): the first three timeline events are 'Initial Title' 20260201T100000, 'Edited Title' 20260202T100000 and 'Initial Title' 20260203T100000, each starting at 10:00 America/New_York
26
  Check: Final file state: the agent writes the first three occurrences of the series (summary, recurrence id, start with time zone) to a file; each of the three records must equal the expected summary, recurrence id string and start datetime exactly, in order. No screenshot
27
  Tools: Python ical (IcsCalendarStream.calendar_from_ics, calendar.timeline) or icalendar plus recurring-ical-events; a script expands the series and writes the first occurrences (summary, recurrence id, start) to a CSV or JSON file
28
  Companions (not included): tests/examples/test_ics_examples.py
 
8
 
9
  **Self-contained:** The file is self-contained. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Python ical (IcsCalendarStream.calendar_from_ics, calendar.timeline) or icalendar plus recurring-ical-events; a script expands the series and writes the first occurrences (summary, recurrence id, start) to a CSV or JSON file
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `tests/examples/testdata/recurring_with_single_change.ics`: Expand a Google Calendar daily series with one renamed occurrence.
20
  Format: iCalendar (.ics) file; expected values in a pytest file
21
  Origin: allenporter/ical example test fixture (tool_fixture)
 
 
22
  Check: Final file state: the agent writes the first three occurrences of the series (summary, recurrence id, start with time zone) to a file; each of the three records must equal the expected summary, recurrence id string and start datetime exactly, in order. No screenshot
23
  Tools: Python ical (IcsCalendarStream.calendar_from_ics, calendar.timeline) or icalendar plus recurring-ical-events; a script expands the series and writes the first occurrences (summary, recurrence id, start) to a CSV or JSON file
24
  Companions (not included): tests/examples/test_ics_examples.py
campaign_files/2026-10-gt-browser-tests-2/apache-poi-formula-eval-testdata/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_cell_values. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every formula cell into the EXPECTED VALUE row directly below it in a copy of the workbook (or into a CSV keyed by cell address)
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `test-data/spreadsheet/FormulaEvalTestData.xls`: Evaluate 1,276 formulas covering 17 operators and 144 worksheet functions in an Excel 97 workbook.
22
  Format: Excel 97-2003 workbook (.xls, BIFF8)
23
  Origin: Apache POI general formula evaluation test workbook authored in Excel (tool_fixture)
24
- Reference result: expected_cell_values
25
- Reference source: for every function block (4 rows each, function name in column B) the row below the FORMULA row is the EXPECTED VALUE row; from column D on, each formula cell has its expected value in the same column of that row; poi/src/test/java/org/apache/poi/ss/formula/eval/TestFormulasFromSpreadsheet.java lines 136-155 and 160-215 check them
26
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each formula cell and compare it with the EXPECTED VALUE cell below it in the original under the parameters below; score the share of cases matched. No screenshot. Eight cached formula results differ from the expected row (for example ADDRESS with a quoted sheet name, AND, MODE and SUMPRODUCT error cases), so the hidden cache must not be used as the key; the expected row is what the test checks.
27
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every formula cell into the EXPECTED VALUE row directly below it in a copy of the workbook (or into a CSV keyed by cell address)
28
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
 
8
 
9
  **Self-contained:** The file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every formula cell into the EXPECTED VALUE row directly below it in a copy of the workbook (or into a CSV keyed by cell address)
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `test-data/spreadsheet/FormulaEvalTestData.xls`: Evaluate 1,276 formulas covering 17 operators and 144 worksheet functions in an Excel 97 workbook.
20
  Format: Excel 97-2003 workbook (.xls, BIFF8)
21
  Origin: Apache POI general formula evaluation test workbook authored in Excel (tool_fixture)
 
 
22
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each formula cell and compare it with the EXPECTED VALUE cell below it in the original under the parameters below; score the share of cases matched. No screenshot. Eight cached formula results differ from the expected row (for example ADDRESS with a quoted sheet name, AND, MODE and SUMPRODUCT error cases), so the hidden cache must not be used as the key; the expected row is what the test checks.
23
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every formula cell into the EXPECTED VALUE row directly below it in a copy of the workbook (or into a CSV keyed by cell address)
24
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
campaign_files/2026-10-gt-browser-tests-2/apache-poi-function-testdata/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_cell_values. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,168 +19,126 @@ Files:
21
  - `test-data/spreadsheet/LookupFunctionsTestCaseData.xls`: Evaluate 207 VLOOKUP, HLOOKUP and LOOKUP formulas in an Excel 97 workbook.
22
  Format: Excel 97-2003 workbook (.xls, BIFF8)
23
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
24
- Reference result: expected_cell_values
25
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestLookupFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
26
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. three cached formula results on sheet VL-oddArgs (rows 13, 25 and 26) differ from column C (Excel cache from the authoring version); column C is the answer the test checks, the cache is hidden anyway
27
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
28
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
29
  - `test-data/spreadsheet/IndexFunctionTestCaseData.xls`: Evaluate 98 INDEX formulas in an Excel 97 workbook.
30
  Format: Excel 97-2003 workbook (.xls, BIFF8)
31
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
32
- Reference result: expected_cell_values
33
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestIndexFunctionFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
34
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
35
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
36
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
37
  - `test-data/spreadsheet/IndirectFunctionTestCaseData.xls`: Evaluate 11 INDIRECT formulas (A1 and R1C1 references, other sheets) in an Excel 97 workbook.
38
  Format: Excel 97-2003 workbook (.xls, BIFF8)
39
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
40
- Reference result: expected_cell_values
41
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestIndirectFunctionFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
42
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
43
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
44
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
45
  - `test-data/spreadsheet/Trend.xls`: Evaluate 41 TREND regression formulas in an Excel 97 workbook.
46
  Format: Excel 97-2003 workbook (.xls, BIFF8)
47
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
48
- Reference result: expected_cell_values
49
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestTrendFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
50
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
51
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
52
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
53
  - `test-data/spreadsheet/DStar.xls`: Evaluate 16 DMIN, DMAX and DSUM database formulas in an Excel 97 workbook.
54
  Format: Excel 97-2003 workbook (.xls, BIFF8)
55
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
56
- Reference result: expected_cell_values
57
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestDStarFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
58
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. cell style names are Polish (from the authoring locale); all cell content is English
59
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
60
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
61
  - `test-data/spreadsheet/DateTimeToNumberTestCases.xls`: Evaluate 15 DATEVALUE and TIMEVALUE style formulas that parse date and time strings in an Excel 97 workbook.
62
  Format: Excel 97-2003 workbook (.xls, BIFF8)
63
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
64
- Reference result: expected_cell_values
65
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestDateTimeToNumberFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
66
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. the strings are US-format dates and times; the task statement names the en-US locale, and column C (computed in that locale) is the answer
67
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
68
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
69
  - `test-data/spreadsheet/RomanFunctionTestCaseData.xls`: Evaluate 35 ROMAN formulas (all five conciseness forms) in an Excel 97 workbook.
70
  Format: Excel 97-2003 workbook (.xls, BIFF8)
71
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
72
- Reference result: expected_cell_values
73
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestRomanFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
74
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
75
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
76
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
77
  - `test-data/spreadsheet/ComplexFunctionTestCaseData.xls`: Evaluate 12 COMPLEX formulas (real and imaginary parts, i or j suffix, error cases) in an Excel 97 workbook.
78
  Format: Excel 97-2003 workbook (.xls, BIFF8)
79
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
80
- Reference result: expected_cell_values
81
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestComplexFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
82
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
83
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
84
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
85
  - `test-data/spreadsheet/MatchFunctionTestCaseData.xls`: Evaluate 14 MATCH formulas (exact, ascending and descending match types) in an Excel 97 workbook.
86
  Format: Excel 97-2003 workbook (.xls, BIFF8)
87
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
88
- Reference result: expected_cell_values
89
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestMatchFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
90
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. the Read Me sheet text in A1 was copied from the lookup workbook and still names lookup functions; the test class name in A3 is correct
91
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
92
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
93
  - `test-data/spreadsheet/WeekNumFunctionTestCaseData.xls`: Evaluate 7 WEEKNUM formulas with different return types in an Excel 97 workbook.
94
  Format: Excel 97-2003 workbook (.xls, BIFF8)
95
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
96
- Reference result: expected_cell_values
97
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestWeekNumFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
98
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
99
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
100
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
101
  - `test-data/spreadsheet/57003-FixedFunctionTestCaseData.xls`: Evaluate 20 FIXED number-formatting formulas in an Excel 97 workbook.
102
  Format: Excel 97-2003 workbook (.xls, BIFF8)
103
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
104
- Reference result: expected_cell_values
105
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestFixedFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
106
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. contributed with POI bug 57003; the cached formula results were computed in a German locale (for example 1.234,57) while column C holds the en-US strings (for example 1,234.57); the task statement names en-US and the judge reads column C
107
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
108
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
109
  - `test-data/spreadsheet/IfFunctionTestCaseData.xls`: Evaluate 45 IF formulas (nested, missing arguments, error propagation) in an Excel 97 workbook.
110
  Format: Excel 97-2003 workbook (.xls, BIFF8)
111
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
112
- Reference result: expected_cell_values
113
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestIFFunctionFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
114
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. the Read Me sheet text in A1 was copied from another function workbook; the test class name in A3 is correct
115
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
116
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
117
  - `test-data/spreadsheet/ReptFunctionTestCaseData.xls`: Evaluate 5 REPT formulas in an Excel 97 workbook.
118
  Format: Excel 97-2003 workbook (.xls, BIFF8)
119
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
120
- Reference result: expected_cell_values
121
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestReptFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
122
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
123
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
124
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
125
  - `test-data/spreadsheet/DGet.xls`: Evaluate 87 DGET database formulas in an Excel 97 workbook.
126
  Format: Excel 97-2003 workbook (.xls, BIFF8)
127
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
128
- Reference result: expected_cell_values
129
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestDGetFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
130
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. two cached formula results (Excel rows 18 and 52) differ from column C (cache from the authoring Excel version: #VALUE! where column C holds Scrooge and #NUM!); column C is the answer the test checks, the cache is hidden anyway; a few case comments in column D are German (in der Zielzelle)
131
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
132
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
133
  - `test-data/spreadsheet/LogicalFunctionsTestCaseData.xls`: Evaluate 15 formulas built on the IS functions (ISERROR, ISBLANK and related) in an Excel 97 workbook.
134
  Format: Excel 97-2003 workbook (.xls, BIFF8)
135
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
136
- Reference result: expected_cell_values
137
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestLogicalFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
138
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
139
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
140
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
141
  - `test-data/spreadsheet/BooleanFunctionsTestCaseData.xls`: Evaluate 28 AND, OR and NOT formulas (array arguments, counts of non-blank values) in an Excel 97 workbook.
142
  Format: Excel 97-2003 workbook (.xls, BIFF8)
143
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
144
- Reference result: expected_cell_values
145
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestBooleanFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
146
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
147
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
148
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
149
  - `test-data/spreadsheet/DeltaFunctionTestCaseData.xls`: Evaluate 7 DELTA formulas (equality test of two numbers, error cases) in an Excel 97 workbook.
150
  Format: Excel 97-2003 workbook (.xls, BIFF8)
151
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
152
- Reference result: expected_cell_values
153
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestDeltaFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
154
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
155
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
156
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
157
  - `test-data/spreadsheet/FactDoubleFunctionTestCaseData.xls`: Evaluate 8 FACTDOUBLE formulas (double factorial, nonnumeric and negative arguments) in an Excel 97 workbook.
158
  Format: Excel 97-2003 workbook (.xls, BIFF8)
159
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
160
- Reference result: expected_cell_values
161
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestFactDoubleFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
162
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
163
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
164
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
165
  - `test-data/spreadsheet/QuotientFunctionTestCaseData.xls`: Evaluate 8 QUOTIENT formulas (integer part of a division, nonnumeric arguments, division by zero) in an Excel 97 workbook.
166
  Format: Excel 97-2003 workbook (.xls, BIFF8)
167
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
168
- Reference result: expected_cell_values
169
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestQuotientFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
170
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
171
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
172
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
173
  - `test-data/spreadsheet/IfNaTestCaseData.xls`: Evaluate 8 IFNA formulas (numeric, string and blank values, #N/A with numeric, string or error defaults) in an Excel 97 workbook.
174
  Format: Excel 97-2003 workbook (.xls, BIFF8)
175
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
176
- Reference result: expected_cell_values
177
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestIfnaFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
178
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
179
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
180
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
181
  - `test-data/spreadsheet/TwoOperandNumericFunctionTestCaseData.xls`: Evaluate 34 two-operand numeric formulas (the Read Me sheet names +, -, * and POWER; the commented cases apply +, -, * and / to array arguments containing error values) in an Excel 97 workbook.
182
  Format: Excel 97-2003 workbook (.xls, BIFF8)
183
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
184
- Reference result: expected_cell_values
185
- Reference source: column C of every case row on each test sheet (rows from 5 to the <end> marker; <skip> rows excluded) is the expected value that TestTwoOperandNumericFunctionsFromSpreadsheet checks through BaseTestFunctionsFromSpreadsheet.java
186
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
187
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
188
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
 
8
 
9
  **Self-contained:** Each file is self-contained. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `test-data/spreadsheet/LookupFunctionsTestCaseData.xls`: Evaluate 207 VLOOKUP, HLOOKUP and LOOKUP formulas in an Excel 97 workbook.
20
  Format: Excel 97-2003 workbook (.xls, BIFF8)
21
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
22
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. three cached formula results on sheet VL-oddArgs (rows 13, 25 and 26) differ from column C (Excel cache from the authoring version); column C is the answer the test checks, the cache is hidden anyway
23
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
24
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
25
  - `test-data/spreadsheet/IndexFunctionTestCaseData.xls`: Evaluate 98 INDEX formulas in an Excel 97 workbook.
26
  Format: Excel 97-2003 workbook (.xls, BIFF8)
27
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
28
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
29
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
30
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
31
  - `test-data/spreadsheet/IndirectFunctionTestCaseData.xls`: Evaluate 11 INDIRECT formulas (A1 and R1C1 references, other sheets) in an Excel 97 workbook.
32
  Format: Excel 97-2003 workbook (.xls, BIFF8)
33
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
34
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
35
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
36
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
37
  - `test-data/spreadsheet/Trend.xls`: Evaluate 41 TREND regression formulas in an Excel 97 workbook.
38
  Format: Excel 97-2003 workbook (.xls, BIFF8)
39
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
40
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
41
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
42
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
43
  - `test-data/spreadsheet/DStar.xls`: Evaluate 16 DMIN, DMAX and DSUM database formulas in an Excel 97 workbook.
44
  Format: Excel 97-2003 workbook (.xls, BIFF8)
45
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
46
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. cell style names are Polish (from the authoring locale); all cell content is English
47
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
48
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
49
  - `test-data/spreadsheet/DateTimeToNumberTestCases.xls`: Evaluate 15 DATEVALUE and TIMEVALUE style formulas that parse date and time strings in an Excel 97 workbook.
50
  Format: Excel 97-2003 workbook (.xls, BIFF8)
51
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
52
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. the strings are US-format dates and times; the task statement names the en-US locale, and column C (computed in that locale) is the answer
53
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
54
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
55
  - `test-data/spreadsheet/RomanFunctionTestCaseData.xls`: Evaluate 35 ROMAN formulas (all five conciseness forms) in an Excel 97 workbook.
56
  Format: Excel 97-2003 workbook (.xls, BIFF8)
57
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
58
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
59
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
60
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
61
  - `test-data/spreadsheet/ComplexFunctionTestCaseData.xls`: Evaluate 12 COMPLEX formulas (real and imaginary parts, i or j suffix, error cases) in an Excel 97 workbook.
62
  Format: Excel 97-2003 workbook (.xls, BIFF8)
63
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
64
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
65
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
66
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
67
  - `test-data/spreadsheet/MatchFunctionTestCaseData.xls`: Evaluate 14 MATCH formulas (exact, ascending and descending match types) in an Excel 97 workbook.
68
  Format: Excel 97-2003 workbook (.xls, BIFF8)
69
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
70
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. the Read Me sheet text in A1 was copied from the lookup workbook and still names lookup functions; the test class name in A3 is correct
71
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
72
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
73
  - `test-data/spreadsheet/WeekNumFunctionTestCaseData.xls`: Evaluate 7 WEEKNUM formulas with different return types in an Excel 97 workbook.
74
  Format: Excel 97-2003 workbook (.xls, BIFF8)
75
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
76
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
77
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
78
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
79
  - `test-data/spreadsheet/57003-FixedFunctionTestCaseData.xls`: Evaluate 20 FIXED number-formatting formulas in an Excel 97 workbook.
80
  Format: Excel 97-2003 workbook (.xls, BIFF8)
81
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
82
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. contributed with POI bug 57003; the cached formula results were computed in a German locale (for example 1.234,57) while column C holds the en-US strings (for example 1,234.57); the task statement names en-US and the judge reads column C
83
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
84
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
85
  - `test-data/spreadsheet/IfFunctionTestCaseData.xls`: Evaluate 45 IF formulas (nested, missing arguments, error propagation) in an Excel 97 workbook.
86
  Format: Excel 97-2003 workbook (.xls, BIFF8)
87
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
88
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. the Read Me sheet text in A1 was copied from another function workbook; the test class name in A3 is correct
89
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
90
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
91
  - `test-data/spreadsheet/ReptFunctionTestCaseData.xls`: Evaluate 5 REPT formulas in an Excel 97 workbook.
92
  Format: Excel 97-2003 workbook (.xls, BIFF8)
93
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
94
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
95
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
96
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
97
  - `test-data/spreadsheet/DGet.xls`: Evaluate 87 DGET database formulas in an Excel 97 workbook.
98
  Format: Excel 97-2003 workbook (.xls, BIFF8)
99
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
100
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot. two cached formula results (Excel rows 18 and 52) differ from column C (cache from the authoring Excel version: #VALUE! where column C holds Scrooge and #NUM!); column C is the answer the test checks, the cache is hidden anyway; a few case comments in column D are German (in der Zielzelle)
101
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
102
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
103
  - `test-data/spreadsheet/LogicalFunctionsTestCaseData.xls`: Evaluate 15 formulas built on the IS functions (ISERROR, ISBLANK and related) in an Excel 97 workbook.
104
  Format: Excel 97-2003 workbook (.xls, BIFF8)
105
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
106
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
107
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
108
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
109
  - `test-data/spreadsheet/BooleanFunctionsTestCaseData.xls`: Evaluate 28 AND, OR and NOT formulas (array arguments, counts of non-blank values) in an Excel 97 workbook.
110
  Format: Excel 97-2003 workbook (.xls, BIFF8)
111
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
112
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
113
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
114
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
115
  - `test-data/spreadsheet/DeltaFunctionTestCaseData.xls`: Evaluate 7 DELTA formulas (equality test of two numbers, error cases) in an Excel 97 workbook.
116
  Format: Excel 97-2003 workbook (.xls, BIFF8)
117
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
118
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
119
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
120
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
121
  - `test-data/spreadsheet/FactDoubleFunctionTestCaseData.xls`: Evaluate 8 FACTDOUBLE formulas (double factorial, nonnumeric and negative arguments) in an Excel 97 workbook.
122
  Format: Excel 97-2003 workbook (.xls, BIFF8)
123
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
124
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
125
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
126
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
127
  - `test-data/spreadsheet/QuotientFunctionTestCaseData.xls`: Evaluate 8 QUOTIENT formulas (integer part of a division, nonnumeric arguments, division by zero) in an Excel 97 workbook.
128
  Format: Excel 97-2003 workbook (.xls, BIFF8)
129
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
130
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
131
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
132
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
133
  - `test-data/spreadsheet/IfNaTestCaseData.xls`: Evaluate 8 IFNA formulas (numeric, string and blank values, #N/A with numeric, string or error defaults) in an Excel 97 workbook.
134
  Format: Excel 97-2003 workbook (.xls, BIFF8)
135
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
136
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
137
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
138
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
139
  - `test-data/spreadsheet/TwoOperandNumericFunctionTestCaseData.xls`: Evaluate 34 two-operand numeric formulas (the Read Me sheet names +, -, * and POWER; the commented cases apply +, -, * and / to array arguments containing error values) in an Excel 97 workbook.
140
  Format: Excel 97-2003 workbook (.xls, BIFF8)
141
  Origin: Apache POI formula function test workbook authored in Excel (tool_fixture)
 
 
142
  Check: Final file state: open the agent output workbook (or CSV), take the value recorded for each case row and compare it with column C of the original under the parameters below; score the share of cases matched. No screenshot.
143
  Tools: soffice --headless --convert-to xlsx (recalculates on load) then openpyxl reads the values, or Apache POI HSSFFormulaEvaluator; the agent writes the evaluated value of every column B formula into column C of a copy of the workbook (or into a CSV keyed by sheet and row)
144
  Dependencies: org.apache.poi:poi at commit ae62bb5116b9aee19ebd5834e3a82066132c9f7f
campaign_files/2026-10-gt-browser-tests-2/biweekly-issues/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_values_in_test_code. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** biweekly (Java) VEvent.getDateIterator, or Python icalendar plus recurring-ical-events or dateutil.rrule; the script writes the occurrence start times to a file; biweekly (Java) TimezoneInfo plus VEvent.getDateIterator, or Python icalendar plus recurring-ical-events after mapping the TZID /America/New_York; the script writes the occurrence start times to a file
14
 
15
  **Licence (SPDX):** BSD-2-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `src/test/resources/biweekly/issues/issue101.ics`: Expand a daily event whose UNTIL in UTC falls between two local occurrences.
22
  Format: iCalendar (.ics) with VTIMEZONE; expected values in Java test code
23
  Origin: biweekly regression test fixture for issue 101 (tool_fixture)
24
- Reference result: expected_values_in_test_code
25
- Reference source: src/test/java/biweekly/issues/Issue101.java lines 43-50: six start times 2020-10-04T17:00:00 to 2020-10-09T17:00:00 local time, default time zone CET (line 27)
26
  Check: Final file state: the agent writes one occurrence start per line as local wall time in CET; the list must equal the six times in Issue101.java lines 44-49, in order, with nothing extra. No screenshot
27
  Tools: biweekly (Java) VEvent.getDateIterator, or Python icalendar plus recurring-ical-events or dateutil.rrule; the script writes the occurrence start times to a file
28
  Companions (not included): src/test/java/biweekly/issues/Issue101.java
@@ -30,8 +26,6 @@ Files:
30
  - `src/test/resources/biweekly/issues/pr63.ics`: Expand a Google Calendar daily event whose TZID has a leading slash.
31
  Format: iCalendar (.ics) without VTIMEZONE; expected values in Java test code
32
  Origin: biweekly regression test fixture for pull request 63 (tool_fixture)
33
- Reference result: expected_values_in_test_code
34
- Reference source: src/test/java/biweekly/issues/PR63.java lines 89-95 (pr63_original): five start times 2016-12-19T09:00:00 to 2016-12-23T09:00:00 in America/New_York
35
  Check: Final file state: the agent writes one occurrence start per line as wall time in America/New_York; the list must equal the five times in PR63.java lines 90-94, in order. No screenshot
36
  Tools: biweekly (Java) TimezoneInfo plus VEvent.getDateIterator, or Python icalendar plus recurring-ical-events after mapping the TZID /America/New_York; the script writes the occurrence start times to a file
37
  Companions (not included): src/test/java/biweekly/issues/PR63.java
 
8
 
9
  **Self-contained:** Each file is self-contained. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** biweekly (Java) VEvent.getDateIterator, or Python icalendar plus recurring-ical-events or dateutil.rrule; the script writes the occurrence start times to a file; biweekly (Java) TimezoneInfo plus VEvent.getDateIterator, or Python icalendar plus recurring-ical-events after mapping the TZID /America/New_York; the script writes the occurrence start times to a file
12
 
13
  **Licence (SPDX):** BSD-2-Clause
 
19
  - `src/test/resources/biweekly/issues/issue101.ics`: Expand a daily event whose UNTIL in UTC falls between two local occurrences.
20
  Format: iCalendar (.ics) with VTIMEZONE; expected values in Java test code
21
  Origin: biweekly regression test fixture for issue 101 (tool_fixture)
 
 
22
  Check: Final file state: the agent writes one occurrence start per line as local wall time in CET; the list must equal the six times in Issue101.java lines 44-49, in order, with nothing extra. No screenshot
23
  Tools: biweekly (Java) VEvent.getDateIterator, or Python icalendar plus recurring-ical-events or dateutil.rrule; the script writes the occurrence start times to a file
24
  Companions (not included): src/test/java/biweekly/issues/Issue101.java
 
26
  - `src/test/resources/biweekly/issues/pr63.ics`: Expand a Google Calendar daily event whose TZID has a leading slash.
27
  Format: iCalendar (.ics) without VTIMEZONE; expected values in Java test code
28
  Origin: biweekly regression test fixture for pull request 63 (tool_fixture)
 
 
29
  Check: Final file state: the agent writes one occurrence start per line as wall time in America/New_York; the list must equal the five times in PR63.java lines 90-94, in order. No screenshot
30
  Tools: biweekly (Java) TimezoneInfo plus VEvent.getDateIterator, or Python icalendar plus recurring-ical-events after mapping the TZID /America/New_York; the script writes the occurrence start times to a file
31
  Companions (not included): src/test/java/biweekly/issues/PR63.java
campaign_files/2026-10-gt-browser-tests-2/browsertime-har-unit-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_json. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** a script that reads the HAR JSON and the browsertime result JSON and writes the enriched or merged HAR, or browsertime run headless with --har (its HAR export); the DevTools Network panel exports the same HAR through Export HAR and shows the page timings in its Timing view; a script that reads the HAR JSON (exported by Chrome headless over CDP, chrome-har or browsertime) and extracts the main document; jq over the .har file works for the redirect chain
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `test/unittests/harBuilderTest.js`: HAR building for a browser measurement run: add the creator block, merge two HAR files with page id renaming, and copy visual metrics, CPU long tasks, first paint, Largest Contentful Paint, long-task ranges and user timings into the HAR page and pageTimings.
22
  Format: ava test file (ES module JavaScript) that builds the input HAR and browsertime results inline and states the expected HAR fragments inline; each input can be written out as JSON for the task
23
  Origin: unit tests of the browsertime HAR builder (tool_fixture); the library source is not fetched, not bundled and not run
24
- Reference result: expected_json
25
- Reference source: test/unittests/harBuilderTest.js: creator name Browsertime and comment (lines 73 to 83); merged pages and entries with ids page_0, page_0-1, page_1 and version 1.2 (lines 117 to 142); the enriched page with _meta.connectivity native, _visualMetrics with VisualProgress as a timestamp-to-percent map, pageTimings._largestContentfulPaint 200 and _cpu (lines 157 to 182); the variant without visual metrics (lines 194 to 206); _longTasks [[100, 180], [250, 310]] (lines 230 to 233); _userTimings (line 262); empty and absent _longTasks (lines 289 to 311); _documentURL on the first entry (line 321)
26
  Check: Final file state: the agent receives the input HAR and the browsertime result written out as JSON and writes the resulting HAR; compare the named HAR fragments (log.creator, log.pages, log.entries, pages[0].pageTimings) with the inline expected objects by deep equality (the convention is ava t.deepEqual and t.is, harBuilderTest.js lines 117 and 157); the field naming (underscore custom fields) follows the HAR 1.2 rule that custom fields start with an underscore, which the expected objects show
27
  Tools: a script that reads the HAR JSON and the browsertime result JSON and writes the enriched or merged HAR, or browsertime run headless with --har (its HAR export); the DevTools Network panel exports the same HAR through Export HAR and shows the page timings in its Timing view
28
  Companions (not included): LICENSE
@@ -30,8 +26,6 @@ Files:
30
  - `test/unittests/mainDocumentTest.js`: Main document of a page from a HAR file: follow the redirect chain, count redirects, lowercase and join header names, drop sensitive headers (Set-Cookie, Authorization) and skip pages without a URL.
31
  Format: ava test file (ES module JavaScript) that builds a small HAR 1.2 object inline (pages and entries with requests, responses, statuses and headers) and states the expected main document inline
32
  Origin: unit tests of the browsertime HAR main document extraction (tool_fixture); the library source is not fetched, not bundled and not run
33
- Reference result: expected_json
34
- Reference source: test/unittests/mainDocumentTest.js: one document for en.m.wikipedia.org/wiki/Sweden with the deep mainDocument object at status 200 and 1 redirect (lines 59 to 71); lowercase header names (lines 75 to 76); set-cookie and authorization dropped (lines 86 to 87); 0 redirects (lines 94 to 95); joined repeated headers (line 105); pages without a URL skipped (line 118)
35
  Check: Final file state: the agent receives the HAR as a .har JSON file and writes the main document summary to a result JSON; compare field by field with the inline expected values (ava t.is, t.deepEqual, t.true, t.false)
36
  Tools: a script that reads the HAR JSON (exported by Chrome headless over CDP, chrome-har or browsertime) and extracts the main document; jq over the .har file works for the redirect chain
37
  Companions (not included): LICENSE
 
8
 
9
  **Self-contained:** Each file is self-contained. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** a script that reads the HAR JSON and the browsertime result JSON and writes the enriched or merged HAR, or browsertime run headless with --har (its HAR export); the DevTools Network panel exports the same HAR through Export HAR and shows the page timings in its Timing view; a script that reads the HAR JSON (exported by Chrome headless over CDP, chrome-har or browsertime) and extracts the main document; jq over the .har file works for the redirect chain
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `test/unittests/harBuilderTest.js`: HAR building for a browser measurement run: add the creator block, merge two HAR files with page id renaming, and copy visual metrics, CPU long tasks, first paint, Largest Contentful Paint, long-task ranges and user timings into the HAR page and pageTimings.
20
  Format: ava test file (ES module JavaScript) that builds the input HAR and browsertime results inline and states the expected HAR fragments inline; each input can be written out as JSON for the task
21
  Origin: unit tests of the browsertime HAR builder (tool_fixture); the library source is not fetched, not bundled and not run
 
 
22
  Check: Final file state: the agent receives the input HAR and the browsertime result written out as JSON and writes the resulting HAR; compare the named HAR fragments (log.creator, log.pages, log.entries, pages[0].pageTimings) with the inline expected objects by deep equality (the convention is ava t.deepEqual and t.is, harBuilderTest.js lines 117 and 157); the field naming (underscore custom fields) follows the HAR 1.2 rule that custom fields start with an underscore, which the expected objects show
23
  Tools: a script that reads the HAR JSON and the browsertime result JSON and writes the enriched or merged HAR, or browsertime run headless with --har (its HAR export); the DevTools Network panel exports the same HAR through Export HAR and shows the page timings in its Timing view
24
  Companions (not included): LICENSE
 
26
  - `test/unittests/mainDocumentTest.js`: Main document of a page from a HAR file: follow the redirect chain, count redirects, lowercase and join header names, drop sensitive headers (Set-Cookie, Authorization) and skip pages without a URL.
27
  Format: ava test file (ES module JavaScript) that builds a small HAR 1.2 object inline (pages and entries with requests, responses, statuses and headers) and states the expected main document inline
28
  Origin: unit tests of the browsertime HAR main document extraction (tool_fixture); the library source is not fetched, not bundled and not run
 
 
29
  Check: Final file state: the agent receives the HAR as a .har JSON file and writes the main document summary to a result JSON; compare field by field with the inline expected values (ava t.is, t.deepEqual, t.true, t.false)
30
  Tools: a script that reads the HAR JSON (exported by Chrome headless over CDP, chrome-har or browsertime) and extracts the main document; jq over the .har file works for the redirect chain
31
  Companions (not included): LICENSE
campaign_files/2026-10-gt-browser-tests-2/browsertime-trace-unit-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained. 10 of 10 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_numbers; expected_table; expected_json; expected_number. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** A script that reads the trace JSON (traceEvents) and computes the figures, or browsertime's trace analysis run headless on a saved trace (browsertime --chrome.timeline with the trace exported to JSON); the DevTools Performance panel shows the same events in its main-thread and frames tracks. Other files: see the notes.
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `test/unittests/traceAnalysisTest.js`: Total Blocking Time and its attribution from a synthetic Chrome trace: blocking time per URL and per kind of work, the window before Largest Contentful Paint, and the first-party versus third-party domain split.
22
  Format: ava test file (ES module JavaScript) that builds the Chrome trace events inline and states the expected results inline; the input trace can be written out as JSON for the task
23
  Origin: unit tests of the browsertime Chrome trace analysis (tool_fixture); the library source is not fetched, not bundled and not run
24
- Reference result: expected_numbers
25
- Reference source: test/unittests/traceAnalysisTest.js lines 124 to 190: t.is(result.totalBlockingTime, 270), t.is(result.tasks, 3), t.deepEqual on result.urls, result.kinds and result.domains, t.is(result.firstParty, 150), t.is(result.thirdParty, 160)
26
  Check: Final file state: the agent receives the trace written out as JSON and writes the figures to a result JSON; compare field by field with the inline expected values (ava t.is and t.deepEqual, exact)
27
  Tools: a script that reads the trace JSON (traceEvents) and computes the figures, or browsertime's trace analysis run headless on a saved trace (browsertime --chrome.timeline with the trace exported to JSON); the DevTools Performance panel shows the same events in its main-thread and frames tracks
28
  Companions (not included): LICENSE
@@ -30,8 +26,6 @@ Files:
30
  - `test/unittests/frameStabilityTest.js`: Frame stability from a synthetic Chrome trace of PipelineReporter events: presented, partial and dropped frames, effective fps and the longest gap between presented frames.
31
  Format: ava test file (ES module JavaScript) that builds the Chrome trace events inline and states the expected results inline; the input trace can be written out as JSON for the task
32
  Origin: unit tests of the browsertime Chrome trace analysis (tool_fixture); the library source is not fetched, not bundled and not run
33
- Reference result: expected_numbers
34
- Reference source: test/unittests/frameStabilityTest.js lines 85 to 184: t.is(result.presented, 3), t.is(result.effectiveFps, 20), t.deepEqual(result.longestGap, { duration: 83.3, startTime: 116.7 }) and the dropped, forked and ignored cases
35
  Check: Final file state: the agent receives the trace written out as JSON and writes the figures to a result JSON; compare field by field with the inline expected values (ava t.is and t.deepEqual, exact)
36
  Tools: a script that reads the trace JSON (traceEvents) and computes the figures, or browsertime's trace analysis run headless on a saved trace (browsertime --chrome.timeline with the trace exported to JSON); the DevTools Performance panel shows the same events in its main-thread and frames tracks
37
  Companions (not included): LICENSE
@@ -39,8 +33,6 @@ Files:
39
  - `test/unittests/styleInvalidationsTest.js`: Style and layout invalidations from a synthetic Chrome trace: recalculation and layout counts, reasons, triggering selectors and source scripts, split at first paint and at Largest Contentful Paint.
40
  Format: ava test file (ES module JavaScript) that builds the Chrome trace events inline and states the expected results inline; the input trace can be written out as JSON for the task
41
  Origin: unit tests of the browsertime Chrome trace analysis (tool_fixture); the library source is not fetched, not bundled and not run
42
- Reference result: expected_table
43
- Reference source: test/unittests/styleInvalidationsTest.js lines 18 to 208: t.is(result.styleRecalcs, 3), t.is(result.layoutInvalidations, 1), t.deepEqual on recalcReasons, layoutReasons, triggers and sources, and the after-first-paint counts
44
  Check: Final file state: the agent receives the trace written out as JSON and writes the figures to a result JSON; compare field by field with the inline expected values (ava t.is and t.deepEqual, exact)
45
  Tools: a script that reads the trace JSON (traceEvents) and computes the figures, or browsertime's trace analysis run headless on a saved trace (browsertime --chrome.timeline with the trace exported to JSON); the DevTools Performance panel shows the same events in its main-thread and frames tracks
46
  Companions (not included): LICENSE
@@ -48,8 +40,6 @@ Files:
48
  - `test/unittests/functionCostsTest.js`: Per-function self and total CPU time from the V8 sampling profiler chunks of a Chrome trace: aggregation across chunks, callee breakdown, bundle module resolution, recovery of URLs truncated at 1024 characters and a 1 ms noise limit.
49
  Format: ava test file (ES module JavaScript) that builds the Chrome trace events (ProfileChunk with cpuProfile nodes, samples and timeDeltas) inline and states the expected results inline; the input trace can be written out as JSON for the task
50
  Origin: unit tests of the browsertime function cost analysis of Chrome traces (tool_fixture); the library source is not fetched, not bundled and not run
51
- Reference result: expected_json
52
- Reference source: test/unittests/functionCostsTest.js: inner selfTime 4.7 and totalTime 4.7, outer selfTime 1.5 and totalTime 6.2 (lines 147 to 157); inner line 10 and column 5 (lines 161 to 163); outer.callees (lines 172 to 175 and 196); outer.module module.one (line 208); the recovered long URL (lines 229 to 230); only inner left under the noise limit (lines 240 to 243); [] for other processes and for no profiler data (lines 253 and 258)
53
  Check: Final file state: the agent receives the trace written out as JSON and writes the function costs to a result JSON; compare field by field with the inline expected values (the convention is ava t.deepEqual and t.is, functionCostsTest.js line 147); the conventions behind the numbers are stated in the test: each sample is weighted by its timeDeltas entry in microseconds and reported in ms, meta frames such as (garbage collector) and (root) are skipped (lines 142 to 144), positions are 1-based (line 160), negative deltas are skipped (lines 235 to 236)
54
  Tools: a script that reads the trace JSON (traceEvents with ProfileChunk samples) and writes the per-function costs, or browsertime run headless with --chrome.timeline on a saved trace; the DevTools Performance panel shows the same self and total times in its Bottom-Up and Call Tree views
55
  Companions (not included): LICENSE
@@ -57,8 +47,6 @@ Files:
57
  - `test/unittests/coverageTest.js`: Used JavaScript bytes from V8 block coverage (Chrome DevTools Profiler.takePreciseCoverage ranges): nested ranges where the innermost range wins, sibling functions, dead branches and zero-length scripts.
58
  Format: ava test file (ES module JavaScript) that states each V8 coverage object (functions with ranges of startOffset, endOffset and count) and the script length inline, with the expected used byte count inline
59
  Origin: unit tests of the browsertime JavaScript coverage computation (tool_fixture); the library source is not fetched, not bundled and not run
60
- Reference result: expected_number
61
- Reference source: test/unittests/coverageTest.js: expected used byte counts at lines 5, 17, 32, 51, 66, 84, 104, 108, 112 and 133 (for example 100 for a whole used function, 0 for an unused one, 80 when an inner count 0 range punches a hole, line 46)
62
  Check: Final file state: the agent receives each coverage object and script length as JSON and writes the used byte count per case to a result JSON; compare per case with the inline expected value by exact integer equality (ava t.is); the rule stated in the test is that the innermost range wins (lines 27 to 29)
63
  Tools: a script that reads the V8 precise coverage JSON (from Profiler.takePreciseCoverage over CDP, or a saved coverage file) and computes used bytes per script; browsertime run headless with its coverage option produces the same figure
64
  Companions (not included): LICENSE
@@ -66,8 +54,6 @@ Files:
66
  - `test/unittests/selectorStatsTest.js`: CSS selector cost from the SelectorStats events of a Chrome trace: totals across style recalculations, per-selector aggregation in ms and the list of wasted selectors (many match attempts, no match).
67
  Format: ava test file (ES module JavaScript) that builds the SelectorStats trace events inline (selector, elapsed, match_attempts, match_count, style_sheet_id) and states the expected results inline
68
  Origin: unit tests of the browsertime selector statistics from Chrome traces (tool_fixture); the library source is not fetched, not bundled and not run
69
- Reference result: expected_json
70
- Reference source: test/unittests/selectorStatsTest.js: totalElapsed 1.6, totalMatchAttempts 850, totalMatchCount 10 (lines 33 to 35); the first selector .slow > * with elapsed 1.5, matchAttempts 800, matchCount 0 (lines 36 to 42); wastedSelectors [.slow > *] (lines 43 to 46); further cases at lines 59 to 64 and 68
71
  Check: Final file state: the agent receives the trace events as JSON and writes the selector statistics to a result JSON; compare field by field with the inline expected values (ava t.is and t.deepEqual); elapsed is converted from microseconds to ms as the expected values show
72
  Tools: a script that reads the Chrome trace JSON (traceEvents) and writes the result JSON, or browsertime run headless with --chrome.timeline and its trace post-processing on a saved trace; the DevTools Performance panel shows the same figures interactively
73
  Companions (not included): LICENSE
@@ -75,8 +61,6 @@ Files:
75
  - `test/unittests/timerCostsTest.js`: JavaScript timer costs from a Chrome trace: TimerInstall and TimerFire events joined per timer, fire time, zero-timeout and recurring installs, grouped by script URL and by install site.
76
  Format: ava test file (ES module JavaScript) that builds the timer trace events inline and states the expected results inline
77
  Origin: unit tests of the browsertime timer cost analysis of Chrome traces (tool_fixture); the library source is not fetched, not bundled and not run
78
- Reference result: expected_json
79
- Reference source: test/unittests/timerCostsTest.js: per-URL row installs 1, fires 1, fireTime 25, timeoutZeroInstalls 1, recurringInstalls 1 (lines 94 to 103); per-site row with timeout 0, recurring true, line 1 and column 1 (lines 105 to 117); undefined when there is no timer data (line 124)
80
  Check: Final file state: the agent receives the trace events as JSON and writes the timer cost tables to a result JSON; compare field by field with the inline expected values (ava t.deepEqual); install sites are 1-based line and column as the expected rows show
81
  Tools: a script that reads the Chrome trace JSON (traceEvents) and writes the result JSON, or browsertime run headless with --chrome.timeline and its trace post-processing on a saved trace; the DevTools Performance panel shows the same figures interactively
82
  Companions (not included): LICENSE
@@ -84,8 +68,6 @@ Files:
84
  - `test/unittests/nonCompositedAnimationsTest.js`: Non-composited animations from Chrome trace Animation events: dedupe by animation id and decode the compositeFailed bitmask into named failure reasons, keeping unknown bits.
85
  Format: ava test file (ES module JavaScript) that builds the Animation trace events inline and states the expected reasons inline
86
  Origin: unit tests of the browsertime non-composited animation analysis of Chrome traces (tool_fixture); the library source is not fetched, not bundled and not run
87
- Reference result: expected_json
88
- Reference source: test/unittests/nonCompositedAnimationsTest.js: 2 animations after dedupe (line 24); reasons [unsupported CSS property, animation has no visible change] for bits 13 and 17 (lines 25 to 28); unsupportedProperties [top] (line 29); [unknown reason (bit 21)] (line 30); [] with no animations (line 37)
89
  Check: Final file state: the agent receives the trace events as JSON and writes the animation list to a result JSON; compare with the inline expected values (ava t.is and t.deepEqual); the bit to reason table of Blink must be given in the task text for the bits the test uses (13 and 17), other bits map to unknown reason (bit N) as line 30 shows
90
  Tools: a script that reads the Chrome trace JSON (traceEvents) and writes the result JSON, or browsertime run headless with --chrome.timeline and its trace post-processing on a saved trace; the DevTools Performance panel shows the same figures interactively
91
  Companions (not included): LICENSE
@@ -93,8 +75,6 @@ Files:
93
  - `test/unittests/extraMetricsTest.js`: Visual metrics milestones from a visual progress curve of a page load: VisualComplete85, 95 and 99 (including a curve that paints backward) and VisualReadiness.
94
  Format: ava test file (ES module JavaScript) that states each visual progress curve and first and last visual change inline with the expected milestones inline
95
  Origin: unit tests of the browsertime extra visual metrics (tool_fixture); the library source is not fetched, not bundled and not run
96
- Reference result: expected_json
97
- Reference source: test/unittests/extraMetricsTest.js: VisualComplete85, 95 and 99 = 1000, 1500, 2000 (lines 10 to 12); VisualReadiness 1500 (line 21); 3000 each after a drop (lines 34 to 36); 500, 3000, 3000 for a dip below 95 (lines 46 to 48)
98
  Check: Final file state: the agent receives each progress curve as JSON and writes the milestones to a result JSON; compare with the inline expected values by exact equality (ava t.is); the rules are named in the test titles (a milestone is the first time the curve reaches the level and stays above it, line 24; VisualReadiness is Last minus First Visual Change, line 15)
99
  Tools: a script that reads the visual progress series (as written by browsertime --visualMetrics or WebPageTest) and computes the milestones; browsertime run headless with --visualMetrics produces them from a video
100
  Companions (not included): LICENSE
@@ -102,8 +82,6 @@ Files:
102
  - `test/unittests/mediawikiResourceLoaderTest.js`: MediaWiki ResourceLoader bundles in browser coverage: split a load.php JavaScript bundle into per-module used and unused byte buckets, and label load.php request URLs by their modules and only= parameters.
103
  Format: ava test file (ES module JavaScript) that builds a load.php style bundle string and coverage ranges inline, plus load.php URLs, with expected buckets and labels inline
104
  Origin: unit tests of the browsertime MediaWiki ResourceLoader attribution (tool_fixture); the library source is not fetched, not bundled and not run
105
- Reference result: expected_json
106
- Reference source: test/unittests/mediawikiResourceLoaderTest.js: per-module buckets and usedBytes 36 for the half used module (lines 62 to 105); bundle edge cases (lines 111 to 119); load.php labels such as load.php[startup] and load.php[scripts:jquery] (lines 125 to 204, for example line 164)
107
  Check: Final file state: the agent receives the bundle text, coverage ranges and URLs and writes the buckets and labels to a result JSON; compare with the inline expected values (ava t.is and t.deepEqual)
108
  Tools: a script that reads the coverage JSON and the bundle text and splits by mw.loader.impl boundaries, plus a URL parser for load.php query strings; browsertime run headless with its coverage option on a Wikipedia page
109
  Companions (not included): LICENSE
 
8
 
9
  **Self-contained:** Each file is self-contained. 10 of 10 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** A script that reads the trace JSON (traceEvents) and computes the figures, or browsertime's trace analysis run headless on a saved trace (browsertime --chrome.timeline with the trace exported to JSON); the DevTools Performance panel shows the same events in its main-thread and frames tracks. Other files: see the notes.
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `test/unittests/traceAnalysisTest.js`: Total Blocking Time and its attribution from a synthetic Chrome trace: blocking time per URL and per kind of work, the window before Largest Contentful Paint, and the first-party versus third-party domain split.
20
  Format: ava test file (ES module JavaScript) that builds the Chrome trace events inline and states the expected results inline; the input trace can be written out as JSON for the task
21
  Origin: unit tests of the browsertime Chrome trace analysis (tool_fixture); the library source is not fetched, not bundled and not run
 
 
22
  Check: Final file state: the agent receives the trace written out as JSON and writes the figures to a result JSON; compare field by field with the inline expected values (ava t.is and t.deepEqual, exact)
23
  Tools: a script that reads the trace JSON (traceEvents) and computes the figures, or browsertime's trace analysis run headless on a saved trace (browsertime --chrome.timeline with the trace exported to JSON); the DevTools Performance panel shows the same events in its main-thread and frames tracks
24
  Companions (not included): LICENSE
 
26
  - `test/unittests/frameStabilityTest.js`: Frame stability from a synthetic Chrome trace of PipelineReporter events: presented, partial and dropped frames, effective fps and the longest gap between presented frames.
27
  Format: ava test file (ES module JavaScript) that builds the Chrome trace events inline and states the expected results inline; the input trace can be written out as JSON for the task
28
  Origin: unit tests of the browsertime Chrome trace analysis (tool_fixture); the library source is not fetched, not bundled and not run
 
 
29
  Check: Final file state: the agent receives the trace written out as JSON and writes the figures to a result JSON; compare field by field with the inline expected values (ava t.is and t.deepEqual, exact)
30
  Tools: a script that reads the trace JSON (traceEvents) and computes the figures, or browsertime's trace analysis run headless on a saved trace (browsertime --chrome.timeline with the trace exported to JSON); the DevTools Performance panel shows the same events in its main-thread and frames tracks
31
  Companions (not included): LICENSE
 
33
  - `test/unittests/styleInvalidationsTest.js`: Style and layout invalidations from a synthetic Chrome trace: recalculation and layout counts, reasons, triggering selectors and source scripts, split at first paint and at Largest Contentful Paint.
34
  Format: ava test file (ES module JavaScript) that builds the Chrome trace events inline and states the expected results inline; the input trace can be written out as JSON for the task
35
  Origin: unit tests of the browsertime Chrome trace analysis (tool_fixture); the library source is not fetched, not bundled and not run
 
 
36
  Check: Final file state: the agent receives the trace written out as JSON and writes the figures to a result JSON; compare field by field with the inline expected values (ava t.is and t.deepEqual, exact)
37
  Tools: a script that reads the trace JSON (traceEvents) and computes the figures, or browsertime's trace analysis run headless on a saved trace (browsertime --chrome.timeline with the trace exported to JSON); the DevTools Performance panel shows the same events in its main-thread and frames tracks
38
  Companions (not included): LICENSE
 
40
  - `test/unittests/functionCostsTest.js`: Per-function self and total CPU time from the V8 sampling profiler chunks of a Chrome trace: aggregation across chunks, callee breakdown, bundle module resolution, recovery of URLs truncated at 1024 characters and a 1 ms noise limit.
41
  Format: ava test file (ES module JavaScript) that builds the Chrome trace events (ProfileChunk with cpuProfile nodes, samples and timeDeltas) inline and states the expected results inline; the input trace can be written out as JSON for the task
42
  Origin: unit tests of the browsertime function cost analysis of Chrome traces (tool_fixture); the library source is not fetched, not bundled and not run
 
 
43
  Check: Final file state: the agent receives the trace written out as JSON and writes the function costs to a result JSON; compare field by field with the inline expected values (the convention is ava t.deepEqual and t.is, functionCostsTest.js line 147); the conventions behind the numbers are stated in the test: each sample is weighted by its timeDeltas entry in microseconds and reported in ms, meta frames such as (garbage collector) and (root) are skipped (lines 142 to 144), positions are 1-based (line 160), negative deltas are skipped (lines 235 to 236)
44
  Tools: a script that reads the trace JSON (traceEvents with ProfileChunk samples) and writes the per-function costs, or browsertime run headless with --chrome.timeline on a saved trace; the DevTools Performance panel shows the same self and total times in its Bottom-Up and Call Tree views
45
  Companions (not included): LICENSE
 
47
  - `test/unittests/coverageTest.js`: Used JavaScript bytes from V8 block coverage (Chrome DevTools Profiler.takePreciseCoverage ranges): nested ranges where the innermost range wins, sibling functions, dead branches and zero-length scripts.
48
  Format: ava test file (ES module JavaScript) that states each V8 coverage object (functions with ranges of startOffset, endOffset and count) and the script length inline, with the expected used byte count inline
49
  Origin: unit tests of the browsertime JavaScript coverage computation (tool_fixture); the library source is not fetched, not bundled and not run
 
 
50
  Check: Final file state: the agent receives each coverage object and script length as JSON and writes the used byte count per case to a result JSON; compare per case with the inline expected value by exact integer equality (ava t.is); the rule stated in the test is that the innermost range wins (lines 27 to 29)
51
  Tools: a script that reads the V8 precise coverage JSON (from Profiler.takePreciseCoverage over CDP, or a saved coverage file) and computes used bytes per script; browsertime run headless with its coverage option produces the same figure
52
  Companions (not included): LICENSE
 
54
  - `test/unittests/selectorStatsTest.js`: CSS selector cost from the SelectorStats events of a Chrome trace: totals across style recalculations, per-selector aggregation in ms and the list of wasted selectors (many match attempts, no match).
55
  Format: ava test file (ES module JavaScript) that builds the SelectorStats trace events inline (selector, elapsed, match_attempts, match_count, style_sheet_id) and states the expected results inline
56
  Origin: unit tests of the browsertime selector statistics from Chrome traces (tool_fixture); the library source is not fetched, not bundled and not run
 
 
57
  Check: Final file state: the agent receives the trace events as JSON and writes the selector statistics to a result JSON; compare field by field with the inline expected values (ava t.is and t.deepEqual); elapsed is converted from microseconds to ms as the expected values show
58
  Tools: a script that reads the Chrome trace JSON (traceEvents) and writes the result JSON, or browsertime run headless with --chrome.timeline and its trace post-processing on a saved trace; the DevTools Performance panel shows the same figures interactively
59
  Companions (not included): LICENSE
 
61
  - `test/unittests/timerCostsTest.js`: JavaScript timer costs from a Chrome trace: TimerInstall and TimerFire events joined per timer, fire time, zero-timeout and recurring installs, grouped by script URL and by install site.
62
  Format: ava test file (ES module JavaScript) that builds the timer trace events inline and states the expected results inline
63
  Origin: unit tests of the browsertime timer cost analysis of Chrome traces (tool_fixture); the library source is not fetched, not bundled and not run
 
 
64
  Check: Final file state: the agent receives the trace events as JSON and writes the timer cost tables to a result JSON; compare field by field with the inline expected values (ava t.deepEqual); install sites are 1-based line and column as the expected rows show
65
  Tools: a script that reads the Chrome trace JSON (traceEvents) and writes the result JSON, or browsertime run headless with --chrome.timeline and its trace post-processing on a saved trace; the DevTools Performance panel shows the same figures interactively
66
  Companions (not included): LICENSE
 
68
  - `test/unittests/nonCompositedAnimationsTest.js`: Non-composited animations from Chrome trace Animation events: dedupe by animation id and decode the compositeFailed bitmask into named failure reasons, keeping unknown bits.
69
  Format: ava test file (ES module JavaScript) that builds the Animation trace events inline and states the expected reasons inline
70
  Origin: unit tests of the browsertime non-composited animation analysis of Chrome traces (tool_fixture); the library source is not fetched, not bundled and not run
 
 
71
  Check: Final file state: the agent receives the trace events as JSON and writes the animation list to a result JSON; compare with the inline expected values (ava t.is and t.deepEqual); the bit to reason table of Blink must be given in the task text for the bits the test uses (13 and 17), other bits map to unknown reason (bit N) as line 30 shows
72
  Tools: a script that reads the Chrome trace JSON (traceEvents) and writes the result JSON, or browsertime run headless with --chrome.timeline and its trace post-processing on a saved trace; the DevTools Performance panel shows the same figures interactively
73
  Companions (not included): LICENSE
 
75
  - `test/unittests/extraMetricsTest.js`: Visual metrics milestones from a visual progress curve of a page load: VisualComplete85, 95 and 99 (including a curve that paints backward) and VisualReadiness.
76
  Format: ava test file (ES module JavaScript) that states each visual progress curve and first and last visual change inline with the expected milestones inline
77
  Origin: unit tests of the browsertime extra visual metrics (tool_fixture); the library source is not fetched, not bundled and not run
 
 
78
  Check: Final file state: the agent receives each progress curve as JSON and writes the milestones to a result JSON; compare with the inline expected values by exact equality (ava t.is); the rules are named in the test titles (a milestone is the first time the curve reaches the level and stays above it, line 24; VisualReadiness is Last minus First Visual Change, line 15)
79
  Tools: a script that reads the visual progress series (as written by browsertime --visualMetrics or WebPageTest) and computes the milestones; browsertime run headless with --visualMetrics produces them from a video
80
  Companions (not included): LICENSE
 
82
  - `test/unittests/mediawikiResourceLoaderTest.js`: MediaWiki ResourceLoader bundles in browser coverage: split a load.php JavaScript bundle into per-module used and unused byte buckets, and label load.php request URLs by their modules and only= parameters.
83
  Format: ava test file (ES module JavaScript) that builds a load.php style bundle string and coverage ranges inline, plus load.php URLs, with expected buckets and labels inline
84
  Origin: unit tests of the browsertime MediaWiki ResourceLoader attribution (tool_fixture); the library source is not fetched, not bundled and not run
 
 
85
  Check: Final file state: the agent receives the bundle text, coverage ranges and URLs and writes the buckets and labels to a result JSON; compare with the inline expected values (ava t.is and t.deepEqual)
86
  Tools: a script that reads the coverage JSON and the bundle text and splits by mw.loader.impl boundaries, plus a URL parser for load.php query strings; browsertime run headless with its coverage option on a Wikipedia page
87
  Companions (not included): LICENSE
campaign_files/2026-10-gt-browser-tests-2/browsertime-visualmetrics-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_json. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** python3 visualmetrics-portable.py headless over the frames directory (as browsertime --visualMetrics does after recording with ffmpeg), or a script with Pillow and numpy computing colour histograms per frame
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `visualmetrics/test_visualmetrics.py`: Visual metrics of a page load from 18 filmstrip frames (PNG, named by ms): First and Last Visual Change, Speed Index and the visual progress series.
22
  Format: Python unittest file plus 18 PNG frames named ms_NNNNNN.png (time in ms) of one page load filmstrip
23
  Origin: unit tests and frames of the browsertime visual metrics script (tool_fixture); the script is not fetched, not bundled and not run
24
- Reference result: expected_json
25
- Reference source: visualmetrics/test_visualmetrics.py: First Visual Change 920 and Last Visual Change 6000 (lines 69 to 70); Speed Index 1125 (line 73); the visual progress string (lines 33 to 37, checked at line 76); speed index 1500 for a 3 point curve (lines 78 to 84); frame diff times (line 107)
26
  Check: Final file state: the agent receives the 18 frames and writes the metrics to a result JSON; compare with the expected values by exact equality (unittest assertEqual); the progress values depend on the histogram method of visualmetrics-portable.py, which the task text must state; Contentful and Perceptual Speed Index (lines 86 to 94) need OpenCV and pyssim and are excluded.
27
  Tools: python3 visualmetrics-portable.py headless over the frames directory (as browsertime --visualMetrics does after recording with ffmpeg), or a script with Pillow and numpy computing colour histograms per frame
28
  Companions (not included): LICENSE; visualmetrics/test_data/ms_000000.png; visualmetrics/test_data/ms_000920.png; visualmetrics/test_data/ms_001000.png; visualmetrics/test_data/ms_001080.png; visualmetrics/test_data/ms_001200.png; visualmetrics/test_data/ms_001240.png; visualmetrics/test_data/ms_001280.png; visualmetrics/test_data/ms_001360.png; visualmetrics/test_data/ms_001400.png; visualmetrics/test_data/ms_001520.png; visualmetrics/test_data/ms_002040.png; visualmetrics/test_data/ms_002600.png; visualmetrics/test_data/ms_003160.png; visualmetrics/test_data/ms_003720.png; visualmetrics/test_data/ms_004280.png; visualmetrics/test_data/ms_004880.png; visualmetrics/test_data/ms_005440.png; visualmetrics/test_data/ms_006000.png
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** python3 visualmetrics-portable.py headless over the frames directory (as browsertime --visualMetrics does after recording with ffmpeg), or a script with Pillow and numpy computing colour histograms per frame
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `visualmetrics/test_visualmetrics.py`: Visual metrics of a page load from 18 filmstrip frames (PNG, named by ms): First and Last Visual Change, Speed Index and the visual progress series.
20
  Format: Python unittest file plus 18 PNG frames named ms_NNNNNN.png (time in ms) of one page load filmstrip
21
  Origin: unit tests and frames of the browsertime visual metrics script (tool_fixture); the script is not fetched, not bundled and not run
 
 
22
  Check: Final file state: the agent receives the 18 frames and writes the metrics to a result JSON; compare with the expected values by exact equality (unittest assertEqual); the progress values depend on the histogram method of visualmetrics-portable.py, which the task text must state; Contentful and Perceptual Speed Index (lines 86 to 94) need OpenCV and pyssim and are excluded.
23
  Tools: python3 visualmetrics-portable.py headless over the frames directory (as browsertime --visualMetrics does after recording with ffmpeg), or a script with Pillow and numpy computing colour histograms per frame
24
  Companions (not included): LICENSE; visualmetrics/test_data/ms_000000.png; visualmetrics/test_data/ms_000920.png; visualmetrics/test_data/ms_001000.png; visualmetrics/test_data/ms_001080.png; visualmetrics/test_data/ms_001200.png; visualmetrics/test_data/ms_001240.png; visualmetrics/test_data/ms_001280.png; visualmetrics/test_data/ms_001360.png; visualmetrics/test_data/ms_001400.png; visualmetrics/test_data/ms_001520.png; visualmetrics/test_data/ms_002040.png; visualmetrics/test_data/ms_002600.png; visualmetrics/test_data/ms_003160.png; visualmetrics/test_data/ms_003720.png; visualmetrics/test_data/ms_004280.png; visualmetrics/test_data/ms_004880.png; visualmetrics/test_data/ms_005440.png; visualmetrics/test_data/ms_006000.png
campaign_files/2026-10-gt-browser-tests-2/calamine-tests-ods/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_values_in_test_asserts. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** calamine Rust crate (open_workbook_auto on .ods), or Python odfpy / pandas.read_excel(engine='odf'); a script writes the used range of each sheet as JSON
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `tests/number_rows_repeated.ods`: Read the used range of each sheet of an ODS file whose rows are stored with number-rows-repeated.
22
  Format: OpenDocument spreadsheet (.ods) as input; per-sheet used range (start cell and cell grid) as answer
23
  Origin: test fixture of calamine for ODS repeated rows (tool_fixture)
24
- Reference result: expected_values_in_test_asserts
25
- Reference source: tests/test.rs lines 1779-1849 (fn ods_number_rows_repeated: three sheets with 0, 1 and 2 leading empty rows; cropped range rows A B / C D / C D / empty / empty / C D / empty / C D)
26
  Check: Final file state: the agent's JSON of each sheet's start offset and cell grid compared to tests/test.rs lines 1779-1849 with exact equality (range_eq).
27
  Tools: calamine Rust crate (open_workbook_auto on .ods), or Python odfpy / pandas.read_excel(engine='odf'); a script writes the used range of each sheet as JSON
28
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** calamine Rust crate (open_workbook_auto on .ods), or Python odfpy / pandas.read_excel(engine='odf'); a script writes the used range of each sheet as JSON
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `tests/number_rows_repeated.ods`: Read the used range of each sheet of an ODS file whose rows are stored with number-rows-repeated.
20
  Format: OpenDocument spreadsheet (.ods) as input; per-sheet used range (start cell and cell grid) as answer
21
  Origin: test fixture of calamine for ODS repeated rows (tool_fixture)
 
 
22
  Check: Final file state: the agent's JSON of each sheet's start offset and cell grid compared to tests/test.rs lines 1779-1849 with exact equality (range_eq).
23
  Tools: calamine Rust crate (open_workbook_auto on .ods), or Python odfpy / pandas.read_excel(engine='odf'); a script writes the used range of each sheet as JSON
24
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
campaign_files/2026-10-gt-browser-tests-2/calamine-tests-xlsx/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 6 of 6 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_values_in_test_asserts. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `tests/temperature-table.xlsx`: List the Excel tables of a workbook (Temperature and OtherTable) with their column names and data.
22
  Format: Office Open XML workbook (.xlsx) as input; table names, columns and data values as answer
23
  Origin: test fixture of calamine built for its Excel table reader (tool_fixture)
24
- Reference result: expected_values_in_test_asserts
25
- Reference source: tests/test.rs lines 689-730 (fn table: tables Temperature with columns label, value and rows celsius 22.2222, fahrenheit 72.0; OtherTable with columns label2, value2 and rows something 12.5, else 64.0) and lines 733-822 (fn table_by_ref: the same tables read by reference)
26
  Check: Final file state: the agent writes a JSON of table names, column names and data rows; compare to the values asserted in tests/test.rs lines 689-822 with exact equality (floats as written in the asserts). Upstream judge: run the two calamine tests at the pinned commit.
27
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
28
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
@@ -30,8 +26,6 @@ Files:
30
  - `tests/merged_range.xlsx`: List the merged cell regions of every sheet of a workbook.
31
  Format: .xlsx as input; list of merged regions (sheet name, start and end cell) as answer
32
  Origin: test fixture of calamine for merged-cell reading (tool_fixture)
33
- Reference result: expected_values_in_test_asserts
34
- Reference source: tests/test.rs lines 1191-1379 and lines 1540-1559 (fn merge_cells_all_sheets_by_name: 15 regions, 9 in Sheet1 and 6 in Sheet2)
35
  Check: Final file state: the agent's JSON of merged regions compared as a set per sheet to the regions asserted in tests/test.rs lines 1191-1379, and the count to 15 (lines 1540-1559); exact equality.
36
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
37
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
@@ -39,8 +33,6 @@ Files:
39
  - `tests/date.xlsx`: Read the typed values (date, duration, number) of a small workbook range, with its 1900 date system.
40
  Format: .xlsx as input; typed cell values of A1:B3 as answer
41
  Origin: test fixture of calamine for date and duration typing (tool_fixture)
42
- Reference result: expected_values_in_test_asserts
43
- Reference source: tests/test.rs lines 897-930 (fn date_xlsx: A1 DateTime serial 44197.0 = 2021-01-01, A3 TimeDelta 10.6320601851852 = 255h10m10s, not 1904 epoch), lines 2152-2157 (fn issue_401_empty_tables: no tables) and lines 2527-2560 (fn test_ref_xlsx: 3x2 range [44197 DateTime, 15], [44198, 16], [10.632... TimeDelta, 17])
44
  Check: Final file state: the agent's JSON gives each cell's type and value; compare to tests/test.rs lines 897-930 and 2527-2560 with exact equality of the serial values as written in the asserts and of the types (date versus duration versus number).
45
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
46
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
@@ -48,8 +40,6 @@ Files:
48
  - `tests/shared_formula_simple.xlsx`: Expand a shared formula saved by Excel: report the formula and value of each cell of B1:B3.
49
  Format: .xlsx as input; per-cell formula text, value and shared-formula range as answer
50
  Origin: test fixture of calamine for shared formulas (tool_fixture)
51
- Reference result: expected_values_in_test_asserts
52
- Reference source: tests/test.rs lines 2178-2204 (B1 = A1*2 value 2, B2 = A2*2 value 4, B3 = A3*2 value 6), lines 2207-2240 (shared index 0, range B1:B3) and lines 2243-2250 (worksheet_formula expands the same formulas)
53
  Check: Final file state: the agent's JSON gives formula and value per cell; compare formula strings and values to tests/test.rs lines 2178-2250 with exact equality.
54
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
55
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
@@ -57,8 +47,6 @@ Files:
57
  - `tests/header-row.xlsx`: Report the data range of a sheet under three header-row choices (first non-empty row, row 0, row 8).
58
  Format: .xlsx as input; start cell, end cell, first row and cell count per choice as answer
59
  Origin: test fixture of calamine for header-row selection (tool_fixture)
60
- Reference result: expected_values_in_test_asserts
61
- Reference source: tests/test.rs lines 2599-2601 (rstest cases) and lines 2606-2632 (fn test_header_row_xlsx: FirstNonEmptyRow start (2,0) end (9,3), first row [Empty, Empty, "Note 1", Empty], 32 cells; Row(0) (0,0) to (9,3), 40 cells; Row(8) start (8,0), first row [Columns, Column A, Column B, Column C], 8 cells)
62
  Check: Final file state: the agent's JSON per choice compared to tests/test.rs lines 2599-2632 with exact equality of start, end, first row and cell count.
63
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
64
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
@@ -66,8 +54,6 @@ Files:
66
  - `tests/column_row_ranges.xlsx`: Read the formulas with whole-column and whole-row ranges of two sheets.
67
  Format: .xlsx as input; formula grid of sheets Column ranges and Row ranges as answer
68
  Origin: test fixture of calamine for column and row range formulas (tool_fixture)
69
- Reference result: expected_values_in_test_asserts
70
- Reference source: tests/test.rs lines 2437-2484 (fn column_row_ranges: sheet Column ranges 5x3 formulas SUM(E:F), SUM(F:G), SUM(G:H); sheet Row ranges rows SUM(5:6), SUM(6:7), SUM(7:8), 5 columns each)
71
  Check: Final file state: the agent's JSON formula grid compared to tests/test.rs lines 2437-2484 with exact string equality (range_eq).
72
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
73
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 6 of 6 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `tests/temperature-table.xlsx`: List the Excel tables of a workbook (Temperature and OtherTable) with their column names and data.
20
  Format: Office Open XML workbook (.xlsx) as input; table names, columns and data values as answer
21
  Origin: test fixture of calamine built for its Excel table reader (tool_fixture)
 
 
22
  Check: Final file state: the agent writes a JSON of table names, column names and data rows; compare to the values asserted in tests/test.rs lines 689-822 with exact equality (floats as written in the asserts). Upstream judge: run the two calamine tests at the pinned commit.
23
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
24
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
 
26
  - `tests/merged_range.xlsx`: List the merged cell regions of every sheet of a workbook.
27
  Format: .xlsx as input; list of merged regions (sheet name, start and end cell) as answer
28
  Origin: test fixture of calamine for merged-cell reading (tool_fixture)
 
 
29
  Check: Final file state: the agent's JSON of merged regions compared as a set per sheet to the regions asserted in tests/test.rs lines 1191-1379, and the count to 15 (lines 1540-1559); exact equality.
30
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
31
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
 
33
  - `tests/date.xlsx`: Read the typed values (date, duration, number) of a small workbook range, with its 1900 date system.
34
  Format: .xlsx as input; typed cell values of A1:B3 as answer
35
  Origin: test fixture of calamine for date and duration typing (tool_fixture)
 
 
36
  Check: Final file state: the agent's JSON gives each cell's type and value; compare to tests/test.rs lines 897-930 and 2527-2560 with exact equality of the serial values as written in the asserts and of the types (date versus duration versus number).
37
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
38
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
 
40
  - `tests/shared_formula_simple.xlsx`: Expand a shared formula saved by Excel: report the formula and value of each cell of B1:B3.
41
  Format: .xlsx as input; per-cell formula text, value and shared-formula range as answer
42
  Origin: test fixture of calamine for shared formulas (tool_fixture)
 
 
43
  Check: Final file state: the agent's JSON gives formula and value per cell; compare formula strings and values to tests/test.rs lines 2178-2250 with exact equality.
44
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
45
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
 
47
  - `tests/header-row.xlsx`: Report the data range of a sheet under three header-row choices (first non-empty row, row 0, row 8).
48
  Format: .xlsx as input; start cell, end cell, first row and cell count per choice as answer
49
  Origin: test fixture of calamine for header-row selection (tool_fixture)
 
 
50
  Check: Final file state: the agent's JSON per choice compared to tests/test.rs lines 2599-2632 with exact equality of start, end, first row and cell count.
51
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
52
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
 
54
  - `tests/column_row_ranges.xlsx`: Read the formulas with whole-column and whole-row ranges of two sheets.
55
  Format: .xlsx as input; formula grid of sheets Column ranges and Row ranges as answer
56
  Origin: test fixture of calamine for column and row range formulas (tool_fixture)
 
 
57
  Check: Final file state: the agent's JSON formula grid compared to tests/test.rs lines 2437-2484 with exact string equality (range_eq).
58
  Tools: calamine Rust crate (open_workbook_auto, worksheet_range, ...), or Python openpyxl / odfpy / pandas.read_excel; a script reads the workbook and writes the asked values to a JSON file
59
  Companions (not included): Cargo.toml (licence line 8 and version 0.36.1); LICENSE-MIT.md (licence text); tests/test.rs (the comparing test functions; the expected values live in its asserts, hide from the agent)
campaign_files/2026-10-gt-browser-tests-2/chrome-devtools-mcp-heap-fixtures/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 2 of 2 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_numbers; expected_table. Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** a script that parses the .heapsnapshot JSON (snapshot.meta node and edge fields, nodes, edges and strings arrays) and diffs objects per constructor between two snapshots; or the DevTools heap snapshot model run under Node (chrome-devtools-mcp HeapSnapshotManager, memlab heap analysis); a script that parses the .heapsnapshot JSON and aggregates nodes per constructor (count, shallow size, retained size from the dominator tree); or the DevTools heap snapshot model run under Node (chrome-devtools-mcp HeapSnapshotManager, memlab heap analysis)
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,8 +19,6 @@ Units:
21
  - `tests/fixtures/` (3 files): Heap snapshot comparison: count of NewObject instances added between V8 heap snapshots heap-1 and heap-3, and heap-2 and heap-3.
22
  Format: V8 heap snapshot JSON (.heapsnapshot: snapshot.meta, nodes, edges, strings), about 1.2 MB each
23
  Origin: test fixtures of the Chrome DevTools MCP server heap snapshot processor, recorded from the d8 shell (tool_fixture); the server source is not fetched, not bundled and not run
24
- Reference result: expected_numbers
25
- Reference source: tests/processors/HeapSnapshotManager.test.ts lines 55 to 73: getClassDiffs(heap-1, heap-3) and getClassDiffs(heap-2, heap-3), then assert.equal(firstNewObjectDiff.addedCount, 7) and assert.equal(secondNewObjectDiff.addedCount, 5)
26
  Check: Final file state: the agent writes, for class NewObject, the number of objects added from heap-1 to heap-3 and from heap-2 to heap-3 (the # New column of the DevTools Memory panel Comparison view) to a result file; require exact integer equality with 7 and 5
27
  Tools: a script that parses the .heapsnapshot JSON (snapshot.meta node and edge fields, nodes, edges and strings arrays) and diffs objects per constructor between two snapshots; or the DevTools heap snapshot model run under Node (chrome-devtools-mcp HeapSnapshotManager, memlab heap analysis)
28
  Companions (not included): tests/processors/HeapSnapshotManager.test.ts; LICENSE
@@ -30,8 +26,6 @@ Units:
30
  - `tests/fixtures/example.heapsnapshot`: Heap snapshot summary of an example.com page: per-constructor count, shallow size and retained size (the DevTools Memory panel Summary view), plus native context sizes.
31
  Format: V8 heap snapshot JSON (.heapsnapshot), about 2.4 MB
32
  Origin: test fixture of the Chrome DevTools MCP server heap snapshot formatter (tool_fixture); the server source is not fetched, not bundled and not run
33
- Reference result: expected_table
34
- Reference source: tests/formatters/HeapSnapshotFormatter.test.js.snapshot, the entries under 'with real fixtures' for example.heapsnapshot: aggregates with default options (line 433 onward: id,name,count,selfSize,retainedSize), with the filters objectsRetainedByContexts, sharedNativeContext and attributedToSpecificNativeContext with objectId 7249, and native context sizes; HeapSnapshotFormatter.test.ts lines 459 to 520 produce them with t.assert.snapshot
35
  Check: Final file state: the agent writes a CSV with one line per constructor (name, count, self size, retained size); require the count column to equal the snapshot entry exactly and the two size columns to equal it after rounding the bytes the way the snapshot prints them (the kB display of the stored snapshot, as in '199 kB' and '0.0 kB'); row order by retained size as stored
36
  Tools: a script that parses the .heapsnapshot JSON and aggregates nodes per constructor (count, shallow size, retained size from the dominator tree); or the DevTools heap snapshot model run under Node (chrome-devtools-mcp HeapSnapshotManager, memlab heap analysis)
37
  Companions (not included): tests/formatters/HeapSnapshotFormatter.test.ts; tests/formatters/HeapSnapshotFormatter.test.js.snapshot; LICENSE
 
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 2 of 2 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** a script that parses the .heapsnapshot JSON (snapshot.meta node and edge fields, nodes, edges and strings arrays) and diffs objects per constructor between two snapshots; or the DevTools heap snapshot model run under Node (chrome-devtools-mcp HeapSnapshotManager, memlab heap analysis); a script that parses the .heapsnapshot JSON and aggregates nodes per constructor (count, shallow size, retained size from the dominator tree); or the DevTools heap snapshot model run under Node (chrome-devtools-mcp HeapSnapshotManager, memlab heap analysis)
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `tests/fixtures/` (3 files): Heap snapshot comparison: count of NewObject instances added between V8 heap snapshots heap-1 and heap-3, and heap-2 and heap-3.
20
  Format: V8 heap snapshot JSON (.heapsnapshot: snapshot.meta, nodes, edges, strings), about 1.2 MB each
21
  Origin: test fixtures of the Chrome DevTools MCP server heap snapshot processor, recorded from the d8 shell (tool_fixture); the server source is not fetched, not bundled and not run
 
 
22
  Check: Final file state: the agent writes, for class NewObject, the number of objects added from heap-1 to heap-3 and from heap-2 to heap-3 (the # New column of the DevTools Memory panel Comparison view) to a result file; require exact integer equality with 7 and 5
23
  Tools: a script that parses the .heapsnapshot JSON (snapshot.meta node and edge fields, nodes, edges and strings arrays) and diffs objects per constructor between two snapshots; or the DevTools heap snapshot model run under Node (chrome-devtools-mcp HeapSnapshotManager, memlab heap analysis)
24
  Companions (not included): tests/processors/HeapSnapshotManager.test.ts; LICENSE
 
26
  - `tests/fixtures/example.heapsnapshot`: Heap snapshot summary of an example.com page: per-constructor count, shallow size and retained size (the DevTools Memory panel Summary view), plus native context sizes.
27
  Format: V8 heap snapshot JSON (.heapsnapshot), about 2.4 MB
28
  Origin: test fixture of the Chrome DevTools MCP server heap snapshot formatter (tool_fixture); the server source is not fetched, not bundled and not run
 
 
29
  Check: Final file state: the agent writes a CSV with one line per constructor (name, count, self size, retained size); require the count column to equal the snapshot entry exactly and the two size columns to equal it after rounding the bytes the way the snapshot prints them (the kB display of the stored snapshot, as in '199 kB' and '0.0 kB'); row order by retained size as stored
30
  Tools: a script that parses the .heapsnapshot JSON and aggregates nodes per constructor (count, shallow size, retained size from the dominator tree); or the DevTools heap snapshot model run under Node (chrome-devtools-mcp HeapSnapshotManager, memlab heap analysis)
31
  Companions (not included): tests/formatters/HeapSnapshotFormatter.test.ts; tests/formatters/HeapSnapshotFormatter.test.js.snapshot; LICENSE
campaign_files/2026-10-gt-browser-tests-2/chrome-har-perflogs/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_har_properties. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Node chrome-har `harFromMessages(perflog)` or any script that turns Chrome DevTools protocol Network events into HAR 1.2 JSON; headless Chrome produces the perflog via chromedriver or CDP
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/perflogs/akamai-h2push.json`: Convert a Chrome DevTools protocol performance log to HAR and identify pushed assets.
22
  Format: JSON array of 7138 DevTools protocol messages (Network.* and Tracing.dataCollected) as input; HAR 1.2 JSON as output
23
  Origin: regression perflog and test of the chrome-har converter used by sitespeed.io and browsertime (tool_fixture)
24
- Reference result: expected_har_properties
25
- Reference source: test/tests.js "Includes pushed assets": the HAR has 1 page, 361 entries for the tile images, 3 entries marked as pushed
26
  Check: Final file state: the agent writes a HAR file; validate it as HAR 1.2 JSON, then check the asserted properties (page count 1, tile entry count 361, pushed entry count 3 via the _was_pushed field). No screenshot needed
27
  Tools: Node chrome-har `harFromMessages(perflog)` or any script that turns Chrome DevTools protocol Network events into HAR 1.2 JSON; headless Chrome produces the perflog via chromedriver or CDP
28
  Companions (not included): test/tests.js test "Includes pushed assets" (lines 341 to 359) holds the assertions
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Node chrome-har `harFromMessages(perflog)` or any script that turns Chrome DevTools protocol Network events into HAR 1.2 JSON; headless Chrome produces the perflog via chromedriver or CDP
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/perflogs/akamai-h2push.json`: Convert a Chrome DevTools protocol performance log to HAR and identify pushed assets.
20
  Format: JSON array of 7138 DevTools protocol messages (Network.* and Tracing.dataCollected) as input; HAR 1.2 JSON as output
21
  Origin: regression perflog and test of the chrome-har converter used by sitespeed.io and browsertime (tool_fixture)
 
 
22
  Check: Final file state: the agent writes a HAR file; validate it as HAR 1.2 JSON, then check the asserted properties (page count 1, tile entry count 361, pushed entry count 3 via the _was_pushed field). No screenshot needed
23
  Tools: Node chrome-har `harFromMessages(perflog)` or any script that turns Chrome DevTools protocol Network events into HAR 1.2 JSON; headless Chrome produces the perflog via chromedriver or CDP
24
  Companions (not included): test/tests.js test "Includes pushed assets" (lines 341 to 359) holds the assertions
campaign_files/2026-10-gt-browser-tests-2/chrome-recorder-cypress-export-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_generated_code. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** @cypress/chrome-recorder CLI (npx @cypress/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a Cypress command
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/CypressStringifyExtension_test.ts`: Chrome DevTools Recorder export to Cypress: 15 cases that turn Recorder steps (setViewport, navigate, click, change, keyDown, scroll, hover, waitForElement, assertions) into Cypress test code.
22
  Format: TypeScript test file; each case builds a Recorder step or flow object (Chrome DevTools Recorder JSON schema of @puppeteer/replay) inline and states the expected Cypress code string inline
23
  Origin: unit tests of the Cypress exporter for the Chrome DevTools Recorder (tool_fixture); the exporter source is not fetched, not bundled and not run
24
- Reference result: expected_generated_code
25
- Reference source: the expected strings written inline in test/CypressStringifyExtension_test.ts (15 cases); assertions use chai assert.equal(result.toString(), ...)
26
  Check: Final file state: the agent receives a Recorder JSON recording (the step objects of a case) and writes the Cypress test file the Recorder panel Export menu would produce; require each step's line to equal the expected string exactly (substring match only for the one header case that upstream checks with toContain), then a syntax check with node --check or tsc --noEmit
27
  Tools: @cypress/chrome-recorder CLI (npx @cypress/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a Cypress command
28
  Companions (not included): LICENSE
 
8
 
9
  **Self-contained:** The file is self-contained. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** @cypress/chrome-recorder CLI (npx @cypress/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a Cypress command
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/CypressStringifyExtension_test.ts`: Chrome DevTools Recorder export to Cypress: 15 cases that turn Recorder steps (setViewport, navigate, click, change, keyDown, scroll, hover, waitForElement, assertions) into Cypress test code.
20
  Format: TypeScript test file; each case builds a Recorder step or flow object (Chrome DevTools Recorder JSON schema of @puppeteer/replay) inline and states the expected Cypress code string inline
21
  Origin: unit tests of the Cypress exporter for the Chrome DevTools Recorder (tool_fixture); the exporter source is not fetched, not bundled and not run
 
 
22
  Check: Final file state: the agent receives a Recorder JSON recording (the step objects of a case) and writes the Cypress test file the Recorder panel Export menu would produce; require each step's line to equal the expected string exactly (substring match only for the one header case that upstream checks with toContain), then a syntax check with node --check or tsc --noEmit
23
  Tools: @cypress/chrome-recorder CLI (npx @cypress/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a Cypress command
24
  Companions (not included): LICENSE
campaign_files/2026-10-gt-browser-tests-2/chrome-recorder-nightwatch-export-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_generated_code. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** @nightwatch/chrome-recorder CLI (npx @nightwatch/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a Nightwatch command
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/nightwatchStringifyExtension.test.ts`: Chrome DevTools Recorder export to Nightwatch: 15 cases that turn Recorder steps (setViewport, navigate, click, change, keyDown, scroll, hover, waitForElement, assertions) into Nightwatch test code.
22
  Format: TypeScript test file; each case builds a Recorder step or flow object (Chrome DevTools Recorder JSON schema of @puppeteer/replay) inline and states the expected Nightwatch code string inline
23
  Origin: unit tests of the Nightwatch exporter for the Chrome DevTools Recorder (tool_fixture); the exporter source is not fetched, not bundled and not run
24
- Reference result: expected_generated_code
25
- Reference source: the expected strings written inline in test/nightwatchStringifyExtension.test.ts (15 cases); assertions use chai expect(writer.toString()).to.equal(...)
26
  Check: Final file state: the agent receives a Recorder JSON recording (the step objects of a case) and writes the Nightwatch test file the Recorder panel Export menu would produce; require each step's line to equal the expected string exactly (substring match only for the one header case that upstream checks with toContain), then a syntax check with node --check or tsc --noEmit
27
  Tools: @nightwatch/chrome-recorder CLI (npx @nightwatch/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a Nightwatch command
28
  Companions (not included): test/InMemoryLineWriter.ts; LICENSE.md
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** @nightwatch/chrome-recorder CLI (npx @nightwatch/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a Nightwatch command
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/nightwatchStringifyExtension.test.ts`: Chrome DevTools Recorder export to Nightwatch: 15 cases that turn Recorder steps (setViewport, navigate, click, change, keyDown, scroll, hover, waitForElement, assertions) into Nightwatch test code.
20
  Format: TypeScript test file; each case builds a Recorder step or flow object (Chrome DevTools Recorder JSON schema of @puppeteer/replay) inline and states the expected Nightwatch code string inline
21
  Origin: unit tests of the Nightwatch exporter for the Chrome DevTools Recorder (tool_fixture); the exporter source is not fetched, not bundled and not run
 
 
22
  Check: Final file state: the agent receives a Recorder JSON recording (the step objects of a case) and writes the Nightwatch test file the Recorder panel Export menu would produce; require each step's line to equal the expected string exactly (substring match only for the one header case that upstream checks with toContain), then a syntax check with node --check or tsc --noEmit
23
  Tools: @nightwatch/chrome-recorder CLI (npx @nightwatch/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a Nightwatch command
24
  Companions (not included): test/InMemoryLineWriter.ts; LICENSE.md
campaign_files/2026-10-gt-browser-tests-2/chrome-recorder-webdriverio-export-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_generated_code. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** @wdio/chrome-recorder CLI (npx @wdio/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a WebdriverIO command
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/stringifyExtension.test.ts`: Chrome DevTools Recorder export to WebdriverIO: 23 cases that turn Recorder steps (setViewport, navigate, click, change, keyDown, scroll, hover, waitForElement, assertions) into WebdriverIO test code.
22
  Format: TypeScript test file; each case builds a Recorder step or flow object (Chrome DevTools Recorder JSON schema of @puppeteer/replay) inline and states the expected WebdriverIO code string inline
23
  Origin: unit tests of the WebdriverIO exporter for the Chrome DevTools Recorder (tool_fixture); the exporter source is not fetched, not bundled and not run
24
- Reference result: expected_generated_code
25
- Reference source: the expected strings written inline in test/stringifyExtension.test.ts (23 cases); assertions use vitest expect(writer.toString()).toBe(...) for most cases and toContain for the test header (line 13)
26
  Check: Final file state: the agent receives a Recorder JSON recording (the step objects of a case) and writes the WebdriverIO test file the Recorder panel Export menu would produce; require each step's line to equal the expected string exactly (substring match only for the one header case that upstream checks with toContain), then a syntax check with node --check or tsc --noEmit
27
  Tools: @wdio/chrome-recorder CLI (npx @wdio/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a WebdriverIO command
28
  Companions (not included): test/InMemoryLineWriter.ts; LICENSE.md
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** @wdio/chrome-recorder CLI (npx @wdio/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a WebdriverIO command
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/stringifyExtension.test.ts`: Chrome DevTools Recorder export to WebdriverIO: 23 cases that turn Recorder steps (setViewport, navigate, click, change, keyDown, scroll, hover, waitForElement, assertions) into WebdriverIO test code.
20
  Format: TypeScript test file; each case builds a Recorder step or flow object (Chrome DevTools Recorder JSON schema of @puppeteer/replay) inline and states the expected WebdriverIO code string inline
21
  Origin: unit tests of the WebdriverIO exporter for the Chrome DevTools Recorder (tool_fixture); the exporter source is not fetched, not bundled and not run
 
 
22
  Check: Final file state: the agent receives a Recorder JSON recording (the step objects of a case) and writes the WebdriverIO test file the Recorder panel Export menu would produce; require each step's line to equal the expected string exactly (substring match only for the one header case that upstream checks with toContain), then a syntax check with node --check or tsc --noEmit
23
  Tools: @wdio/chrome-recorder CLI (npx @wdio/chrome-recorder recording.json) or its Node API on the Recorder JSON, or a script that maps each @puppeteer/replay step to a WebdriverIO command
24
  Companions (not included): test/InMemoryLineWriter.ts; LICENSE.md
campaign_files/2026-10-gt-browser-tests-2/chromium-bidi-unit-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** 3 of 4 files are self-contained; the others are used together with other upstream files. 4 of 4 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_json; expected_text; expected_bool. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** a script (TypeScript or any language) that implements the conversions between WebDriver BiDi and Chrome DevTools protocol network headers and cookies and the BiDi URL pattern match, and writes the results; the DevTools Network panel shows the same headers, its request blocking takes URL patterns and the Application panel edits the same cookies; a script that formats the console arguments (WebDriver BiDi script.RemoteValue objects) with the console format specifiers %d %f %s %o %O %c and writes the message text; the DevTools Console panel renders the same console.log calls; a script using the WHATWG URL parser (node URL or python urllib.parse) that classifies the URLs; chromium-bidi runs headless as the BiDi mapper of Chrome; a script using Intl.Segmenter with granularity grapheme (node) or the regex module \X (python); chromium-bidi runs headless as the BiDi mapper
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `src/bidiMapper/modules/network/NetworkUtils.test.ts`: WebDriver BiDi to Chrome DevTools protocol network mapping: header size, header conversion in both directions (Network and Fetch domains), timing clamp, BiDi URL pattern matching and BiDi cookie to CDP cookie with partition keys.
22
  Format: node:test test file (TypeScript) with chai assertions; inputs and expected values inline
23
  Origin: unit tests of the Chromium WebDriver BiDi mapper (tool_fixture); the mapper source is not fetched, not bundled and not run
24
- Reference result: expected_json
25
- Reference source: src/bidiMapper/modules/network/NetworkUtils.test.ts: header size 0 and 46 (lines 37 and 59); converted header lists and maps (lines 69 to 237); getTiming 0, 0, 1 (lines 243 to 251); URL pattern match results true or false (lines 260 to 389); CDP cookie partitionKey objects and the UnableToSetCookieException (lines 399 to 452)
26
  Check: Final file state: the agent writes, for each input of the test, the converted object or the match result to a result JSON; compare with the inline expected values by exact and deep equality (the convention is chai assert.equal and assert.deepEqual, NetworkUtils.test.ts lines 37 and 69); the conventions behind the values come from the WebDriver BiDi specification (network.Header, network.UrlPattern, storage cookies) and the Chrome DevTools protocol Network and Fetch domains, and the header size 46 for the two headers at lines 41 to 56 is name, colon, space, value and CRLF per header as in HTTP/1.1 header framing
27
  Tools: a script (TypeScript or any language) that implements the conversions between WebDriver BiDi and Chrome DevTools protocol network headers and cookies and the BiDi URL pattern match, and writes the results; the DevTools Network panel shows the same headers, its request blocking takes URL patterns and the Application panel edits the same cookies
28
  Companions (not included): LICENSE
@@ -30,8 +26,6 @@ Files:
30
  - `src/bidiMapper/modules/log/logHelper.test.ts`: Console message text from WebDriver BiDi remote values: the format specifiers %d %f %s %o %O %c applied to undefined, null, strings, numbers (NaN, -0, Infinity), booleans, objects, arrays, maps and sets, plus the joined text of plain arguments and the errors for too many or too few values.
31
  Format: node:test test file (TypeScript) with a table of cases (STRING_FORMAT_TEST_CASES: input remote value and the expected text per specifier) and chai assertions
32
  Origin: unit tests of the Chromium WebDriver BiDi mapper log helper (tool_fixture); the mapper source is not fetched, not bundled and not run
33
- Reference result: expected_text
34
- Reference source: src/bidiMapper/modules/log/logHelper.test.ts: the expected text per specifier in STRING_FORMAT_TEST_CASES (lines 25 to 280, 19 input values), the joined texts at lines 304 to 320, and the error messages at lines 386 to 403
35
  Check: Final file state: the agent writes the formatted message text for each case to a result file; compare each with the expected string by exact equality (the convention is chai assert.equal, logHelper.test.ts line 295); the formatting rules behind the expected strings are those of the WHATWG Console Standard formatter (%d and %f parse as integer and float, %s as string, %o %O %c as the object preview), as the test cases show; for the two error cases require the stated message text
36
  Tools: a script that formats the console arguments (WebDriver BiDi script.RemoteValue objects) with the console format specifiers %d %f %s %o %O %c and writes the message text; the DevTools Console panel renders the same console.log calls
37
  Companions (not included): LICENSE
@@ -39,8 +33,6 @@ Files:
39
  - `src/utils/urlHelpers.test.ts`: Which URLs count as about:blank in WebDriver BiDi navigation: 5 matching cases (query, fragment, case) and 10 non-matching ones (about:srcdoc, about:blank with path, other schemes).
40
  Format: mocha and chai test file (TypeScript) with each URL and the expected boolean inline
41
  Origin: unit tests of the chromium-bidi URL helpers (tool_fixture); the library source is not fetched, not bundled and not run
42
- Reference result: expected_bool
43
- Reference source: src/utils/urlHelpers.test.ts: assert.isTrue for 5 URLs (lines 27 to 43) and assert.isFalse for 10 URLs (lines 49 to 85)
44
  Check: Final file state: the agent writes a JSON map from each of the 15 URLs to true or false; compare per URL with the inline expected boolean
45
  Tools: a script using the WHATWG URL parser (node URL or python urllib.parse) that classifies the URLs; chromium-bidi runs headless as the BiDi mapper of Chrome
46
  Companions (not included): LICENSE
@@ -48,8 +40,6 @@ Files:
48
  - `src/utils/graphemeTools.test.ts`: Single grapheme checks for WebDriver BiDi input.performActions key values: whether a string is exactly one user-perceived character (surrogate pairs, Tamil and Hangul clusters, emoji with variation selectors) and whether it is a complex one.
49
  Format: mocha and chai test file (TypeScript) with each string (as escapes) and the expected boolean inline
50
  Origin: unit tests of the chromium-bidi grapheme tools (tool_fixture); the library source is not fetched, not bundled and not run
51
- Reference result: expected_bool
52
- Reference source: src/utils/graphemeTools.test.ts: isSingleGrapheme cases at lines 27 to 75 and isSingleComplexGrapheme cases at lines 83 to 113
53
  Check: Final file state: the agent writes a JSON map from each case string to the two booleans; compare per case with the inline expected values; the upstream case at line 75 reuses the emoji pair under a different title, so the case list is taken from the asserted strings, not the titles
54
  Tools: a script using Intl.Segmenter with granularity grapheme (node) or the regex module \X (python); chromium-bidi runs headless as the BiDi mapper
55
  Companions (not included): LICENSE
 
8
 
9
  **Self-contained:** 3 of 4 files are self-contained; the others are used together with other upstream files. 4 of 4 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** a script (TypeScript or any language) that implements the conversions between WebDriver BiDi and Chrome DevTools protocol network headers and cookies and the BiDi URL pattern match, and writes the results; the DevTools Network panel shows the same headers, its request blocking takes URL patterns and the Application panel edits the same cookies; a script that formats the console arguments (WebDriver BiDi script.RemoteValue objects) with the console format specifiers %d %f %s %o %O %c and writes the message text; the DevTools Console panel renders the same console.log calls; a script using the WHATWG URL parser (node URL or python urllib.parse) that classifies the URLs; chromium-bidi runs headless as the BiDi mapper of Chrome; a script using Intl.Segmenter with granularity grapheme (node) or the regex module \X (python); chromium-bidi runs headless as the BiDi mapper
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `src/bidiMapper/modules/network/NetworkUtils.test.ts`: WebDriver BiDi to Chrome DevTools protocol network mapping: header size, header conversion in both directions (Network and Fetch domains), timing clamp, BiDi URL pattern matching and BiDi cookie to CDP cookie with partition keys.
20
  Format: node:test test file (TypeScript) with chai assertions; inputs and expected values inline
21
  Origin: unit tests of the Chromium WebDriver BiDi mapper (tool_fixture); the mapper source is not fetched, not bundled and not run
 
 
22
  Check: Final file state: the agent writes, for each input of the test, the converted object or the match result to a result JSON; compare with the inline expected values by exact and deep equality (the convention is chai assert.equal and assert.deepEqual, NetworkUtils.test.ts lines 37 and 69); the conventions behind the values come from the WebDriver BiDi specification (network.Header, network.UrlPattern, storage cookies) and the Chrome DevTools protocol Network and Fetch domains, and the header size 46 for the two headers at lines 41 to 56 is name, colon, space, value and CRLF per header as in HTTP/1.1 header framing
23
  Tools: a script (TypeScript or any language) that implements the conversions between WebDriver BiDi and Chrome DevTools protocol network headers and cookies and the BiDi URL pattern match, and writes the results; the DevTools Network panel shows the same headers, its request blocking takes URL patterns and the Application panel edits the same cookies
24
  Companions (not included): LICENSE
 
26
  - `src/bidiMapper/modules/log/logHelper.test.ts`: Console message text from WebDriver BiDi remote values: the format specifiers %d %f %s %o %O %c applied to undefined, null, strings, numbers (NaN, -0, Infinity), booleans, objects, arrays, maps and sets, plus the joined text of plain arguments and the errors for too many or too few values.
27
  Format: node:test test file (TypeScript) with a table of cases (STRING_FORMAT_TEST_CASES: input remote value and the expected text per specifier) and chai assertions
28
  Origin: unit tests of the Chromium WebDriver BiDi mapper log helper (tool_fixture); the mapper source is not fetched, not bundled and not run
 
 
29
  Check: Final file state: the agent writes the formatted message text for each case to a result file; compare each with the expected string by exact equality (the convention is chai assert.equal, logHelper.test.ts line 295); the formatting rules behind the expected strings are those of the WHATWG Console Standard formatter (%d and %f parse as integer and float, %s as string, %o %O %c as the object preview), as the test cases show; for the two error cases require the stated message text
30
  Tools: a script that formats the console arguments (WebDriver BiDi script.RemoteValue objects) with the console format specifiers %d %f %s %o %O %c and writes the message text; the DevTools Console panel renders the same console.log calls
31
  Companions (not included): LICENSE
 
33
  - `src/utils/urlHelpers.test.ts`: Which URLs count as about:blank in WebDriver BiDi navigation: 5 matching cases (query, fragment, case) and 10 non-matching ones (about:srcdoc, about:blank with path, other schemes).
34
  Format: mocha and chai test file (TypeScript) with each URL and the expected boolean inline
35
  Origin: unit tests of the chromium-bidi URL helpers (tool_fixture); the library source is not fetched, not bundled and not run
 
 
36
  Check: Final file state: the agent writes a JSON map from each of the 15 URLs to true or false; compare per URL with the inline expected boolean
37
  Tools: a script using the WHATWG URL parser (node URL or python urllib.parse) that classifies the URLs; chromium-bidi runs headless as the BiDi mapper of Chrome
38
  Companions (not included): LICENSE
 
40
  - `src/utils/graphemeTools.test.ts`: Single grapheme checks for WebDriver BiDi input.performActions key values: whether a string is exactly one user-perceived character (surrogate pairs, Tamil and Hangul clusters, emoji with variation selectors) and whether it is a complex one.
41
  Format: mocha and chai test file (TypeScript) with each string (as escapes) and the expected boolean inline
42
  Origin: unit tests of the chromium-bidi grapheme tools (tool_fixture); the library source is not fetched, not bundled and not run
 
 
43
  Check: Final file state: the agent writes a JSON map from each case string to the two booleans; compare per case with the inline expected values; the upstream case at line 75 reuses the emoji pair under a different title, so the case list is taken from the asserted strings, not the titles
44
  Tools: a script using Intl.Segmenter with granularity grapheme (node) or the regex module \X (python); chromium-bidi runs headless as the BiDi mapper
45
  Companions (not included): LICENSE
campaign_files/2026-10-gt-browser-tests-2/chromium-policy-pref-mapping/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 8 of 8 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_policy_and_pref_values. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
14
 
15
  **Licence (SPDX):** BSD-3-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `components/policy/test/data/pref_mapping/BookmarkBarEnabled.json`: Chrome bookmarks bar setting: BookmarkBarEnabled policy mapped to the bookmark_bar prefs (3 cases).
22
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
23
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
24
- Reference result: expected_policy_and_pref_values
25
- Reference source: components/policy/test/data/pref_mapping/BookmarkBarEnabled.json (whole file): three test cases: policy unset (bookmark_bar.show_on_all_tabs default false, bookmark_bar.visibility_state default 1), BookmarkBarEnabled true (show_on_all_tabs true, visibility_state 0), BookmarkBarEnabled false (show_on_all_tabs false, visibility_state 2); can_be_recommended true; os android, win, linux, mac, chromeos, fuchsia; field meanings (policies, prefs, value, default_value, check_for_recommended, simple_policy_pref_mapping_test) in docs/enterprise/policy_pref_mapping_test.md, sections PrefTestCase and SimplePolicyPrefMappingTest
26
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
27
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
28
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
@@ -30,8 +26,6 @@ Files:
30
  - `components/policy/test/data/pref_mapping/DownloadDirectory.json`: Chrome download location setting: DownloadDirectory policy mapped to download.default_directory and download.prompt_for_download.
31
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
32
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
33
- Reference result: expected_policy_and_pref_values
34
- Reference source: components/policy/test/data/pref_mapping/DownloadDirectory.json (whole file): desktop case: DownloadDirectory "${user_home}/test-downloads" gives download.default_directory "${user_home}/test-downloads" and download.prompt_for_download false (check_for_recommended false); a second ChromeOS-only case maps "${google_drive}/downloads" to gdata.disabled false; ${user_home} and ${google_drive} are Chrome policy path variables that Chrome expands; field meanings (policies, prefs, value, default_value, check_for_recommended, simple_policy_pref_mapping_test) in docs/enterprise/policy_pref_mapping_test.md, sections PrefTestCase and SimplePolicyPrefMappingTest
35
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
36
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
37
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
@@ -39,8 +33,6 @@ Files:
39
  - `components/policy/test/data/pref_mapping/RestoreOnStartup.json`: Chrome on-startup setting: RestoreOnStartup policy values 1 to 6 mapped to session.restore_on_startup.
40
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
41
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
42
- Reference result: expected_policy_and_pref_values
43
- Reference source: components/policy/test/data/pref_mapping/RestoreOnStartup.json (whole file): six test cases: RestoreOnStartup 1, 2, 3, 4, 5 and 6 each give session.restore_on_startup the same value; can_be_recommended true; os win, linux, mac, chromeos, fuchsia; field meanings (policies, prefs, value, default_value, check_for_recommended, simple_policy_pref_mapping_test) in docs/enterprise/policy_pref_mapping_test.md, sections PrefTestCase and SimplePolicyPrefMappingTest
44
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
45
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
46
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
@@ -48,8 +40,6 @@ Files:
48
  - `components/policy/test/data/pref_mapping/ProxySettings.json`: Chrome proxy setting: ProxySettings policy with ProxyMode direct mapped to the proxy pref.
49
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
50
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
51
- Reference result: expected_policy_and_pref_values
52
- Reference source: components/policy/test/data/pref_mapping/ProxySettings.json (whole file): one test case: ProxySettings {"ProxyMode": "direct"} gives the proxy pref the value {"mode": "direct"}; os linux, win, android, chromeos, mac, fuchsia; field meanings (policies, prefs, value, default_value, check_for_recommended, simple_policy_pref_mapping_test) in docs/enterprise/policy_pref_mapping_test.md, sections PrefTestCase and SimplePolicyPrefMappingTest
53
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
54
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
55
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
@@ -57,8 +47,6 @@ Files:
57
  - `components/policy/test/data/pref_mapping/IncognitoModeAvailability.json`: Chrome Incognito availability setting: IncognitoModeAvailability policy mapped to incognito.mode_availability.
58
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
59
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
60
- Reference result: expected_policy_and_pref_values
61
- Reference source: components/policy/test/data/pref_mapping/IncognitoModeAvailability.json (whole file): simple_policy_pref_mapping_test: pref incognito.mode_availability, default_value 0 when the policy is unset, values_to_test 0, 1 and 2 (the doc companion says each tested value must appear in the pref unchanged); os win, linux, mac, chromeos, android, fuchsia; field meanings (policies, prefs, value, default_value, check_for_recommended, simple_policy_pref_mapping_test) in docs/enterprise/policy_pref_mapping_test.md, sections PrefTestCase and SimplePolicyPrefMappingTest
62
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
63
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
64
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
@@ -66,8 +54,6 @@ Files:
66
  - `components/policy/test/data/pref_mapping/DefaultCookiesSetting.json`: Chrome cookie setting: DefaultCookiesSetting policy mapped to profile.managed_default_content_settings.cookies.
67
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
68
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
69
- Reference result: expected_policy_and_pref_values
70
- Reference source: components/policy/test/data/pref_mapping/DefaultCookiesSetting.json (whole file): simple_policy_pref_mapping_test: pref profile.managed_default_content_settings.cookies, default_value 0 when the policy is unset, values_to_test 1, 2 and 4; os win, linux, mac, chromeos, android, fuchsia; field meanings (policies, prefs, value, default_value, check_for_recommended, simple_policy_pref_mapping_test) in docs/enterprise/policy_pref_mapping_test.md, sections PrefTestCase and SimplePolicyPrefMappingTest
71
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
72
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
73
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
@@ -75,8 +61,6 @@ Files:
75
  - `components/policy/test/data/pref_mapping/HomepageLocation.json`: Chrome home page setting: HomepageLocation policy mapped to the homepage pref.
76
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and simple_policy_pref_mapping_test {pref_name, default_value, values_to_test})
77
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
78
- Reference result: expected_policy_and_pref_values
79
- Reference source: components/policy/test/data/pref_mapping/HomepageLocation.json (whole file): simple_policy_pref_mapping_test with pref_name "homepage", default_value "" and values_to_test ["http://chromium.org/"]; can_be_recommended true; os win, linux, mac, chromeos, android, fuchsia; field meanings in docs/enterprise/policy_pref_mapping_test.md, section SimplePolicyPrefMappingTest
80
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that the policy key and value equal the tested value (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that the pref path equals the tested value, or keeps default_value when the policy is unset (JSON equality). Confirming that the policy really yields the pref needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
81
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
82
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
@@ -84,8 +68,6 @@ Files:
84
  - `components/policy/test/data/pref_mapping/ShowHomeButton.json`: Chrome home button setting: ShowHomeButton policy mapped to the browser.show_home_button pref.
85
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and simple_policy_pref_mapping_test {pref_name, default_value, values_to_test})
86
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
87
- Reference result: expected_policy_and_pref_values
88
- Reference source: components/policy/test/data/pref_mapping/ShowHomeButton.json (whole file): simple_policy_pref_mapping_test with pref_name "browser.show_home_button", default_value false and values_to_test [true, false]; can_be_recommended true; os android, win, linux, mac, chromeos, fuchsia; field meanings in docs/enterprise/policy_pref_mapping_test.md, section SimplePolicyPrefMappingTest
89
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that the policy key and value equal the tested value (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that the pref path equals the tested value, or keeps default_value when the policy is unset (JSON equality). Confirming that the policy really yields the pref needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
90
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
91
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 8 of 8 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
12
 
13
  **Licence (SPDX):** BSD-3-Clause
 
19
  - `components/policy/test/data/pref_mapping/BookmarkBarEnabled.json`: Chrome bookmarks bar setting: BookmarkBarEnabled policy mapped to the bookmark_bar prefs (3 cases).
20
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
21
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
 
 
22
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
23
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
24
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
 
26
  - `components/policy/test/data/pref_mapping/DownloadDirectory.json`: Chrome download location setting: DownloadDirectory policy mapped to download.default_directory and download.prompt_for_download.
27
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
28
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
 
 
29
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
30
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
31
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
 
33
  - `components/policy/test/data/pref_mapping/RestoreOnStartup.json`: Chrome on-startup setting: RestoreOnStartup policy values 1 to 6 mapped to session.restore_on_startup.
34
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
35
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
 
 
36
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
37
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
38
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
 
40
  - `components/policy/test/data/pref_mapping/ProxySettings.json`: Chrome proxy setting: ProxySettings policy with ProxyMode direct mapped to the proxy pref.
41
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
42
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
 
 
43
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
44
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
45
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
 
47
  - `components/policy/test/data/pref_mapping/IncognitoModeAvailability.json`: Chrome Incognito availability setting: IncognitoModeAvailability policy mapped to incognito.mode_availability.
48
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
49
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
 
 
50
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
51
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
52
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
 
54
  - `components/policy/test/data/pref_mapping/DefaultCookiesSetting.json`: Chrome cookie setting: DefaultCookiesSetting policy mapped to profile.managed_default_content_settings.cookies.
55
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and either policy_pref_mapping_tests or simple_policy_pref_mapping_test)
56
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
 
 
57
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that its policy keys and values equal the "policies" object of the chosen test case (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that each pref path listed under "prefs" equals its "value" (JSON equality; "default_value" entries mean the pref keeps its default). Confirming that a policy really yields the listed prefs needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
58
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
59
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
 
61
  - `components/policy/test/data/pref_mapping/HomepageLocation.json`: Chrome home page setting: HomepageLocation policy mapped to the homepage pref.
62
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and simple_policy_pref_mapping_test {pref_name, default_value, values_to_test})
63
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
 
 
64
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that the policy key and value equal the tested value (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that the pref path equals the tested value, or keeps default_value when the policy is unset (JSON equality). Confirming that the policy really yields the pref needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
65
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
66
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
 
68
  - `components/policy/test/data/pref_mapping/ShowHomeButton.json`: Chrome home button setting: ShowHomeButton policy mapped to the browser.show_home_button pref.
69
  Format: Chromium policy_pref_mapping JSON (list of PolicyTestCase objects with os, can_be_recommended and simple_policy_pref_mapping_test {pref_name, default_value, values_to_test})
70
  Origin: test data of Chromium's PolicyPrefsTest, which applies each policy and checks the resulting pref values (tool_fixture)
 
 
71
  Check: Final configuration state, no screenshot needed. Policy direction: parse the managed-policy JSON file the agent wrote and check that the policy key and value equal the tested value (JSON equality). User-setting direction: read the profile Preferences JSON after the browser is closed and check that the pref path equals the tested value, or keeps default_value when the policy is unset (JSON equality). Confirming that the policy really yields the pref needs Chromium at the pinned commit (PolicyPrefsTest); the file-level judge needs only a JSON parser
72
  Tools: write the policy as JSON into the managed-policy directory (/etc/opt/chrome/policies/managed/*.json for Chrome or /etc/chromium/policies/managed/*.json for Chromium on Linux, a plist via defaults on macOS, the registry via reg add on Windows) with a text editor, jq or a script; the user-setting direction edits the pref in the profile's Preferences JSON with jq while the browser is closed; Chrome or Chromium can then be started headless (--headless) to load them
73
  Companions (not included): docs/enterprise/policy_pref_mapping_test.md; LICENSE
campaign_files/2026-10-gt-browser-tests-2/coach-core-merge-tests/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: expected_json. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** a script (node or jq) that merges the two JSON advice files; the coach-core CLI or sitespeed.io run headless with --coach produces the merged result
14
 
15
  **Licence (SPDX):** MIT
@@ -21,8 +19,6 @@ Files:
21
  - `test/merge/mergeTest.js`: Merge the web performance advice of a DOM run and a HAR run of the Coach: union of the advice lists, recomputed category and total scores, and HAR advice overriding DOM advice of the same id.
22
  Format: ava test file plus three JSON advice result files (DOM result, HAR result, HAR result with an override)
23
  Origin: unit tests and fixtures of the coach-core advice merge (tool_fixture); the library source is not fetched, not bundled and not run
24
- Reference result: expected_json
25
- Reference source: test/merge/mergeTest.js: number of performance advice = DOM count + HAR count (lines 20 to 23); performance score 99 (line 28); total score 98 (line 33); altImages title altImages from HAR after override (lines 38 to 41)
26
  Check: Final file state: the agent writes the merged advice JSON; compare the advice count, the two scores and the overridden title with the expected values (ava t.is); the scoring rule (mean of advice scores weighted as in coach-core lib/merge.js) must be stated in the task text.
27
  Tools: a script (node or jq) that merges the two JSON advice files; the coach-core CLI or sitespeed.io run headless with --coach produces the merged result
28
  Companions (not included): LICENSE; test/merge/files/domResult.json; test/merge/files/harResult.json; test/merge/files/harResultOverride.json
 
8
 
9
  **Self-contained:** Not self-contained: each file is used together with other files from the same upstream. 1 of 1 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** a script (node or jq) that merges the two JSON advice files; the coach-core CLI or sitespeed.io run headless with --coach produces the merged result
12
 
13
  **Licence (SPDX):** MIT
 
19
  - `test/merge/mergeTest.js`: Merge the web performance advice of a DOM run and a HAR run of the Coach: union of the advice lists, recomputed category and total scores, and HAR advice overriding DOM advice of the same id.
20
  Format: ava test file plus three JSON advice result files (DOM result, HAR result, HAR result with an override)
21
  Origin: unit tests and fixtures of the coach-core advice merge (tool_fixture); the library source is not fetched, not bundled and not run
 
 
22
  Check: Final file state: the agent writes the merged advice JSON; compare the advice count, the two scores and the overridden title with the expected values (ava t.is); the scoring rule (mean of advice scores weighted as in coach-core lib/merge.js) must be stated in the task text.
23
  Tools: a script (node or jq) that merges the two JSON advice files; the coach-core CLI or sitespeed.io run headless with --coach produces the merged result
24
  Companions (not included): LICENSE; test/merge/files/domResult.json; test/merge/files/harResult.json; test/merge/files/harResultOverride.json
campaign_files/2026-10-hw-eda-tests-2/bosl2-the-belfry-openscad-library-v2/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 3 of 3 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: cad_geometry_assert. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** openscad -o out.echo runs the extracted test script headless and the run passes when it exits 0 with no ECHO or WARNING (tests/run-tests.sh does this for each [[test]] block)
14
 
15
  **Licence (SPDX):** BSD-2-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `tests/test_screw_drive.scadtest`: BOSL2 screw drive tests: Torx and other drive recess diameters and depths computed from the standard tables.
22
  Format: OpenSCAD test file in TOML form: [[test]] blocks whose script includes BOSL2 std.scad and calls assert_approx, assert_equal and assert
23
  Origin: BOSL2 library regression test written by BOSL2 maintainers; expected numbers are closed-form geometry or standard dimension tables
24
- Reference result: cad_geometry_assert
25
- Reference source: upstream expected values written in the test file itself: line 8 `assert_approx(torx_diam(10), 2.80);` through line 12 `assert_approx(torx_diam(30), 5.60);` and the depth checks at lines 25-30 (e.g. `assert_approx(torx_depth(10), 1.142);`).
26
  Check: Task: given each test's BOSL2 call with the expected values hidden, compute the returned dimensions, points or transforms (by OpenSCAD or in closed form). Judge: the test script passes by the BOSL2 rule (OpenSCAD exit 0 with no ECHO or WARNING), assert_approx by abs(got - want) <= 1e-9, assert_equal exactly.
27
  Tools: openscad -o out.echo runs the extracted test script headless and the run passes when it exits 0 with no ECHO or WARNING (tests/run-tests.sh does this for each [[test]] block)
28
  Companions (not included): LICENSE
@@ -30,8 +26,6 @@ Files:
30
  - `tests/test_distributors.scadtest`: BOSL2 distributor tests: copy placements along lines, grids, arcs and spheres and check the computed translation and rotation positions.
31
  Format: OpenSCAD test file in TOML form: [[test]] blocks whose script includes BOSL2 std.scad and calls assert_approx, assert_equal and assert
32
  Origin: BOSL2 library regression test written by BOSL2 maintainers; expected numbers are closed-form geometry or standard dimension tables
33
- Reference result: cad_geometry_assert
34
- Reference source: upstream expected values written in the test file itself: assert_approx and assert_equal on the point lists and transform matrices returned by line_copies, grid_copies, arc_copies and the other distributors; closed-form.
35
  Check: Task: given each test's BOSL2 call with the expected values hidden, compute the returned dimensions, points or transforms (by OpenSCAD or in closed form). Judge: the test script passes by the BOSL2 rule (OpenSCAD exit 0 with no ECHO or WARNING), assert_approx by abs(got - want) <= 1e-9, assert_equal exactly.
36
  Tools: openscad -o out.echo runs the extracted test script headless and the run passes when it exits 0 with no ECHO or WARNING (tests/run-tests.sh does this for each [[test]] block)
37
  Companions (not included): LICENSE
@@ -39,8 +33,6 @@ Files:
39
  - `tests/test_drawing.scadtest`: BOSL2 drawing tests: turtle paths, arcs and helices returned as point lists and checked against computed coordinates.
40
  Format: OpenSCAD test file in TOML form: [[test]] blocks whose script includes BOSL2 std.scad and calls assert_approx, assert_equal and assert
41
  Origin: BOSL2 library regression test written by BOSL2 maintainers; expected numbers are closed-form geometry or standard dimension tables
42
- Reference result: cad_geometry_assert
43
- Reference source: upstream expected values written in the test file itself: 17 assert_approx calls on arc and turtle point lists, e.g.
44
  Check: Task: given each test's BOSL2 call with the expected values hidden, compute the returned dimensions, points or transforms (by OpenSCAD or in closed form). Judge: the test script passes by the BOSL2 rule (OpenSCAD exit 0 with no ECHO or WARNING), assert_approx by abs(got - want) <= 1e-9, assert_equal exactly.
45
  Tools: openscad -o out.echo runs the extracted test script headless and the run passes when it exits 0 with no ECHO or WARNING (tests/run-tests.sh does this for each [[test]] block)
46
  Companions (not included): LICENSE
 
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 3 of 3 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** openscad -o out.echo runs the extracted test script headless and the run passes when it exits 0 with no ECHO or WARNING (tests/run-tests.sh does this for each [[test]] block)
12
 
13
  **Licence (SPDX):** BSD-2-Clause
 
19
  - `tests/test_screw_drive.scadtest`: BOSL2 screw drive tests: Torx and other drive recess diameters and depths computed from the standard tables.
20
  Format: OpenSCAD test file in TOML form: [[test]] blocks whose script includes BOSL2 std.scad and calls assert_approx, assert_equal and assert
21
  Origin: BOSL2 library regression test written by BOSL2 maintainers; expected numbers are closed-form geometry or standard dimension tables
 
 
22
  Check: Task: given each test's BOSL2 call with the expected values hidden, compute the returned dimensions, points or transforms (by OpenSCAD or in closed form). Judge: the test script passes by the BOSL2 rule (OpenSCAD exit 0 with no ECHO or WARNING), assert_approx by abs(got - want) <= 1e-9, assert_equal exactly.
23
  Tools: openscad -o out.echo runs the extracted test script headless and the run passes when it exits 0 with no ECHO or WARNING (tests/run-tests.sh does this for each [[test]] block)
24
  Companions (not included): LICENSE
 
26
  - `tests/test_distributors.scadtest`: BOSL2 distributor tests: copy placements along lines, grids, arcs and spheres and check the computed translation and rotation positions.
27
  Format: OpenSCAD test file in TOML form: [[test]] blocks whose script includes BOSL2 std.scad and calls assert_approx, assert_equal and assert
28
  Origin: BOSL2 library regression test written by BOSL2 maintainers; expected numbers are closed-form geometry or standard dimension tables
 
 
29
  Check: Task: given each test's BOSL2 call with the expected values hidden, compute the returned dimensions, points or transforms (by OpenSCAD or in closed form). Judge: the test script passes by the BOSL2 rule (OpenSCAD exit 0 with no ECHO or WARNING), assert_approx by abs(got - want) <= 1e-9, assert_equal exactly.
30
  Tools: openscad -o out.echo runs the extracted test script headless and the run passes when it exits 0 with no ECHO or WARNING (tests/run-tests.sh does this for each [[test]] block)
31
  Companions (not included): LICENSE
 
33
  - `tests/test_drawing.scadtest`: BOSL2 drawing tests: turtle paths, arcs and helices returned as point lists and checked against computed coordinates.
34
  Format: OpenSCAD test file in TOML form: [[test]] blocks whose script includes BOSL2 std.scad and calls assert_approx, assert_equal and assert
35
  Origin: BOSL2 library regression test written by BOSL2 maintainers; expected numbers are closed-form geometry or standard dimension tables
 
 
36
  Check: Task: given each test's BOSL2 call with the expected values hidden, compute the returned dimensions, points or transforms (by OpenSCAD or in closed form). Judge: the test script passes by the BOSL2 rule (OpenSCAD exit 0 with no ECHO or WARNING), assert_approx by abs(got - want) <= 1e-9, assert_equal exactly.
37
  Tools: openscad -o out.echo runs the extracted test script headless and the run passes when it exits 0 with no ECHO or WARNING (tests/run-tests.sh does this for each [[test]] block)
38
  Companions (not included): LICENSE
campaign_files/2026-10-hw-eda-tests-2/buserror-simavr-tests-github/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 6 of 6 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: emulator_stdout_exact; emulator_cycle_window. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Tools are named in the check written per file in the notes.
14
 
15
  **Licence (SPDX):** GPL-3.0-or-later AND MIT
@@ -21,48 +19,36 @@ Files:
21
  - `tests/atmega88_example.c`: Simavr: EEPROM write and read back of two 32-bit words on an emulated ATmega88, printed over the UART; produce the exact UART output.
22
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
23
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
24
- Reference result: emulator_stdout_exact
25
- Reference source: tests/test_atmega88_example.c lines 6-8 (two lines: "Read from eeprom 0xdeadbeef, should be 0xdeadbeef\r\n" and "Read from eeprom 0xcafef00d, should be 0xcafef00d\r\n"); the expected string is written by the project in the test program, which compares it with the UART text of the emulated run
26
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega88_example.c against the simavr library, run it; it loads atmega88_example.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
27
  Companions (not included): tests/test_atmega88_example.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
28
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
29
  - `tests/atmega48_watchdog_test.c`: Simavr: watchdog timer on an emulated ATmega48: arm the watchdog, let it fire its interrupt, then let it reset the MCU twice, reporting each stage over the UART; produce the exact UART output.
30
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
31
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
32
- Reference result: emulator_stdout_exact
33
- Reference source: tests/test_atmega48_watchdog_test.c lines 6-11 (five lines from "Watchdog is active.\r\n" to "Watchdog reset MCU again.\r\n"); the expected string is written by the project in the test program, which compares it with the UART text of the emulated run
34
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega48_watchdog_test.c against the simavr library, run it; it loads atmega48_watchdog_test.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
35
  Companions (not included): tests/test_atmega48_watchdog_test.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
36
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
37
  - `tests/atmega644_adc_test.c`: Simavr: ADC on an emulated ATmega644: read 8 channels with interrupts, then the 1.1 V band-gap in polling mode with AREF at 2200 mV, printing the converted values; produce the exact UART output.
38
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
39
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
40
- Reference result: emulator_stdout_exact
41
- Reference source: tests/test_atmega644_adc_test.c lines 4-8 (four lines, including "Read ADC value 0155 = 1098 mvolts, ought to be 1098\r\n"); the expected string is written by the project in the test program, which compares it with the UART text of the emulated run
42
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega644_adc_test.c against the simavr library, run it; it loads atmega644_adc_test.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
43
  Companions (not included): tests/test_atmega644_adc_test.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
44
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
45
  - `tests/atmega88_uart_echo.c`: Simavr: UART loopback on an emulated ATmega88: send a line, receive it back through the receive interrupt and print what was received; produce the exact UART output.
46
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
47
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
48
- Reference result: emulator_stdout_exact
49
- Reference source: tests/test_atmega88_uart_echo.c lines 6-8 (two lines: "Hey there, this should be received back\r\n" and "Received: Hey there, this should be received back\r\r\n"); the expected string is written by the project in the test program, which compares it with the UART text of the emulated run
50
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega88_uart_echo.c against the simavr library, run it; it loads atmega88_uart_echo.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
51
  Companions (not included): tests/test_atmega88_uart_echo.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
52
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
53
  - `tests/atmega644p_signature_read.c`: Simavr: read the device signature bytes of an emulated ATmega644P with the boot loader signature-read instruction and print them; produce the exact UART output.
54
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
55
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
56
- Reference result: emulator_stdout_exact
57
- Reference source: tests/test_atmega644p_signature_read.c line 4 ("Signature is 0x1e 96 09\r\n"); the expected string is written by the project in the test program, which compares it with the UART text of the emulated run
58
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega644p_signature_read.c against the simavr library, run it; it loads atmega644p_signature_read.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
59
  Companions (not included): tests/test_atmega644p_signature_read.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
60
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
61
  - `tests/atmega88_timer16.c`: Simavr: 16-bit Timer1 on an emulated ATmega88 drives a fixed sequence and then stops the simulation; the run must end by itself within a cycle window.
62
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
63
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
64
- Reference result: emulator_cycle_window
65
- Reference source: tests/test_atmega88_timer16.c lines 7-17: the run must end by special deinit (not by the cycle timer) and the final cycle count must lie in [12500000, 12500300]; the bounds are written by the project
66
  Check: Task: given the firmware (or its specification), predict or reproduce its timing so that it ends the simulation within the stated cycle window. Judge: build the firmware with avr-gcc and the test program against the simavr library, run it; it fails if the cycle timer fires first or if the cycle count at the end is outside the window.
67
  Companions (not included): tests/test_atmega88_timer16.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
68
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
 
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 6 of 6 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** Tools are named in the check written per file in the notes.
12
 
13
  **Licence (SPDX):** GPL-3.0-or-later AND MIT
 
19
  - `tests/atmega88_example.c`: Simavr: EEPROM write and read back of two 32-bit words on an emulated ATmega88, printed over the UART; produce the exact UART output.
20
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
21
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
 
 
22
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega88_example.c against the simavr library, run it; it loads atmega88_example.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
23
  Companions (not included): tests/test_atmega88_example.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
24
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
25
  - `tests/atmega48_watchdog_test.c`: Simavr: watchdog timer on an emulated ATmega48: arm the watchdog, let it fire its interrupt, then let it reset the MCU twice, reporting each stage over the UART; produce the exact UART output.
26
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
27
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
 
 
28
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega48_watchdog_test.c against the simavr library, run it; it loads atmega48_watchdog_test.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
29
  Companions (not included): tests/test_atmega48_watchdog_test.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
30
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
31
  - `tests/atmega644_adc_test.c`: Simavr: ADC on an emulated ATmega644: read 8 channels with interrupts, then the 1.1 V band-gap in polling mode with AREF at 2200 mV, printing the converted values; produce the exact UART output.
32
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
33
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
 
 
34
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega644_adc_test.c against the simavr library, run it; it loads atmega644_adc_test.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
35
  Companions (not included): tests/test_atmega644_adc_test.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
36
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
37
  - `tests/atmega88_uart_echo.c`: Simavr: UART loopback on an emulated ATmega88: send a line, receive it back through the receive interrupt and print what was received; produce the exact UART output.
38
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
39
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
 
 
40
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega88_uart_echo.c against the simavr library, run it; it loads atmega88_uart_echo.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
41
  Companions (not included): tests/test_atmega88_uart_echo.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
42
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
43
  - `tests/atmega644p_signature_read.c`: Simavr: read the device signature bytes of an emulated ATmega644P with the boot loader signature-read instruction and print them; produce the exact UART output.
44
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
45
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
 
 
46
  Check: Task: given the firmware (or its specification) and the MCU, produce the exact UART text the firmware prints, or write firmware whose run matches it. Judge: build the firmware with avr-gcc, build tests/test_atmega644p_signature_read.c against the simavr library, run it; it loads atmega644p_signature_read.axf into the simulated MCU, captures UART0 and fails unless the simulation ends by itself within the budget and the captured text equals the expected string exactly.
47
  Companions (not included): tests/test_atmega644p_signature_read.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
48
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
49
  - `tests/atmega88_timer16.c`: Simavr: 16-bit Timer1 on an emulated ATmega88 drives a fixed sequence and then stops the simulation; the run must end by itself within a cycle window.
50
  Format: AVR C firmware file (with simavr section macros) compiled by avr-gcc to .axf, plus a host C test program linked against simavr; reference output is UART text in the test program
51
  Origin: simavr's own regression tests: an AVR firmware file plus a host test program that runs it in the emulator and checks the UART text or cycle count; written by the project
 
 
52
  Check: Task: given the firmware (or its specification), predict or reproduce its timing so that it ends the simulation within the stated cycle window. Judge: build the firmware with avr-gcc and the test program against the simavr library, run it; it fails if the cycle timer fires first or if the cycle count at the end is outside the window.
53
  Companions (not included): tests/test_atmega88_timer16.c; tests/tests.c; tests/tests.h; tests/Makefile; simavr/sim/avr/avr_mcu_section.h; COPYING
54
  Dependencies: simavr library and headers built from buserror/simavr at commit d6aed536 (GPL-3.0-or-later; the emulator itself, linked by the test program; not packed); avr-gcc and avr-libc to build the firmware .axf; host C compiler and libelf for the test program
campaign_files/2026-10-hw-eda-tests-2/cadquery-plugins-community-plugin-collection-for-cadquery/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 5 of 5 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: cad_geometry_assert. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** pytest runs the module from the repository root (pytest tests/test_gear_generator.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API; pytest runs the module from the repository root (pytest tests/test_heatserts.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API; pytest runs the module from the repository root (pytest tests/test_fragment.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API; pytest runs the module from the repository root (pytest tests/test_teardrop.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API; pytest runs the module from the repository root (pytest tests/test_apply_to_each_face.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
14
 
15
  **Licence (SPDX):** Apache-2.0
@@ -21,8 +19,6 @@ Files:
21
  - `tests/test_gear_generator.py`: Cadquery-plugins gear generator tests: build spur, helical, bevel and rack gears from module, tooth count and width and check their volumes.
22
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
23
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
24
- Reference result: cad_geometry_assert
25
- Reference source: upstream expected values written in the test file itself: line 29 volume 2505.7966225157024; lines 53-54 pinion 483.3443067836134 and gear 1126.5921554741049; lines 68-69; line 86 10113 within 70; line 91 10033.574314576623. Recorded kernel results.
26
  Check: Task: given each test's gear parameters (module, teeth, width, pressure angle, helix angle) with the expected volumes hidden, build the gear and report its volume. Judge: assertAlmostEqual with places=1 passes when round(got - want, 1) == 0 (CPython unittest/case.py line 876); line 86 passes when 10113 - 70 < volume < 10113 + 70. Note line 32 re-checks gear1 instead of gear2 upstream; the expected value there is for gear2 and is judged against gear2.
27
  Tools: pytest runs the module from the repository root (pytest tests/test_gear_generator.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
28
  Companions (not included): plugins/gear_generator/gear_generator/__init__.py; plugins/gear_generator/gear_generator/helpers.py; plugins/gear_generator/gear_generator/main.py; LICENSE
@@ -30,8 +26,6 @@ Files:
30
  - `tests/test_heatserts.py`: Cadquery-plugins heat-set insert hole tests: cut insert holes of standard sizes into a box and check removed volume, hole radii and chamfer depth.
31
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
32
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
33
- Reference result: cad_geometry_assert
34
- Reference source: expected values computed in the test file from the plugin's size table: line 33 removed volume > heatsert volume - 0.1; lines 44-46 extra bolt-clearance volume; line 72 hole radii within 0.001; lines 77-78 chamfer depth within 0.01.
35
  Check: Task: given an insert size and options, cut the hole and report the removed volume, hole radii and chamfer depth. Judge: removed volume must exceed the closed-form insert volume minus 0.1 (line 33) and the clearance extension minus 0.1 (line 46); each expected radius must match some circle radius within 0.001 (line 72); chamfer depth within 0.01 (line 78).
36
  Tools: pytest runs the module from the repository root (pytest tests/test_heatserts.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
37
  Companions (not included): plugins/heatserts/__init__.py; plugins/heatserts/heatserts.py; LICENSE
@@ -39,8 +33,6 @@ Files:
39
  - `tests/test_fragment.py`: Cadquery-plugins fragment tests: split overlapping solids into non-overlapping fragments and check the fragment volumes and counts.
40
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
41
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
42
- Reference result: cad_geometry_assert
43
- Reference source: upstream expected values written in the test file itself: lines 13-15 fragment volumes 7875.0, 125.0 and 875.0; lines 26-62 fragment counts. Closed-form box volumes.
44
  Check: Task: given the plate and the overlapping cutters, fragment them and report each fragment's volume. Judge: each volume within 1e-6 of 7875, 125 and 875 (lines 13-15). The upstream check is one-sided, (abs(V) - 7875.0) < 1e-6, so it only rejects values above; Fragment counts must equal the expected counts (lines 26, 37, 48, 62).
45
  Tools: pytest runs the module from the repository root (pytest tests/test_fragment.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
46
  Companions (not included): plugins/fragment/__init__.py; plugins/fragment/fragment.py; LICENSE
@@ -48,8 +40,6 @@ Files:
48
  - `tests/test_teardrop.py`: Cadquery-plugins teardrop hole tests: cut 3D-printable teardrop holes and check the computed vertex positions of the profile.
49
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
50
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
51
- Reference result: cad_geometry_assert
52
- Reference source: upstream expected values written in the test file itself: line 23 left vertex x = -5.65685 (4 * sqrt(2)); line 27 symmetry; line 32 top vertex z = 4.1. Closed-form.
53
  Check: Task: given the hole radius and clip value, build the teardrop profile and report its extreme vertex coordinates. Judge: pytest.approx with rel=1e-3 for x = -5.65685 (line 23) and z = 4.1 (line 32); abs(x1 + x2) < 0.001 for symmetry (line 27).
54
  Tools: pytest runs the module from the repository root (pytest tests/test_teardrop.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
55
  Companions (not included): plugins/teardrop/__init__.py; plugins/teardrop/teardrop.py; LICENSE
@@ -57,8 +47,6 @@ Files:
57
  - `tests/test_apply_to_each_face.py`: Cadquery-plugins apply-to-each-face tests: build features on every face of a solid with computed local frames and check that the frame axes and positions match the faces.
58
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
59
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
60
- Reference result: cad_geometry_assert
61
- Reference source: expected values computed in the test file from the faces themselves: lines 40-47 and 85-86 projections of frame axes on face normals are zero; line 99 offset vectors have zero length; lines 122 and 142 face counts 36.
62
  Check: Task: given the solid and the per-face operation, build the result and report the frame projections, offsets and face count. Judge: every projection and offset length must be below TOLERANCE (lines 40-47, 85-86, 99); face counts must equal 3 * 6 * 2 (lines 122, 142).
63
  Tools: pytest runs the module from the repository root (pytest tests/test_apply_to_each_face.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
64
  Companions (not included): plugins/apply_to_each_face/__init__.py; plugins/apply_to_each_face/apply_to_each_face.py; LICENSE
 
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 5 of 5 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** pytest runs the module from the repository root (pytest tests/test_gear_generator.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API; pytest runs the module from the repository root (pytest tests/test_heatserts.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API; pytest runs the module from the repository root (pytest tests/test_fragment.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API; pytest runs the module from the repository root (pytest tests/test_teardrop.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API; pytest runs the module from the repository root (pytest tests/test_apply_to_each_face.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
12
 
13
  **Licence (SPDX):** Apache-2.0
 
19
  - `tests/test_gear_generator.py`: Cadquery-plugins gear generator tests: build spur, helical, bevel and rack gears from module, tooth count and width and check their volumes.
20
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
21
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
 
 
22
  Check: Task: given each test's gear parameters (module, teeth, width, pressure angle, helix angle) with the expected volumes hidden, build the gear and report its volume. Judge: assertAlmostEqual with places=1 passes when round(got - want, 1) == 0 (CPython unittest/case.py line 876); line 86 passes when 10113 - 70 < volume < 10113 + 70. Note line 32 re-checks gear1 instead of gear2 upstream; the expected value there is for gear2 and is judged against gear2.
23
  Tools: pytest runs the module from the repository root (pytest tests/test_gear_generator.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
24
  Companions (not included): plugins/gear_generator/gear_generator/__init__.py; plugins/gear_generator/gear_generator/helpers.py; plugins/gear_generator/gear_generator/main.py; LICENSE
 
26
  - `tests/test_heatserts.py`: Cadquery-plugins heat-set insert hole tests: cut insert holes of standard sizes into a box and check removed volume, hole radii and chamfer depth.
27
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
28
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
 
 
29
  Check: Task: given an insert size and options, cut the hole and report the removed volume, hole radii and chamfer depth. Judge: removed volume must exceed the closed-form insert volume minus 0.1 (line 33) and the clearance extension minus 0.1 (line 46); each expected radius must match some circle radius within 0.001 (line 72); chamfer depth within 0.01 (line 78).
30
  Tools: pytest runs the module from the repository root (pytest tests/test_heatserts.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
31
  Companions (not included): plugins/heatserts/__init__.py; plugins/heatserts/heatserts.py; LICENSE
 
33
  - `tests/test_fragment.py`: Cadquery-plugins fragment tests: split overlapping solids into non-overlapping fragments and check the fragment volumes and counts.
34
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
35
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
 
 
36
  Check: Task: given the plate and the overlapping cutters, fragment them and report each fragment's volume. Judge: each volume within 1e-6 of 7875, 125 and 875 (lines 13-15). The upstream check is one-sided, (abs(V) - 7875.0) < 1e-6, so it only rejects values above; Fragment counts must equal the expected counts (lines 26, 37, 48, 62).
37
  Tools: pytest runs the module from the repository root (pytest tests/test_fragment.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
38
  Companions (not included): plugins/fragment/__init__.py; plugins/fragment/fragment.py; LICENSE
 
40
  - `tests/test_teardrop.py`: Cadquery-plugins teardrop hole tests: cut 3D-printable teardrop holes and check the computed vertex positions of the profile.
41
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
42
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
 
 
43
  Check: Task: given the hole radius and clip value, build the teardrop profile and report its extreme vertex coordinates. Judge: pytest.approx with rel=1e-3 for x = -5.65685 (line 23) and z = 4.1 (line 32); abs(x1 + x2) < 0.001 for symmetry (line 27).
44
  Tools: pytest runs the module from the repository root (pytest tests/test_teardrop.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
45
  Companions (not included): plugins/teardrop/__init__.py; plugins/teardrop/teardrop.py; LICENSE
 
47
  - `tests/test_apply_to_each_face.py`: Cadquery-plugins apply-to-each-face tests: build features on every face of a solid with computed local frames and check that the frame axes and positions match the faces.
48
  Format: Python test module (pytest or unittest) importing one plugin package from the plugins/ directory, run against CadQuery
49
  Origin: test written by the plugin author to check a community CadQuery plugin; expected numbers are closed-form geometry or recorded kernel results
 
 
50
  Check: Task: given the solid and the per-face operation, build the result and report the frame projections, offsets and face count. Judge: every projection and offset length must be below TOLERANCE (lines 40-47, 85-86, 99); face counts must equal 3 * 6 * 2 (lines 122, 142).
51
  Tools: pytest runs the module from the repository root (pytest tests/test_apply_to_each_face.py); volumes, areas and coordinates read from Volume(), Area() and Center() in the CadQuery Python API
52
  Companions (not included): plugins/apply_to_each_face/__init__.py; plugins/apply_to_each_face/apply_to_each_face.py; LICENSE
campaign_files/2026-10-hw-eda-tests-2/camotics-gcodetool-regression-tests-github-cauldrondevelopmentllc-camotics/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** 0 of 7 files are self-contained; for the others it is not stated. 7 of 7 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: normalised_gcode_with_messages. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/GlobalTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code. Other files: see the notes.
14
 
15
  **Licence (SPDX):** GPL-2.0-or-later AND BSD-3-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `tests/offsetTests/GlobalTest/data/stdin`: G92, G92.1, G92.2 and G92.3 global offset set, suspend, restore and clear with parameters #5210 to #5212 and #5420, #5421: run gcodetool and produce the normalised program with machine coordinates and the DEBUG values.
22
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
23
  Origin: a small upstream regression program written to exercise the interpreter feature named in the title
24
- Reference result: normalised_gcode_with_messages
25
- Reference source: tests/offsetTests/GlobalTest/expect/stdout: the gcodetool output recorded upstream (20 lines), with expect/stderr empty and expect/return 0; Y20.; G92.2 gives GO=0 with the offset kept; G92.3 gives GO=1 and X=-10; G92.1 zeroes #5211 and #5212; every MSG value and every output coordinate of expect/stdout agrees
26
  Check: Task: given the program, produce what gcodetool prints (the normalised program with N numbers, coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in). Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare it with expect/stdout line by line after stripping CR and LF at line ends; stderr must be empty and the return code 0
27
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/GlobalTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
28
  Companions (not included): LICENSE; COPYING; tests/test.json; tests/testHarness; tests/offsetTests/test.json; tests/offsetTests/GlobalTest/expect/stdout; tests/offsetTests/GlobalTest/expect/stderr; tests/offsetTests/GlobalTest/expect/return
@@ -30,8 +26,6 @@ Files:
30
  - `tests/offsetTests/CSMultiTest/data/stdin`: G10 L20 P2 sets coordinate system 2 so the current point gets the given value, repeated, then G55 is selected: run gcodetool and produce the normalised program in machine coordinates and the DEBUG values of #5241, #5242, #5220, #5420, #5421.
31
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
32
  Origin: a small upstream regression program written to exercise the interpreter feature named in the title
33
- Reference result: normalised_gcode_with_messages
34
- Reference source: tests/offsetTests/CSMultiTest/expect/stdout: the gcodetool output recorded upstream (12 lines), with expect/stderr empty and expect/return 0; after G55 the current X reads 30; G0 X10 in G55 becomes N10 G0 X-10.; at machine -10, G10 L20 P2 X40 gives -50; G0 X15 becomes N14 G0 X-35.; all agree with expect/stdout
35
  Check: Task: given the program, produce what gcodetool prints (the normalised program with N numbers, coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in). Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare it with expect/stdout line by line after stripping CR and LF at line ends; stderr must be empty and the return code 0
36
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/CSMultiTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
37
  Companions (not included): LICENSE; COPYING; tests/test.json; tests/testHarness; tests/offsetTests/test.json; tests/offsetTests/CSMultiTest/expect/stdout; tests/offsetTests/CSMultiTest/expect/stderr; tests/offsetTests/CSMultiTest/expect/return
@@ -39,8 +33,6 @@ Files:
39
  - `tests/offsetTests/G52Test/data/stdin`: G52 local offset applied twice: run gcodetool and produce the normalised program in machine coordinates and the DEBUG values of #5210 to #5212, #5420, #5421.
40
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
41
  Origin: a small upstream regression program written to exercise the interpreter feature named in the title
42
- Reference result: normalised_gcode_with_messages
43
- Reference source: tests/offsetTests/G52Test/expect/stdout: the gcodetool output recorded upstream (11 lines), with expect/stderr empty and expect/return 0; G0 X10 and X15 become X30. and X35.; G52 X40 Y40 at machine 35 gives X=-5; G0 X0 becomes X40.; all agree with expect/stdout
44
  Check: Task: given the program, produce what gcodetool prints (the normalised program with N numbers, coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in). Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare it with expect/stdout line by line after stripping CR and LF at line ends; stderr must be empty and the return code 0
45
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/G52Test/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
46
  Companions (not included): LICENSE; COPYING; tests/test.json; tests/testHarness; tests/offsetTests/test.json; tests/offsetTests/G52Test/expect/stdout; tests/offsetTests/G52Test/expect/stderr; tests/offsetTests/G52Test/expect/return
@@ -48,8 +40,6 @@ Files:
48
  - `tests/oCodeTests/ElseIfTest/data/stdin`: O-word subroutine with if, elseif and else branches called with 0, 1 and 2: run gcodetool and produce the normalised program (the MSG lines in branch order).
49
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
50
  Origin: a small upstream regression program written to exercise the interpreter feature named in the title
51
- Reference result: normalised_gcode_with_messages
52
- Reference source: tests/oCodeTests/ElseIfTest/expect/stdout: the gcodetool output recorded upstream (14 lines), with expect/stderr empty and expect/return 0; expect/stdout lists exactly sub, zero, endsub, sub, one, endsub, sub, other, endsub with separators between G21 and M2
53
  Check: Task: given the program, produce what gcodetool prints (the normalised program with N numbers, coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in). Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare it with expect/stdout line by line after stripping CR and LF at line ends; stderr must be empty and the return code 0
54
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/oCodeTests/ElseIfTest/data/stdin`; tests/testHarness runs the command of tests/oCodeTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
55
  Companions (not included): LICENSE; COPYING; tests/test.json; tests/testHarness; tests/oCodeTests/test.json; tests/oCodeTests/ElseIfTest/expect/stdout; tests/oCodeTests/ElseIfTest/expect/stderr; tests/oCodeTests/ElseIfTest/expect/return
@@ -57,8 +47,6 @@ Files:
57
  - `tests/offsetTests/G53Test/data/stdin`: G53 machine coordinate move under a G92 offset: run gcodetool and produce the normalised program in machine coordinates with the DEBUG values.
58
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
59
  Origin: a small upstream regression program for G53 with an active G92 offset
60
- Reference result: normalised_gcode_with_messages
61
- Reference source: tests/offsetTests/G53Test/expect/stdout: the gcodetool output recorded upstream (8 lines), expect/return 0, expect/stderr empty; MSG GO=1, X=20, Y=20; G0 X0 Y0 then G53 G0 X10 Y10 print as the machine coordinates a hand trace gives; expect/stderr empty, return 0
62
  Check: Task: given the program, produce what gcodetool prints (the normalised program with coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in), its stderr and return code. Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
63
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/G53Test/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
64
  Companions (not included): tests/offsetTests/G53Test/expect/return; tests/offsetTests/G53Test/expect/stderr; tests/offsetTests/G53Test/expect/stdout; tests/offsetTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
@@ -66,8 +54,6 @@ Files:
66
  - `tests/offsetTests/ToolOffsetTest/data/stdin`
67
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
68
  Origin: a small upstream regression program for G10 L1 and G43 tool length compensation
69
- Reference result: normalised_gcode_with_messages
70
- Reference source: tests/offsetTests/ToolOffsetTest/expect/stdout: the gcodetool output recorded upstream (7 lines), expect/return 0, expect/stderr one warning line; Expected stderr is not empty here: it holds the tool auto-create warning. G0 Z0 prints as Z1.5; expect/stderr is the one warning "Auto-creating missing tool 1" at stdin line 4; note: CAMotics reports #5422 as 10 after G43, where LinuxCNC would report the value less the tool length; the reference is CAMotics behaviour
71
  Check: Task: given the program, produce what gcodetool prints (the normalised program with coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in), its stderr and return code. Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
72
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/ToolOffsetTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
73
  Companions (not included): tests/offsetTests/ToolOffsetTest/expect/return; tests/offsetTests/ToolOffsetTest/expect/stderr; tests/offsetTests/ToolOffsetTest/expect/stdout; tests/offsetTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
@@ -75,8 +61,6 @@ Files:
75
  - `tests/offsetTests/ToolRelativeOffsetTest/data/stdin`
76
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
77
  Origin: a small upstream regression program for G10 L10 relative tool length setting
78
- Reference result: normalised_gcode_with_messages
79
- Reference source: tests/offsetTests/ToolRelativeOffsetTest/expect/stdout: the gcodetool output recorded upstream (8 lines), expect/return 0, expect/stderr one warning line; Expected stderr holds the tool auto-create warning. the program then prints G0 Z8.5 and G0 Z10 as a hand trace gives; expect/stderr holds the tool 1 auto-create warning; same #5422 behaviour note as the tool offset test
80
  Check: Task: given the program, produce what gcodetool prints (the normalised program with coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in), its stderr and return code. Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
81
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/ToolRelativeOffsetTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
82
  Companions (not included): tests/offsetTests/ToolRelativeOffsetTest/expect/return; tests/offsetTests/ToolRelativeOffsetTest/expect/stderr; tests/offsetTests/ToolRelativeOffsetTest/expect/stdout; tests/offsetTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
 
8
 
9
  **Self-contained:** 0 of 7 files are self-contained; for the others it is not stated. 7 of 7 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/GlobalTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code. Other files: see the notes.
12
 
13
  **Licence (SPDX):** GPL-2.0-or-later AND BSD-3-Clause
 
19
  - `tests/offsetTests/GlobalTest/data/stdin`: G92, G92.1, G92.2 and G92.3 global offset set, suspend, restore and clear with parameters #5210 to #5212 and #5420, #5421: run gcodetool and produce the normalised program with machine coordinates and the DEBUG values.
20
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
21
  Origin: a small upstream regression program written to exercise the interpreter feature named in the title
 
 
22
  Check: Task: given the program, produce what gcodetool prints (the normalised program with N numbers, coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in). Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare it with expect/stdout line by line after stripping CR and LF at line ends; stderr must be empty and the return code 0
23
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/GlobalTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
24
  Companions (not included): LICENSE; COPYING; tests/test.json; tests/testHarness; tests/offsetTests/test.json; tests/offsetTests/GlobalTest/expect/stdout; tests/offsetTests/GlobalTest/expect/stderr; tests/offsetTests/GlobalTest/expect/return
 
26
  - `tests/offsetTests/CSMultiTest/data/stdin`: G10 L20 P2 sets coordinate system 2 so the current point gets the given value, repeated, then G55 is selected: run gcodetool and produce the normalised program in machine coordinates and the DEBUG values of #5241, #5242, #5220, #5420, #5421.
27
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
28
  Origin: a small upstream regression program written to exercise the interpreter feature named in the title
 
 
29
  Check: Task: given the program, produce what gcodetool prints (the normalised program with N numbers, coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in). Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare it with expect/stdout line by line after stripping CR and LF at line ends; stderr must be empty and the return code 0
30
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/CSMultiTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
31
  Companions (not included): LICENSE; COPYING; tests/test.json; tests/testHarness; tests/offsetTests/test.json; tests/offsetTests/CSMultiTest/expect/stdout; tests/offsetTests/CSMultiTest/expect/stderr; tests/offsetTests/CSMultiTest/expect/return
 
33
  - `tests/offsetTests/G52Test/data/stdin`: G52 local offset applied twice: run gcodetool and produce the normalised program in machine coordinates and the DEBUG values of #5210 to #5212, #5420, #5421.
34
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
35
  Origin: a small upstream regression program written to exercise the interpreter feature named in the title
 
 
36
  Check: Task: given the program, produce what gcodetool prints (the normalised program with N numbers, coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in). Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare it with expect/stdout line by line after stripping CR and LF at line ends; stderr must be empty and the return code 0
37
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/G52Test/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
38
  Companions (not included): LICENSE; COPYING; tests/test.json; tests/testHarness; tests/offsetTests/test.json; tests/offsetTests/G52Test/expect/stdout; tests/offsetTests/G52Test/expect/stderr; tests/offsetTests/G52Test/expect/return
 
40
  - `tests/oCodeTests/ElseIfTest/data/stdin`: O-word subroutine with if, elseif and else branches called with 0, 1 and 2: run gcodetool and produce the normalised program (the MSG lines in branch order).
41
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
42
  Origin: a small upstream regression program written to exercise the interpreter feature named in the title
 
 
43
  Check: Task: given the program, produce what gcodetool prints (the normalised program with N numbers, coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in). Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare it with expect/stdout line by line after stripping CR and LF at line ends; stderr must be empty and the return code 0
44
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/oCodeTests/ElseIfTest/data/stdin`; tests/testHarness runs the command of tests/oCodeTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
45
  Companions (not included): LICENSE; COPYING; tests/test.json; tests/testHarness; tests/oCodeTests/test.json; tests/oCodeTests/ElseIfTest/expect/stdout; tests/oCodeTests/ElseIfTest/expect/stderr; tests/oCodeTests/ElseIfTest/expect/return
 
47
  - `tests/offsetTests/G53Test/data/stdin`: G53 machine coordinate move under a G92 offset: run gcodetool and produce the normalised program in machine coordinates with the DEBUG values.
48
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
49
  Origin: a small upstream regression program for G53 with an active G92 offset
 
 
50
  Check: Task: given the program, produce what gcodetool prints (the normalised program with coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in), its stderr and return code. Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
51
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/G53Test/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
52
  Companions (not included): tests/offsetTests/G53Test/expect/return; tests/offsetTests/G53Test/expect/stderr; tests/offsetTests/G53Test/expect/stdout; tests/offsetTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
 
54
  - `tests/offsetTests/ToolOffsetTest/data/stdin`
55
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
56
  Origin: a small upstream regression program for G10 L1 and G43 tool length compensation
 
 
57
  Check: Task: given the program, produce what gcodetool prints (the normalised program with coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in), its stderr and return code. Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
58
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/ToolOffsetTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
59
  Companions (not included): tests/offsetTests/ToolOffsetTest/expect/return; tests/offsetTests/ToolOffsetTest/expect/stderr; tests/offsetTests/ToolOffsetTest/expect/stdout; tests/offsetTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
 
61
  - `tests/offsetTests/ToolRelativeOffsetTest/data/stdin`
62
  Format: RS274/NGC G-code text (LinuxCNC-style dialect read by CAMotics); reference output is the gcodetool normalised G-code text with MSG lines
63
  Origin: a small upstream regression program for G10 L10 relative tool length setting
 
 
64
  Check: Task: given the program, produce what gcodetool prints (the normalised program with coordinates converted to machine coordinates, and DEBUG turned into MSG lines with the parameter values filled in), its stderr and return code. Judge: run `gcodetool < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
65
  Tools: CAMotics gcodetool (headless command line G-code parser and evaluator, prints the normalised program): `gcodetool < tests/offsetTests/ToolRelativeOffsetTest/data/stdin`; tests/testHarness runs the command of tests/offsetTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
66
  Companions (not included): tests/offsetTests/ToolRelativeOffsetTest/expect/return; tests/offsetTests/ToolRelativeOffsetTest/expect/stderr; tests/offsetTests/ToolRelativeOffsetTest/expect/stdout; tests/offsetTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
campaign_files/2026-10-hw-eda-tests-2/camotics-tplang-regression-tests-github-cauldrondevelopmentllc-camotics/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: generated_gcode. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** CAMotics tplang (headless command line runner of the TPL tool path language, prints the generated G-code): `tplang < tests/tplTests/RotateTest/data/stdin`; tests/testHarness runs the command of tests/tplTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code; CAMotics tplang (headless command line runner of the TPL tool path language, prints the generated G-code): `tplang < tests/tplTests/StarTest/data/stdin`; tests/testHarness runs the command of tests/tplTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
14
 
15
  **Licence (SPDX):** GPL-2.0-or-later AND BSD-3-Clause
@@ -21,8 +19,6 @@ Files:
21
  - `tests/tplTests/RotateTest/data/stdin`: TPL (CAMotics JavaScript tool path language) program that rotates a square about the (1,1,1) axis 40 times: run tplang and produce the generated G-code.
22
  Format: TPL program text (JavaScript dialect of CAMotics); reference output is the generated G-code text
23
  Origin: a small upstream regression program for the TPL rotate() matrix transform
24
- Reference result: generated_gcode
25
- Reference source: tests/tplTests/RotateTest/expect/stdout: the tplang output recorded upstream (324 lines), expect/return 0, expect/stderr empty; 40 angle comments match
26
  Check: Task: given the TPL program (JavaScript with the CAMotics tool path calls rapid, cut, icut, rotate, feed, speed, tool), produce the G-code tplang prints, its stderr (empty) and return code (0). Judge: run `tplang < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
27
  Tools: CAMotics tplang (headless command line runner of the TPL tool path language, prints the generated G-code): `tplang < tests/tplTests/RotateTest/data/stdin`; tests/testHarness runs the command of tests/tplTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
28
  Companions (not included): tests/tplTests/RotateTest/expect/return; tests/tplTests/RotateTest/expect/stderr; tests/tplTests/RotateTest/expect/stdout; tests/tplTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
@@ -30,8 +26,6 @@ Files:
30
  - `tests/tplTests/StarTest/data/stdin`: TPL program that cuts 10 nested squares 4 times with an accumulated rotate(pi/8): run tplang and produce the generated G-code.
31
  Format: TPL program text (JavaScript dialect of CAMotics); reference output is the generated G-code text
32
  Origin: a small upstream regression program for nested TPL functions and accumulated rotation about Z
33
- Reference result: generated_gcode
34
- Reference source: tests/tplTests/StarTest/expect/stdout: the tplang output recorded upstream (288 lines), expect/return 0, expect/stderr empty; header G21, F400., M6 T1, M3 S2000. and tail M5, M2 as the program gives
35
  Check: Task: given the TPL program (JavaScript with the CAMotics tool path calls rapid, cut, icut, rotate, feed, speed, tool), produce the G-code tplang prints, its stderr (empty) and return code (0). Judge: run `tplang < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
36
  Tools: CAMotics tplang (headless command line runner of the TPL tool path language, prints the generated G-code): `tplang < tests/tplTests/StarTest/data/stdin`; tests/testHarness runs the command of tests/tplTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
37
  Companions (not included): tests/tplTests/StarTest/data/project.xml; tests/tplTests/StarTest/expect/return; tests/tplTests/StarTest/expect/stderr; tests/tplTests/StarTest/expect/stdout; tests/tplTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
 
8
 
9
  **Self-contained:** Whether each file needs other upstream files is not stated. 2 of 2 files name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per file in the notes. Dependencies named upstream are listed per file in the notes.
10
 
 
 
11
  **Tools to open and check:** CAMotics tplang (headless command line runner of the TPL tool path language, prints the generated G-code): `tplang < tests/tplTests/RotateTest/data/stdin`; tests/testHarness runs the command of tests/tplTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code; CAMotics tplang (headless command line runner of the TPL tool path language, prints the generated G-code): `tplang < tests/tplTests/StarTest/data/stdin`; tests/testHarness runs the command of tests/tplTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
12
 
13
  **Licence (SPDX):** GPL-2.0-or-later AND BSD-3-Clause
 
19
  - `tests/tplTests/RotateTest/data/stdin`: TPL (CAMotics JavaScript tool path language) program that rotates a square about the (1,1,1) axis 40 times: run tplang and produce the generated G-code.
20
  Format: TPL program text (JavaScript dialect of CAMotics); reference output is the generated G-code text
21
  Origin: a small upstream regression program for the TPL rotate() matrix transform
 
 
22
  Check: Task: given the TPL program (JavaScript with the CAMotics tool path calls rapid, cut, icut, rotate, feed, speed, tool), produce the G-code tplang prints, its stderr (empty) and return code (0). Judge: run `tplang < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
23
  Tools: CAMotics tplang (headless command line runner of the TPL tool path language, prints the generated G-code): `tplang < tests/tplTests/RotateTest/data/stdin`; tests/testHarness runs the command of tests/tplTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
24
  Companions (not included): tests/tplTests/RotateTest/expect/return; tests/tplTests/RotateTest/expect/stderr; tests/tplTests/RotateTest/expect/stdout; tests/tplTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
 
26
  - `tests/tplTests/StarTest/data/stdin`: TPL program that cuts 10 nested squares 4 times with an accumulated rotate(pi/8): run tplang and produce the generated G-code.
27
  Format: TPL program text (JavaScript dialect of CAMotics); reference output is the generated G-code text
28
  Origin: a small upstream regression program for nested TPL functions and accumulated rotation about Z
 
 
29
  Check: Task: given the TPL program (JavaScript with the CAMotics tool path calls rapid, cut, icut, rotate, feed, speed, tool), produce the G-code tplang prints, its stderr (empty) and return code (0). Judge: run `tplang < data/stdin` from this commit, or take the agent output, and compare stdout, stderr and the return code with expect/stdout, expect/stderr and expect/return line by line after stripping CR and LF at line ends, as testHarness intends.
30
  Tools: CAMotics tplang (headless command line runner of the TPL tool path language, prints the generated G-code): `tplang < tests/tplTests/StarTest/data/stdin`; tests/testHarness runs the command of tests/tplTests/test.json with data/stdin as standard input and checks stdout, stderr and the return code
31
  Companions (not included): tests/tplTests/StarTest/data/project.xml; tests/tplTests/StarTest/expect/return; tests/tplTests/StarTest/expect/stderr; tests/tplTests/StarTest/expect/stdout; tests/tplTests/test.json; tests/test.json; tests/testHarness; LICENSE; COPYING
campaign_files/2026-10-hw-eda-tests-2/cocotb-examples/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 2 of 2 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: self_checking_testbench_pass_fail. Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Check: In a sandbox with the versions listed in dependencies run make in examples/matrix_multiplier/tests. Two variants are judged. Verilog variant: TOPLEVEL_LANG=verilog and SIM=icarus (both the Makefile defaults, tests/Makefile:4-5). VHDL variant: make TOPLEVEL_LANG=vhdl SIM=nvc (NVC is the open-source simulator the Makefile wires for VHDL, tests/Makefile:48-50; with SIM=ghdl the Makefile skips the run, tests/Makefile:63-66 "Skipping since GHDL doesn't support constants effectively", and test_matrix_multiplier.py:14-15 skips the runner test the same way, so GHDL is not used). Each variant passes when make exits 0 and results.xml has no <failure> element, i.e. the one cocotb test test_random passes its final assert tb.checker.errors == 0 (matrix_multiplier_tests.py:438-470); the same testbench and Python reference model judge both variants. For a write-the-RTL task, hide hdl/matrix_multiplier.sv (Verilog variant) or hdl/matrix_multiplier.vhd (VHDL variant) and keep tests/ unchanged (the directory has no model/); whether hdl/matrix_multiplier_pkg.vhd is given or hidden in the VHDL task is decided when the task is written. In a sandbox with the versions listed in dependencies run make in examples/adder/tests (TOPLEVEL_LANG defaults to verilog at tests/Makefile:7; tests/Makefile sets no SIM, so the simulator is the default of cocotb's own makefile, Makefile.sim:84 `SIM ?= icarus` at the same commit); for the VHDL variant run make TOPLEVEL_LANG=vhdl SIM=ghdl. The unit passes when make exits 0 and results.xml has no <failure> element, i.e. both cocotb tests adder_basic_test and adder_randomised_test pass. For a write-the-RTL task, hide the hdl/ files and keep tests/ and model/ unchanged.
14
 
15
  **Licence (SPDX):** CC0-1.0 AND CC0-1.0 AND BSD-3-Clause
@@ -21,16 +19,12 @@ Units:
21
  - `examples/matrix_multiplier/` (6 files): Cocotb example matrix_multiplier: parameterised matrix multiplier in SystemVerilog and VHDL with a cocotb testbench that has a Python reference model and scoreboard.
22
  Format: cocotb example directory: hdl/ DUT sources (SystemVerilog and VHDL), tests/Makefile for the cocotb makefile flow and pytest-runnable Python test module(s)
23
  Origin: example design and testbench shipped with the cocotb verification framework (tool_fixture)
24
- Reference result: self_checking_testbench_pass_fail
25
- Reference source: matrix_multiplier_tests.py computes the expected product of random input matrices in Python, a monitor collects the DUT output, a checker compares expected and actual element by element and the test asserts tb.checker.errors == 0
26
  Check: In a sandbox with the versions listed in dependencies run make in examples/matrix_multiplier/tests. Two variants are judged. Verilog variant: TOPLEVEL_LANG=verilog and SIM=icarus (both the Makefile defaults, tests/Makefile:4-5). VHDL variant: make TOPLEVEL_LANG=vhdl SIM=nvc (NVC is the open-source simulator the Makefile wires for VHDL, tests/Makefile:48-50; with SIM=ghdl the Makefile skips the run, tests/Makefile:63-66 "Skipping since GHDL doesn't support constants effectively", and test_matrix_multiplier.py:14-15 skips the runner test the same way, so GHDL is not used). Each variant passes when make exits 0 and results.xml has no <failure> element, i.e. the one cocotb test test_random passes its final assert tb.checker.errors == 0 (matrix_multiplier_tests.py:438-470); the same testbench and Python reference model judge both variants. For a write-the-RTL task, hide hdl/matrix_multiplier.sv (Verilog variant) or hdl/matrix_multiplier.vhd (VHDL variant) and keep tests/ unchanged (the directory has no model/); whether hdl/matrix_multiplier_pkg.vhd is given or hidden in the VHDL task is decided when the task is written.
27
  Companions (not included): LICENSE
28
  Dependencies: cocotb at commit 1f037538302fbb176e4f84dd249f0415aaf2fec5; open source, not packed; Icarus Verilog v13_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2026-10-05, GitHub release published 2026-03-02, prerelease s20251012 not counted; https://github.com/steveicarus/iverilog/releases/tag/v13_0); open source, not packed; NVC r1.23.0 (VHDL variant only; upstream does not pin it; latest release on or before the unit commit date 2026-10-05, GitHub release published 2026-09-19T17:02:33Z, read from https://api.github.com/repos/nickg/nvc/releases on 2026-10-06; https://github.com/nickg/nvc/releases/tag/r1.23.0); open source, not packed
29
  - `examples/adder/` (6 files): Cocotb example adder: parameterised adder in SystemVerilog and VHDL with a cocotb testbench checking against a Python reference model.
30
  Format: cocotb example directory: hdl/ DUT sources (SystemVerilog and VHDL), tests/Makefile for the cocotb makefile flow and pytest-runnable Python test module(s)
31
  Origin: example design and testbench shipped with the cocotb verification framework (tool_fixture)
32
- Reference result: self_checking_testbench_pass_fail
33
- Reference source: test_adder.py drives one fixed pair (5 + 10, adder_basic_test, lines 19-34) and 10 random pairs with A and B in 0..15 (adder_randomised_test, lines 36-52) and asserts dut.X.value == adder_model(A, B), where adder_model in model/adder_model.py is the Python reference (model/ is put on PYTHONPATH by tests/Makefile:11)
34
  Check: In a sandbox with the versions listed in dependencies run make in examples/adder/tests (TOPLEVEL_LANG defaults to verilog at tests/Makefile:7; tests/Makefile sets no SIM, so the simulator is the default of cocotb's own makefile, Makefile.sim:84 `SIM ?= icarus` at the same commit); for the VHDL variant run make TOPLEVEL_LANG=vhdl SIM=ghdl. The unit passes when make exits 0 and results.xml has no <failure> element, i.e. both cocotb tests adder_basic_test and adder_randomised_test pass. For a write-the-RTL task, hide the hdl/ files and keep tests/ and model/ unchanged.
35
  Companions (not included): LICENSE
36
  Dependencies: cocotb at commit 1f037538302fbb176e4f84dd249f0415aaf2fec5; open source, not packed; Icarus Verilog v13_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2026-10-05, GitHub release published 2026-03-02, prerelease s20251012 not counted; https://github.com/steveicarus/iverilog/releases/tag/v13_0); open source, not packed; GHDL v6.0.0 (VHDL variant only; upstream does not pin it; latest release on or before the unit commit date 2026-10-05, GitHub release published 2026-03-07; https://github.com/ghdl/ghdl/releases/tag/v6.0.0); open source, not packed
 
8
 
9
  **Self-contained:** Not self-contained: each unit is used together with other files from the same upstream. 2 of 2 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: In a sandbox with the versions listed in dependencies run make in examples/matrix_multiplier/tests. Two variants are judged. Verilog variant: TOPLEVEL_LANG=verilog and SIM=icarus (both the Makefile defaults, tests/Makefile:4-5). VHDL variant: make TOPLEVEL_LANG=vhdl SIM=nvc (NVC is the open-source simulator the Makefile wires for VHDL, tests/Makefile:48-50; with SIM=ghdl the Makefile skips the run, tests/Makefile:63-66 "Skipping since GHDL doesn't support constants effectively", and test_matrix_multiplier.py:14-15 skips the runner test the same way, so GHDL is not used). Each variant passes when make exits 0 and results.xml has no <failure> element, i.e. the one cocotb test test_random passes its final assert tb.checker.errors == 0 (matrix_multiplier_tests.py:438-470); the same testbench and Python reference model judge both variants. For a write-the-RTL task, hide hdl/matrix_multiplier.sv (Verilog variant) or hdl/matrix_multiplier.vhd (VHDL variant) and keep tests/ unchanged (the directory has no model/); whether hdl/matrix_multiplier_pkg.vhd is given or hidden in the VHDL task is decided when the task is written. In a sandbox with the versions listed in dependencies run make in examples/adder/tests (TOPLEVEL_LANG defaults to verilog at tests/Makefile:7; tests/Makefile sets no SIM, so the simulator is the default of cocotb's own makefile, Makefile.sim:84 `SIM ?= icarus` at the same commit); for the VHDL variant run make TOPLEVEL_LANG=vhdl SIM=ghdl. The unit passes when make exits 0 and results.xml has no <failure> element, i.e. both cocotb tests adder_basic_test and adder_randomised_test pass. For a write-the-RTL task, hide the hdl/ files and keep tests/ and model/ unchanged.
12
 
13
  **Licence (SPDX):** CC0-1.0 AND CC0-1.0 AND BSD-3-Clause
 
19
  - `examples/matrix_multiplier/` (6 files): Cocotb example matrix_multiplier: parameterised matrix multiplier in SystemVerilog and VHDL with a cocotb testbench that has a Python reference model and scoreboard.
20
  Format: cocotb example directory: hdl/ DUT sources (SystemVerilog and VHDL), tests/Makefile for the cocotb makefile flow and pytest-runnable Python test module(s)
21
  Origin: example design and testbench shipped with the cocotb verification framework (tool_fixture)
 
 
22
  Check: In a sandbox with the versions listed in dependencies run make in examples/matrix_multiplier/tests. Two variants are judged. Verilog variant: TOPLEVEL_LANG=verilog and SIM=icarus (both the Makefile defaults, tests/Makefile:4-5). VHDL variant: make TOPLEVEL_LANG=vhdl SIM=nvc (NVC is the open-source simulator the Makefile wires for VHDL, tests/Makefile:48-50; with SIM=ghdl the Makefile skips the run, tests/Makefile:63-66 "Skipping since GHDL doesn't support constants effectively", and test_matrix_multiplier.py:14-15 skips the runner test the same way, so GHDL is not used). Each variant passes when make exits 0 and results.xml has no <failure> element, i.e. the one cocotb test test_random passes its final assert tb.checker.errors == 0 (matrix_multiplier_tests.py:438-470); the same testbench and Python reference model judge both variants. For a write-the-RTL task, hide hdl/matrix_multiplier.sv (Verilog variant) or hdl/matrix_multiplier.vhd (VHDL variant) and keep tests/ unchanged (the directory has no model/); whether hdl/matrix_multiplier_pkg.vhd is given or hidden in the VHDL task is decided when the task is written.
23
  Companions (not included): LICENSE
24
  Dependencies: cocotb at commit 1f037538302fbb176e4f84dd249f0415aaf2fec5; open source, not packed; Icarus Verilog v13_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2026-10-05, GitHub release published 2026-03-02, prerelease s20251012 not counted; https://github.com/steveicarus/iverilog/releases/tag/v13_0); open source, not packed; NVC r1.23.0 (VHDL variant only; upstream does not pin it; latest release on or before the unit commit date 2026-10-05, GitHub release published 2026-09-19T17:02:33Z, read from https://api.github.com/repos/nickg/nvc/releases on 2026-10-06; https://github.com/nickg/nvc/releases/tag/r1.23.0); open source, not packed
25
  - `examples/adder/` (6 files): Cocotb example adder: parameterised adder in SystemVerilog and VHDL with a cocotb testbench checking against a Python reference model.
26
  Format: cocotb example directory: hdl/ DUT sources (SystemVerilog and VHDL), tests/Makefile for the cocotb makefile flow and pytest-runnable Python test module(s)
27
  Origin: example design and testbench shipped with the cocotb verification framework (tool_fixture)
 
 
28
  Check: In a sandbox with the versions listed in dependencies run make in examples/adder/tests (TOPLEVEL_LANG defaults to verilog at tests/Makefile:7; tests/Makefile sets no SIM, so the simulator is the default of cocotb's own makefile, Makefile.sim:84 `SIM ?= icarus` at the same commit); for the VHDL variant run make TOPLEVEL_LANG=vhdl SIM=ghdl. The unit passes when make exits 0 and results.xml has no <failure> element, i.e. both cocotb tests adder_basic_test and adder_randomised_test pass. For a write-the-RTL task, hide the hdl/ files and keep tests/ and model/ unchanged.
29
  Companions (not included): LICENSE
30
  Dependencies: cocotb at commit 1f037538302fbb176e4f84dd249f0415aaf2fec5; open source, not packed; Icarus Verilog v13_0 (iverilog and vvp; upstream does not pin it; latest release on or before the unit commit date 2026-10-05, GitHub release published 2026-03-02, prerelease s20251012 not counted; https://github.com/steveicarus/iverilog/releases/tag/v13_0); open source, not packed; GHDL v6.0.0 (VHDL variant only; upstream does not pin it; latest release on or before the unit commit date 2026-10-05, GitHub release published 2026-03-07; https://github.com/ghdl/ghdl/releases/tag/v6.0.0); open source, not packed
campaign_files/2026-10-hw-eda-tests-2/cocotb/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The unit is self-contained. Dependencies named upstream are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: self_checking_testbench_pass_fail. Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Check: Build in a sandbox with the versions listed in dependencies: in examples/first_steps run `make` (SIM = icarus). The unit passes when the cocotb results.xml reports no failed test; only test_self_checking compares against expected values, the other seven tests are tutorial steps that pass when they run without error. For a write-the-RTL task, hide or replace counter.sv and keep counter_tests.py unchanged.
14
 
15
  **Licence (SPDX):** CC0-1.0
@@ -21,7 +19,5 @@ Units:
21
  - `examples/first_steps/` (3 files): Cocotb first_steps: 8-bit loadable counter in SystemVerilog with a self-checking cocotb test.
22
  Format: SystemVerilog DUT with a cocotb (Python) testbench and a cocotb make file
23
  Origin: self-checking cocotb test written by the cocotb maintainers for a tutorial DUT (tool_fixture): expected values are fixed in the test
24
- Reference result: self_checking_testbench_pass_fail
25
- Reference source: test_self_checking loads 10, checks the count equals 10, then enables counting and checks the count increases by exactly 1 on each of 20 clock edges
26
  Check: Build in a sandbox with the versions listed in dependencies: in examples/first_steps run `make` (SIM = icarus). The unit passes when the cocotb results.xml reports no failed test; only test_self_checking compares against expected values, the other seven tests are tutorial steps that pass when they run without error. For a write-the-RTL task, hide or replace counter.sv and keep counter_tests.py unchanged.
27
  Dependencies: cocotb at the unit commit f622ed090a46b919548e744725a7d1f23ea93ffe (the make file uses COCOTB_TOPLEVEL and COCOTB_TEST_MODULES, cocotb 2.x names); open source, not packed; Icarus Verilog v13_0 (examples/first_steps/Makefile line 7 `SIM = icarus`; .github/generate-envs.py line 109 names v13_0 as the latest release version used in CI); open source, not packed; Python 3.12
 
8
 
9
  **Self-contained:** The unit is self-contained. Dependencies named upstream are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Check: Build in a sandbox with the versions listed in dependencies: in examples/first_steps run `make` (SIM = icarus). The unit passes when the cocotb results.xml reports no failed test; only test_self_checking compares against expected values, the other seven tests are tutorial steps that pass when they run without error. For a write-the-RTL task, hide or replace counter.sv and keep counter_tests.py unchanged.
12
 
13
  **Licence (SPDX):** CC0-1.0
 
19
  - `examples/first_steps/` (3 files): Cocotb first_steps: 8-bit loadable counter in SystemVerilog with a self-checking cocotb test.
20
  Format: SystemVerilog DUT with a cocotb (Python) testbench and a cocotb make file
21
  Origin: self-checking cocotb test written by the cocotb maintainers for a tutorial DUT (tool_fixture): expected values are fixed in the test
 
 
22
  Check: Build in a sandbox with the versions listed in dependencies: in examples/first_steps run `make` (SIM = icarus). The unit passes when the cocotb results.xml reports no failed test; only test_self_checking compares against expected values, the other seven tests are tutorial steps that pass when they run without error. For a write-the-RTL task, hide or replace counter.sv and keep counter_tests.py unchanged.
23
  Dependencies: cocotb at the unit commit f622ed090a46b919548e744725a7d1f23ea93ffe (the make file uses COCOTB_TOPLEVEL and COCOTB_TEST_MODULES, cocotb 2.x names); open source, not packed; Icarus Verilog v13_0 (examples/first_steps/Makefile line 7 `SIM = icarus`; .github/generate-envs.py line 109 names v13_0 as the latest release version used in CI); open source, not packed; Python 3.12
campaign_files/2026-10-m-issue-trackers-2/academy-bus-nj-gtfs/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the service calendar itself (active services on a date, removed or added dates, day counts). Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -22,8 +20,6 @@ Units:
22
  Read out of the archive: https://content.njtransit.com/sites/default/files/developers-resources/Academy_bus_data.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt)
24
  Origin: Live static GTFS feed published by Academy Bus, published on NJ TRANSIT developer resources at https://www.njtransit.com/Academy_bus_data.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
25
- Reference result: facts computed from the service calendar itself (active services on a date, removed or added dates, day counts)
26
- Reference source: calendar.txt weekday flags with start_date and end_date, calendar_dates.txt exception rows, agency_timezone of agency.txt, all from the pinned members.
27
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
28
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
29
  Companions (not included): calendar_dates.txt; agency.txt; feed_info.txt
 
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
12
 
13
  **Licence (SPDX):** NOASSERTION
 
20
  Read out of the archive: https://content.njtransit.com/sites/default/files/developers-resources/Academy_bus_data.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt)
22
  Origin: Live static GTFS feed published by Academy Bus, published on NJ TRANSIT developer resources at https://www.njtransit.com/Academy_bus_data.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
 
 
23
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
24
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
25
  Companions (not included): calendar_dates.txt; agency.txt; feed_info.txt
campaign_files/2026-10-m-issue-trackers-2/agl-gerrit-changes/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** curl the Gerrit REST API (GET https://gerrit.automotivelinux.org/gerrit/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `changes/31039/detail.json`: AGL/meta-agl-demo change 31039: Move AGL IC demo recipes to meta-agl-demo-shared.
22
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
23
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
24
- Reference result: facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments)
25
- Reference source: The pinned response: changes/31039/detail.json.
26
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: agl-jobbuilder, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
27
  Tools: curl the Gerrit REST API (GET https://gerrit.automotivelinux.org/gerrit/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** curl the Gerrit REST API (GET https://gerrit.automotivelinux.org/gerrit/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `changes/31039/detail.json`: AGL/meta-agl-demo change 31039: Move AGL IC demo recipes to meta-agl-demo-shared.
20
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
21
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
 
 
22
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: agl-jobbuilder, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
23
  Tools: curl the Gerrit REST API (GET https://gerrit.automotivelinux.org/gerrit/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
campaign_files/2026-10-m-issue-trackers-2/altlinux-devel/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m329-04/2019-May.txt.gz`: [devel] q: отключить debuginfo (7 messages, 2019-05-05 to 2019-05-06).
22
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
23
  Origin: Real list traffic in a closed past month of the list's archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole. Message bodies are in Russian; the questions only use headers (Message-ID, In-Reply-To, References, From, Date), so the answers do not depend on the body language.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread, read from the pinned month file.
26
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC (a Date header without zone is read as UTC) and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and '[devel] Проблемы тестовой пересборки пакетов, использующих cmake' (6 messages, 6 people, root <20190515094908.75fd22efe77134279fc302a9@altlinux.org>); message count, Message-ID set, last message time and first-reply seconds agreed on both.
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m329-04/2019-May.txt.gz`: [devel] q: отключить debuginfo (7 messages, 2019-05-05 to 2019-05-06).
20
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
21
  Origin: Real list traffic in a closed past month of the list's archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole. Message bodies are in Russian; the questions only use headers (Message-ID, In-Reply-To, References, From, Date), so the answers do not depend on the body language.
 
 
22
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC (a Date header without zone is read as UTC) and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and '[devel] Проблемы тестовой пересборки пакетов, использующих cmake' (6 messages, 6 people, root <20190515094908.75fd22efe77134279fc302a9@altlinux.org>); message count, Message-ID set, last message time and first-reply seconds agreed on both.
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
campaign_files/2026-10-m-issue-trackers-2/aomedia-gerrit-changes/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** curl the Gerrit REST API (GET https://aomedia-review.googlesource.com/changes/215701/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `changes/215701/detail.json`: Aom change 215701: Add AV1 encoder public API fuzzer.
22
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote, messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
23
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
24
- Reference result: facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments)
25
- Reference source: The pinned response: changes/215701/detail.json.
26
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: AO Media, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler|maloo|janitor|autotest|gerrit). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
27
  Tools: curl the Gerrit REST API (GET https://aomedia-review.googlesource.com/changes/215701/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** curl the Gerrit REST API (GET https://aomedia-review.googlesource.com/changes/215701/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `changes/215701/detail.json`: Aom change 215701: Add AV1 encoder public API fuzzer.
20
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote, messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
21
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
 
 
22
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: AO Media, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler|maloo|janitor|autotest|gerrit). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
23
  Tools: curl the Gerrit REST API (GET https://aomedia-review.googlesource.com/changes/215701/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
campaign_files/2026-10-m-issue-trackers-2/aosp-gerrit-changes/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** curl the Gerrit REST API (GET https://android-review.googlesource.com/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `changes/2716973/detail.json`: platform/frameworks/base change 2716973: Set default values for screen dim feature.
22
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
23
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
24
- Reference result: facts computed from the change's own review timeline (status, submit time and submitter, patch sets, votes, reviewers added or removed, review comments)
25
- Reference source: The pinned response: changes/2716973/detail.json.
26
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: CrystalBall Performance Presubmit, Treehugger Robot, android-build-ayeaye@system.gserviceaccount.com, android-test-infra-android-presubmit-result-watcher@system.gserviceaccount.com, android-test-infra-autosubmit@system.gserviceaccount.com, android-test-infra-presubmit-run-create-handler@system.gserviceaccount.com, android-test-infra-stale-change-watcher@system.gserviceaccount.com, android-test-infra-workplan-finisher@system.gserviceaccount.com, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
27
  Tools: curl the Gerrit REST API (GET https://android-review.googlesource.com/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** curl the Gerrit REST API (GET https://android-review.googlesource.com/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `changes/2716973/detail.json`: platform/frameworks/base change 2716973: Set default values for screen dim feature.
20
  Format: Gerrit REST API ChangeInfo JSON from the /detail endpoint, preceded by the anti-XSSI line ")]}'": status, created, updated, submitted, submitter, owner, labels with every vote (DETAILED_LABELS), messages (author, real_author, date, tag, _revision_number, message text) and reviewer_updates (updated, updated_by, reviewer, state REVIEWER/CC/REMOVED). Patch-set uploads are messages tagged autogenerated:gerrit:newPatchSet; votes appear in message text as 'Patch Set N: Label+V' and in labels.*.all. The endpoint returns all messages and reviewer updates of the change, so the record is complete.
21
  Origin: Live Gerrit REST interface of a closed (merged) code review change, collected as a NOASSERTION pointer row under spec general rule 4: the detail JSON was fetched twice apart and was byte-identical. Unit file kept only outside the repository for checking and then deleted.
 
 
22
  Check: Exact match against values read from the pinned JSON: status, submitted and submitter; reviewer_updates entries (reviewer, state, updated_by, updated); messages tagged autogenerated:gerrit:newPatchSet (patch-set count, _revision_number); votes from labels.*.all (value, _account_id) and from message first lines 'Patch Set N: Label+V'. A human review comment is a message whose author is not a bot or service account, whose tag does not start with 'autogenerated:', and whose text has content beyond the 'Patch Set N: votes' first line or names inline comments '(N comments)'. Bot and service accounts excluded from the comment count: CrystalBall Performance Presubmit, Treehugger Robot, android-build-ayeaye@system.gserviceaccount.com, android-test-infra-android-presubmit-result-watcher@system.gserviceaccount.com, android-test-infra-autosubmit@system.gserviceaccount.com, android-test-infra-presubmit-run-create-handler@system.gserviceaccount.com, android-test-infra-stale-change-watcher@system.gserviceaccount.com, android-test-infra-workplan-finisher@system.gserviceaccount.com, accounts tagged SERVICE_USER, and any account whose name, username or email matches the pattern (bot|zuul|jenkins|ci|luci|gerrit code review|service user|automation|buildbot|lint|sonar|tryjob|cq|autosubmit|copybara|tricium|gerritforge|verifier|checker|scheduler). Deterministic, no model judgement. Before answers are issued, the extraction program must pass on 2 hand-checked units of this source (spec general rule 5); hand-checking is pending for the delivery step.
23
  Tools: curl the Gerrit REST API (GET https://android-review.googlesource.com/changes/{number}/detail), strip the first line ")]}'" and read the JSON with jq or Python json; the git-review CLI and Gerrit's ssh 'gerrit query --format=JSON --all-approvals --comments' reach the same data where ssh access is open
campaign_files/2026-10-m-issue-trackers-2/arin-arin-ppml/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m141-01/2019-March.txt.gz`: [arin-ppml] Policy action menu for Waiting List (4.1.8) - feedback requested. (6 messages, 2019-03-07 to 2019-03-14).
22
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
23
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread, read from the pinned month file.
26
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and '[arin-ppml] Draft Policy ARIN-2019-4: Allow Inter-regional IPv6 Resource Transfers' (15 messages, root <407a9024-9a6e-69c1-eb7c-ce143dd2adf6@arin.net>); message count, people, last sender and time, first-reply seconds and direct replies agreed on both.
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m141-01/2019-March.txt.gz`: [arin-ppml] Policy action menu for Waiting List (4.1.8) - feedback requested. (6 messages, 2019-03-07 to 2019-03-14).
20
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
21
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
 
 
22
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and '[arin-ppml] Draft Policy ARIN-2019-4: Allow Inter-regional IPv6 Resource Transfers' (15 messages, root <407a9024-9a6e-69c1-eb7c-ce143dd2adf6@arin.net>); message count, people, last sender and time, first-reply seconds and direct replies agreed on both.
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
campaign_files/2026-10-m-issue-trackers-2/atlassian-public-jira/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** Each file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the ticket's own timeline (status transitions, assignee changes, comments). Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** curl the Jira Data Center REST API v2 (/rest/api/2/issue/KEY?expand=changelog) on jira.atlassian.com and read it with jq or Python json; the jira Python library or a Jira command line client also work.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,21 +19,15 @@ Files:
21
  - `rest/api/2/issue/CONFCLOUD-72040.json`: Atlassian-public-jira CONFCLOUD-72040: Ability to display Insight Assets data in Confluence (Closed, Fixed, 2024-09-30).
22
  Format: Jira REST API v2 issue JSON with expand=changelog: fields (status, resolution, resolutiondate, assignee, reporter, comment.comments[]) and changelog.histories[] (author, created, items[] with field, fromString, toString).
23
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: two copies fetched apart were byte-identical. Unit file kept only outside the repository for checking and then deleted.
24
- Reference result: facts computed from the ticket's own timeline (status transitions, assignee changes, comments)
25
- Reference source: The pinned response: rest/api/2/issue/CONFCLOUD-72040.json.
26
  Check: Exact match against values read from the pinned JSON: changelog.histories[].items[] with field 'status' or 'assignee' (fromString, toString, created, author.displayName) and fields.comment.comments[] (author.displayName, created). Deterministic, no model judgement. Human comment count excludes bot accounts, recognised by display name against the pattern bot|automation|jenkins|gerrit|jira|system|migrat|importer: No bot account found in this ticket: display names of all comment authors were checked against the pattern and none matched. Timeline events = comments plus changelog histories; changelog.total equals the number of histories returned, so the history is not truncated. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
27
  Tools: curl the Jira Data Center REST API v2 (/rest/api/2/issue/KEY?expand=changelog) on jira.atlassian.com and read it with jq or Python json; the jira Python library or a Jira command line client also work.
28
  - `rest/api/2/issue/CONFSERVER-26011.json`: Atlassian-public-jira CONFSERVER-26011: Broken image when copy paste image into the editor using Mac's Safari, Chrome and Edge (Closed, Fixed, 2026-07-01).
29
  Format: Jira REST API v2 issue JSON with expand=changelog: fields (status, resolution, resolutiondate, assignee, reporter, comment.comments[]) and changelog.histories[] (author, created, items[] with field, fromString, toString).
30
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: two copies fetched apart were byte-identical. Unit file kept only outside the repository for checking and then deleted.
31
- Reference result: facts computed from the ticket's own timeline (status transitions, assignee changes, comments)
32
- Reference source: The pinned response: rest/api/2/issue/CONFSERVER-26011.json.
33
  Check: Exact match against values read from the pinned JSON: changelog.histories[].items[] with field 'status' or 'assignee' (fromString, toString, created, author.displayName) and fields.comment.comments[] (author.displayName, created). Deterministic, no model judgement. Human comment count excludes bot accounts, recognised by display name against the pattern bot|automation|jenkins|gerrit|jira|system|migrat|importer: No bot account found in this ticket: display names of all comment authors were checked against the pattern and none matched. Timeline events = comments plus changelog histories; changelog.total equals the number of histories returned, so the history is not truncated. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
34
  Tools: curl the Jira Data Center REST API v2 (/rest/api/2/issue/KEY?expand=changelog) on jira.atlassian.com and read it with jq or Python json; the jira Python library or a Jira command line client also work.
35
  - `rest/api/2/issue/JRASERVER-79420.json`: Atlassian-public-jira JRASERVER-79420: InvalidDocumentFieldException while reindexing or performing issue operations in Jira DC (Closed, Fixed, 2026-07-08).
36
  Format: Jira REST API v2 issue JSON with expand=changelog: fields (status, resolution, resolutiondate, assignee, reporter, comment.comments[]) and changelog.histories[] (author, created, items[] with field, fromString, toString).
37
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: two copies fetched apart were byte-identical. Unit file kept only outside the repository for checking and then deleted.
38
- Reference result: facts computed from the ticket's own timeline (status transitions, assignee changes, comments)
39
- Reference source: The pinned response: rest/api/2/issue/JRASERVER-79420.json.
40
  Check: Exact match against values read from the pinned JSON: changelog.histories[].items[] with field 'status' or 'assignee' (fromString, toString, created, author.displayName) and fields.comment.comments[] (author.displayName, created). Deterministic, no model judgement. Human comment count excludes bot accounts, recognised by display name against the pattern bot|automation|jenkins|gerrit|jira|system|migrat|importer: No bot account found in this ticket: display names of all comment authors were checked against the pattern and none matched. Timeline events = comments plus changelog histories; changelog.total equals the number of histories returned, so the history is not truncated. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
41
  Tools: curl the Jira Data Center REST API v2 (/rest/api/2/issue/KEY?expand=changelog) on jira.atlassian.com and read it with jq or Python json; the jira Python library or a Jira command line client also work.
 
8
 
9
  **Self-contained:** Each file is self-contained.
10
 
 
 
11
  **Tools to open and check:** curl the Jira Data Center REST API v2 (/rest/api/2/issue/KEY?expand=changelog) on jira.atlassian.com and read it with jq or Python json; the jira Python library or a Jira command line client also work.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `rest/api/2/issue/CONFCLOUD-72040.json`: Atlassian-public-jira CONFCLOUD-72040: Ability to display Insight Assets data in Confluence (Closed, Fixed, 2024-09-30).
20
  Format: Jira REST API v2 issue JSON with expand=changelog: fields (status, resolution, resolutiondate, assignee, reporter, comment.comments[]) and changelog.histories[] (author, created, items[] with field, fromString, toString).
21
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: two copies fetched apart were byte-identical. Unit file kept only outside the repository for checking and then deleted.
 
 
22
  Check: Exact match against values read from the pinned JSON: changelog.histories[].items[] with field 'status' or 'assignee' (fromString, toString, created, author.displayName) and fields.comment.comments[] (author.displayName, created). Deterministic, no model judgement. Human comment count excludes bot accounts, recognised by display name against the pattern bot|automation|jenkins|gerrit|jira|system|migrat|importer: No bot account found in this ticket: display names of all comment authors were checked against the pattern and none matched. Timeline events = comments plus changelog histories; changelog.total equals the number of histories returned, so the history is not truncated. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
23
  Tools: curl the Jira Data Center REST API v2 (/rest/api/2/issue/KEY?expand=changelog) on jira.atlassian.com and read it with jq or Python json; the jira Python library or a Jira command line client also work.
24
  - `rest/api/2/issue/CONFSERVER-26011.json`: Atlassian-public-jira CONFSERVER-26011: Broken image when copy paste image into the editor using Mac's Safari, Chrome and Edge (Closed, Fixed, 2026-07-01).
25
  Format: Jira REST API v2 issue JSON with expand=changelog: fields (status, resolution, resolutiondate, assignee, reporter, comment.comments[]) and changelog.histories[] (author, created, items[] with field, fromString, toString).
26
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: two copies fetched apart were byte-identical. Unit file kept only outside the repository for checking and then deleted.
 
 
27
  Check: Exact match against values read from the pinned JSON: changelog.histories[].items[] with field 'status' or 'assignee' (fromString, toString, created, author.displayName) and fields.comment.comments[] (author.displayName, created). Deterministic, no model judgement. Human comment count excludes bot accounts, recognised by display name against the pattern bot|automation|jenkins|gerrit|jira|system|migrat|importer: No bot account found in this ticket: display names of all comment authors were checked against the pattern and none matched. Timeline events = comments plus changelog histories; changelog.total equals the number of histories returned, so the history is not truncated. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
28
  Tools: curl the Jira Data Center REST API v2 (/rest/api/2/issue/KEY?expand=changelog) on jira.atlassian.com and read it with jq or Python json; the jira Python library or a Jira command line client also work.
29
  - `rest/api/2/issue/JRASERVER-79420.json`: Atlassian-public-jira JRASERVER-79420: InvalidDocumentFieldException while reindexing or performing issue operations in Jira DC (Closed, Fixed, 2026-07-08).
30
  Format: Jira REST API v2 issue JSON with expand=changelog: fields (status, resolution, resolutiondate, assignee, reporter, comment.comments[]) and changelog.histories[] (author, created, items[] with field, fromString, toString).
31
  Origin: Live tracker interface of a closed ticket, collected as a NOASSERTION pointer row under spec general rule 4: two copies fetched apart were byte-identical. Unit file kept only outside the repository for checking and then deleted.
 
 
32
  Check: Exact match against values read from the pinned JSON: changelog.histories[].items[] with field 'status' or 'assignee' (fromString, toString, created, author.displayName) and fields.comment.comments[] (author.displayName, created). Deterministic, no model judgement. Human comment count excludes bot accounts, recognised by display name against the pattern bot|automation|jenkins|gerrit|jira|system|migrat|importer: No bot account found in this ticket: display names of all comment authors were checked against the pattern and none matched. Timeline events = comments plus changelog histories; changelog.total equals the number of histories returned, so the history is not truncated. Before answers are issued, the extraction program must pass on 2 hand-checked tickets of this tracker (spec general rule 5); hand-checking is pending for the delivery step.
33
  Tools: curl the Jira Data Center REST API v2 (/rest/api/2/issue/KEY?expand=changelog) on jira.atlassian.com and read it with jq or Python json; the jira Python library or a Jira command line client also work.
campaign_files/2026-10-m-issue-trackers-2/atoc-gtfs-2021-archive/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the service calendar itself (active services on a date, removed or added dates, day counts). Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -22,8 +20,6 @@ Units:
22
  Read out of the archive: https://openmobilitydata-data.s3.us-west-1.amazonaws.com/public/feeds/association-of-train-operating-companies/284/20210423/gtfs.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
24
  Origin: Live static GTFS feed published by Association of Train Operating Companies timetable data, converted to GTFS and archived by OpenMobilityData (Transitfeeds), now held by MobilityData at https://openmobilitydata-data.s3.us-west-1.amazonaws.com/public/feeds/association-of-train-operating-companies/284/20210423/gtfs.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
25
- Reference result: facts computed from the service calendar itself (active services on a date, removed or added dates, day counts)
26
- Reference source: calendar.txt weekday flags with start_date and end_date, calendar_dates.txt exception rows, agency_timezone of agency.txt, all from the pinned members.
27
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
28
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
29
  Companions (not included): calendar_dates.txt; agency.txt
 
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
12
 
13
  **Licence (SPDX):** NOASSERTION
 
20
  Read out of the archive: https://openmobilitydata-data.s3.us-west-1.amazonaws.com/public/feeds/association-of-train-operating-companies/284/20210423/gtfs.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
22
  Origin: Live static GTFS feed published by Association of Train Operating Companies timetable data, converted to GTFS and archived by OpenMobilityData (Transitfeeds), now held by MobilityData at https://openmobilitydata-data.s3.us-west-1.amazonaws.com/public/feeds/association-of-train-operating-companies/284/20210423/gtfs.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
 
 
23
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
24
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
25
  Companions (not included): calendar_dates.txt; agency.txt
campaign_files/2026-10-m-issue-trackers-2/big-blue-bus-gtfs/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the service calendar itself (active services on a date, removed or added dates, day counts). Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
14
 
15
  **Licence (SPDX):** ODC-By-1.0
@@ -22,8 +20,6 @@ Units:
22
  Read out of the archive: https://gtfs.bigbluebus.com/current.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt)
24
  Origin: Live static GTFS feed published by City of Santa Monica, Big Blue Bus at https://gtfs.bigbluebus.com/current.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
25
- Reference result: facts computed from the service calendar itself (active services on a date, removed or added dates, day counts)
26
- Reference source: calendar.txt weekday flags with start_date and end_date, calendar_dates.txt exception rows, agency_timezone of agency.txt, all from the pinned members.
27
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
28
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
29
  Companions (not included): calendar_dates.txt; agency.txt; feed_info.txt
 
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
12
 
13
  **Licence (SPDX):** ODC-By-1.0
 
20
  Read out of the archive: https://gtfs.bigbluebus.com/current.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt, feed_info.txt)
22
  Origin: Live static GTFS feed published by City of Santa Monica, Big Blue Bus at https://gtfs.bigbluebus.com/current.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
 
 
23
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
24
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
25
  Companions (not included): calendar_dates.txt; agency.txt; feed_info.txt
campaign_files/2026-10-m-issue-trackers-2/bitcoin-dev/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m329-10/t.mbox.gz`: [bitcoin-dev] Signet (9 messages, 2019-03-08 to 2019-03-14).
22
  Format: public-inbox thread mbox, gzip-compressed (t.mbox.gz): every message public-inbox groups into the thread, with full headers (From, Date, Subject, Message-ID, In-Reply-To, References) and bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
23
  Origin: Real traffic of the bitcoin-dev list (Linux Foundation Mailman, since moved to Google Groups as bitcoindev), mirrored in the public-inbox archive at gnusha.org/pi/bitcoindev. The unit is the public-inbox thread mbox t.mbox.gz of the 'Signet' thread (March 2019, closed). The file holds 11 messages: 9 chain back to the root by In-Reply-To/References.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread, read from the pinned t.mbox.gz.
26
  Check: Exact match against values computed from the pinned t.mbox.gz: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: same address, or same name after removing accents and letter case. The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions): the chosen thread (also read message by message) and '[bitcoin-dev] Removal of reject network messages from Bitcoin Core (BIP61)' (t.mbox.gz fetched once, 22 messages, 11 people, root <CAK51vgDO2Tg38XbW0pqAnO3ETJ_qf8owRsUYsTXmrf7H2yGZtw@mail.gmail.com>); message count, Message-ID set, last message time and first-reply seconds agreed on both.
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m329-10/t.mbox.gz`: [bitcoin-dev] Signet (9 messages, 2019-03-08 to 2019-03-14).
20
  Format: public-inbox thread mbox, gzip-compressed (t.mbox.gz): every message public-inbox groups into the thread, with full headers (From, Date, Subject, Message-ID, In-Reply-To, References) and bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
21
  Origin: Real traffic of the bitcoin-dev list (Linux Foundation Mailman, since moved to Google Groups as bitcoindev), mirrored in the public-inbox archive at gnusha.org/pi/bitcoindev. The unit is the public-inbox thread mbox t.mbox.gz of the 'Signet' thread (March 2019, closed). The file holds 11 messages: 9 chain back to the root by In-Reply-To/References.
 
 
22
  Check: Exact match against values computed from the pinned t.mbox.gz: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: same address, or same name after removing accents and letter case. The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions): the chosen thread (also read message by message) and '[bitcoin-dev] Removal of reject network messages from Bitcoin Core (BIP61)' (t.mbox.gz fetched once, 22 messages, 11 people, root <CAK51vgDO2Tg38XbW0pqAnO3ETJ_qf8owRsUYsTXmrf7H2yGZtw@mail.gmail.com>); message count, Message-ID set, last message time and first-reply seconds agreed on both.
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
campaign_files/2026-10-m-issue-trackers-2/blender-bf-committers/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The file is self-contained.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the thread headers. Where the reference result comes from and how a solution is checked is written per file in the notes.
12
-
13
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
14
 
15
  **Licence (SPDX):** NOASSERTION
@@ -21,7 +19,5 @@ Files:
21
  - `m141-02/2019-March.txt.gz`: [Bf-committers] Sculpt Mode development (4 messages, 2019-03-10 to 2019-03-12).
22
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
23
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
24
- Reference result: facts computed from the thread headers
25
- Reference source: Message-ID, In-Reply-To, References, From and Date headers of the messages in the thread, read from the pinned month file.
26
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and '[Bf-committers] Blender 2.79 Nightly builds using different Python versions' (4 messages, root <CA+VsiZpQ44oY4RAqoKNvfyEh=Av0A2dw7iSsRuE1Ra-yo5_nxw@mail.gmail.com>); message count, people, last sender and time, first-reply seconds and direct replies agreed on both.
27
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
 
8
 
9
  **Self-contained:** The file is self-contained.
10
 
 
 
11
  **Tools to open and check:** Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
12
 
13
  **Licence (SPDX):** NOASSERTION
 
19
  - `m141-02/2019-March.txt.gz`: [Bf-committers] Sculpt Mode development (4 messages, 2019-03-10 to 2019-03-12).
20
  Format: Pipermail month archive, mbox text gzip-compressed: full headers (From with obfuscated address and name, Date, Subject, Message-ID, In-Reply-To, References) and plain-text bodies. Cut the thread by root Message-ID through the In-Reply-To closure (last References entry when In-Reply-To is missing).
21
  Origin: Real list traffic in a closed past month of the list's Pipermail archive. The unit is one thread, cut out of the month file by its root Message-ID; the month file is the pinned pointer and is delivered whole.
 
 
22
  Check: Exact match against values computed from the pinned month file: parse it as mbox, drop repeated copies of a Message-ID already seen (first copy kept), take the root Message-ID, collect every message whose In-Reply-To (or last References entry when In-Reply-To is missing) chains back to it, then count, take dates in UTC and compare sender names. Message-IDs are compared after removing angle brackets and double quotes. People are counted by person: From entries with the same address, or the same name after removing accents and letter case, are one person (the names in scale are the first spelling of each). The first-reply interval is the exact difference in whole seconds between the Date headers of the root and the earliest reply, converted to UTC; Date headers have one-second resolution, so no rounding is involved. Deterministic. The program was checked on 2 threads of this source by a second, independent parser (splits the raw text on mbox From lines and reads headers with regular expressions, no mailbox or email module): the chosen thread (also read message by message) and '[Bf-committers] Blender 2.79 Nightly builds using different Python versions' (4 messages, root <CA+VsiZpQ44oY4RAqoKNvfyEh=Av0A2dw7iSsRuE1Ra-yo5_nxw@mail.gmail.com>); message count, people, last sender and time, first-reply seconds and direct replies agreed on both.
23
  Tools: Read the archive as mbox with Python's mailbox and email modules (or mutt -f / notmuch on the decompressed file); rebuild the thread from Message-ID, In-Reply-To and References headers.
campaign_files/2026-10-m-issue-trackers-2/bluestar-passenger-gtfs/README.md CHANGED
@@ -8,8 +8,6 @@
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
11
- **Answers:** No answer files are included. Reference result: Facts computed from the service calendar itself (active services on a date, removed or added dates, day counts). Where the reference result comes from and how a solution is checked is written per unit in the notes.
12
-
13
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
14
 
15
  **Licence (SPDX):** OGL-UK-3.0
@@ -22,8 +20,6 @@ Units:
22
  Read out of the archive: https://s3-eu-west-1.amazonaws.com/passenger-sources/bluestar/gtfs/bluestar_1791379183.zip
23
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
24
  Origin: Live static GTFS feed published by Bluestar (published through Passenger, data.discoverpassenger.com) at https://s3-eu-west-1.amazonaws.com/passenger-sources/bluestar/gtfs/bluestar_1791379183.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
25
- Reference result: facts computed from the service calendar itself (active services on a date, removed or added dates, day counts)
26
- Reference source: calendar.txt weekday flags with start_date and end_date, calendar_dates.txt exception rows, agency_timezone of agency.txt, all from the pinned members.
27
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
28
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
29
  Companions (not included): calendar_dates.txt; agency.txt
 
8
 
9
  **Self-contained:** The unit is self-contained. 1 of 1 units name companion files (other upstream files read alongside); companions are not included in this delivery and are listed per unit in the notes.
10
 
 
 
11
  **Tools to open and check:** Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
12
 
13
  **Licence (SPDX):** OGL-UK-3.0
 
20
  Read out of the archive: https://s3-eu-west-1.amazonaws.com/passenger-sources/bluestar/gtfs/bluestar_1791379183.zip
21
  Format: text/csv (GTFS Schedule calendar.txt; companions calendar_dates.txt, agency.txt)
22
  Origin: Live static GTFS feed published by Bluestar (published through Passenger, data.discoverpassenger.com) at https://s3-eu-west-1.amazonaws.com/passenger-sources/bluestar/gtfs/bluestar_1791379183.zip. calendar.txt and its companions were read by member out of the zip with HTTP Range requests (spec general rule 4: by member); kept as a snapshot pinned by sha256.
 
 
23
  Check: deterministic, no model judgement. Hand-checking 2 samples per source is pending for the delivery step (spec general rule 5).
24
  Tools: Read the CSV files with Python csv or pandas (or sqlite3 .import) and expand service days per GTFS rules; gtfs-kit or partridge compute active service_ids per date headless; dates are local to agency_timezone (Python zoneinfo).
25
  Companions (not included): calendar_dates.txt; agency.txt