Duplicate from mondk/SWE-DockerLess
Browse filesCo-authored-by: mon-dark <mondk@users.noreply.huggingface.co>
- .gitattributes +60 -0
- README.md +120 -0
- SWE-DockerLess.jsonl +4 -0
- build_dataset.py +355 -0
- fixed/cronlite/cronlite/fields.py +82 -0
- fixed/miniconf/miniconf/config.py +92 -0
- fixed/semverlite/semverlite/range.py +80 -0
- fixed/semverlite/semverlite/version.py +96 -0
- fixed/ttlcache/ttlcache/cache.py +94 -0
- newtests/cronlite/tests/test_step_and_sunday.py +43 -0
- newtests/miniconf/tests/test_interpolation_edge_cases.py +64 -0
- newtests/semverlite/tests/test_prerelease_precedence.py +55 -0
- newtests/ttlcache/tests/test_lru_semantics.py +55 -0
- predictions/cross.jsonl +4 -0
- predictions/empty.jsonl +4 -0
- predictions/partial.jsonl +4 -0
- reports/cross.json +45 -0
- reports/empty.json +45 -0
- reports/gold.json +49 -0
- reports/partial.json +66 -0
- repos/cronlite/README.md +43 -0
- repos/cronlite/cronlite/__init__.py +27 -0
- repos/cronlite/cronlite/expr.py +59 -0
- repos/cronlite/cronlite/fields.py +74 -0
- repos/cronlite/tests/test_basic.py +98 -0
- repos/miniconf/README.md +52 -0
- repos/miniconf/miniconf/__init__.py +21 -0
- repos/miniconf/miniconf/config.py +67 -0
- repos/miniconf/miniconf/errors.py +30 -0
- repos/miniconf/miniconf/loader.py +14 -0
- repos/miniconf/tests/test_basic.py +82 -0
- repos/semverlite/README.md +45 -0
- repos/semverlite/semverlite/__init__.py +8 -0
- repos/semverlite/semverlite/range.py +71 -0
- repos/semverlite/semverlite/version.py +84 -0
- repos/semverlite/tests/test_version.py +106 -0
- repos/ttlcache/README.md +40 -0
- repos/ttlcache/tests/test_cache.py +111 -0
- repos/ttlcache/ttlcache/__init__.py +8 -0
- repos/ttlcache/ttlcache/cache.py +83 -0
- repos/ttlcache/ttlcache/clock.py +28 -0
- run_harness.py +238 -0
- tools/make_controls.py +168 -0
- tools/validate.py +174 -0
.gitattributes
ADDED
|
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.avro filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.lz4 filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.mds filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
# Audio files - uncompressed
|
| 40 |
+
*.pcm filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
*.sam filter=lfs diff=lfs merge=lfs -text
|
| 42 |
+
*.raw filter=lfs diff=lfs merge=lfs -text
|
| 43 |
+
# Audio files - compressed
|
| 44 |
+
*.aac filter=lfs diff=lfs merge=lfs -text
|
| 45 |
+
*.flac filter=lfs diff=lfs merge=lfs -text
|
| 46 |
+
*.mp3 filter=lfs diff=lfs merge=lfs -text
|
| 47 |
+
*.ogg filter=lfs diff=lfs merge=lfs -text
|
| 48 |
+
*.wav filter=lfs diff=lfs merge=lfs -text
|
| 49 |
+
# Image files - uncompressed
|
| 50 |
+
*.bmp filter=lfs diff=lfs merge=lfs -text
|
| 51 |
+
*.gif filter=lfs diff=lfs merge=lfs -text
|
| 52 |
+
*.png filter=lfs diff=lfs merge=lfs -text
|
| 53 |
+
*.tiff filter=lfs diff=lfs merge=lfs -text
|
| 54 |
+
# Image files - compressed
|
| 55 |
+
*.jpg filter=lfs diff=lfs merge=lfs -text
|
| 56 |
+
*.jpeg filter=lfs diff=lfs merge=lfs -text
|
| 57 |
+
*.webp filter=lfs diff=lfs merge=lfs -text
|
| 58 |
+
# Video files - compressed
|
| 59 |
+
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 60 |
+
*.webm filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,120 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
tags:
|
| 6 |
+
- swe
|
| 7 |
+
- dockerless
|
| 8 |
+
configs:
|
| 9 |
+
- config_name: swe
|
| 10 |
+
data_files: "SWE-DockerLess.jsonl"
|
| 11 |
+
---
|
| 12 |
+
# SWE-DockerLess
|
| 13 |
+
|
| 14 |
+
A SWE-bench-shaped benchmark that needs **no Docker, no network and no prebuilt
|
| 15 |
+
image**. Every row carries its own repo snapshot, so an instance is reproduced by
|
| 16 |
+
writing files to a temp directory, `git init`, `git apply` and running pytest.
|
| 17 |
+
|
| 18 |
+
Rows are byte-for-byte SWE-bench compatible: same field names, same
|
| 19 |
+
`FAIL_TO_PASS` / `PASS_TO_PASS` JSON-string encoding, same "resolved" rule. One
|
| 20 |
+
extra `dockerless` object carries the snapshot and the test command.
|
| 21 |
+
|
| 22 |
+
```
|
| 23 |
+
SWE-DockerLess.jsonl 4 instances, 62.9 KiB (~16k tokens total)
|
| 24 |
+
build_dataset.py rebuilds the jsonl from repos/ + fixed/ + newtests/ and
|
| 25 |
+
derives F2P/P2P by actually running pytest
|
| 26 |
+
run_harness.py scoring harness (--gold, or --predictions preds.jsonl)
|
| 27 |
+
repos/ the four buggy base repos (the source of the snapshots)
|
| 28 |
+
fixed/ gold-fix versions of the touched files
|
| 29 |
+
newtests/ the tests that become test_patch
|
| 30 |
+
tools/validate.py schema, patch-application and budget checks
|
| 31 |
+
tools/make_controls.py generates empty / cross / partial negative controls
|
| 32 |
+
reports/ harness output for gold and the three controls
|
| 33 |
+
```
|
| 34 |
+
|
| 35 |
+
## Instances
|
| 36 |
+
|
| 37 |
+
| instance_id | repo | the bug | F2P | P2P | gold patch |
|
| 38 |
+
| --- | --- | --- | --- | --- | --- |
|
| 39 |
+
| `swedl__miniconf-1` | swedl/miniconf | `${k}` interpolation recurses on cycles (no `CircularReferenceError`) and `$$` never escapes | 9 | 15 | 41 added lines |
|
| 40 |
+
| `swedl__ttlcache-1` | swedl/ttlcache | LRU order ignores reads and overwrites; `len()` counts expired rows | 5 | 12 | 12 added lines |
|
| 41 |
+
| `swedl__cronlite-1` | swedl/cronlite | `5/15` collapses to `{5}`; day-of-week `7` rejected | 8 | 21 | 11 added lines |
|
| 42 |
+
| `swedl__semverlite-1` | swedl/semverlite | prerelease ordering is lexicographic; prereleases satisfy release ranges | 7 | 19 | 23 added lines |
|
| 43 |
+
|
| 44 |
+
29 FAIL_TO_PASS tests, 67 PASS_TO_PASS tests. Difficulty is SWE-bench-like rather
|
| 45 |
+
than toy-like: each fix needs a behaviour change in a real algorithm (recursive
|
| 46 |
+
descent with cycle detection, an LRU invariant, cron field expansion, semver
|
| 47 |
+
precedence), and the hidden tests pin exact semantics — error message text, the
|
| 48 |
+
0-6 weekday scale, the "prerelease must opt in" range rule — that a shallow
|
| 49 |
+
patch misses. The `partial` control below scores 5/9, 2/5, 6/8 and 3/7 F2P
|
| 50 |
+
without resolving a single instance.
|
| 51 |
+
|
| 52 |
+
## Row schema
|
| 53 |
+
|
| 54 |
+
| field | notes |
|
| 55 |
+
| --- | --- |
|
| 56 |
+
| `instance_id` | `swedl__<repo>-1` |
|
| 57 |
+
| `repo` | `swedl/<repo>` |
|
| 58 |
+
| `base_commit` | real 40-char sha of the snapshot commit (deterministic) |
|
| 59 |
+
| `patch` | gold fix, `git diff` unified diff against `base_commit` |
|
| 60 |
+
| `test_patch` | adds test files only, never touches source |
|
| 61 |
+
| `problem_statement` | issue-style report; states the contract, never names a test |
|
| 62 |
+
| `hints_text` | `""` |
|
| 63 |
+
| `created_at`, `version` | `2026-10-30`, `1.0` |
|
| 64 |
+
| `environment_setup_commit` | equals `base_commit` |
|
| 65 |
+
| `FAIL_TO_PASS`, `PASS_TO_PASS` | JSON-encoded strings, as in SWE-bench |
|
| 66 |
+
| `dockerless.files` | `{path: content}` snapshot of the repo at `base_commit` |
|
| 67 |
+
| `dockerless.test_cmd` | `python -m pytest -v --tb=no -p no:cacheprovider {tests}` |
|
| 68 |
+
| `dockerless.test_files` | the node-id prefixes to run |
|
| 69 |
+
|
| 70 |
+
## Running it
|
| 71 |
+
|
| 72 |
+
```bash
|
| 73 |
+
python -m pip install pytest # the only dependency
|
| 74 |
+
python build_dataset.py # regenerate + re-verify the jsonl
|
| 75 |
+
python run_harness.py --gold # sanity check, expect 4/4
|
| 76 |
+
python run_harness.py --predictions predictions/partial.jsonl --out reports/partial.json
|
| 77 |
+
```
|
| 78 |
+
|
| 79 |
+
A predictions file is JSONL of `{"instance_id", "model_name_or_path",
|
| 80 |
+
"model_patch"}` — the SWE-bench format. `resolved` requires **all** F2P and
|
| 81 |
+
**all** P2P to pass.
|
| 82 |
+
|
| 83 |
+
## Verified results
|
| 84 |
+
|
| 85 |
+
```
|
| 86 |
+
$ python run_harness.py --gold
|
| 87 |
+
instance resolved F2P P2P note
|
| 88 |
+
swedl__miniconf-1 yes 9/9 15/15 ok
|
| 89 |
+
swedl__ttlcache-1 yes 5/5 12/12 ok
|
| 90 |
+
swedl__cronlite-1 yes 8/8 21/21 ok
|
| 91 |
+
swedl__semverlite-1 yes 7/7 19/19 ok
|
| 92 |
+
resolved 4/4 (100.0%) in 10.3s
|
| 93 |
+
|
| 94 |
+
$ python run_harness.py --predictions predictions/empty.jsonl -> 0/4
|
| 95 |
+
$ python run_harness.py --predictions predictions/cross.jsonl -> 0/4
|
| 96 |
+
$ python run_harness.py --predictions predictions/partial.jsonl -> 0/4
|
| 97 |
+
```
|
| 98 |
+
|
| 99 |
+
`build_dataset.py` is the ground truth for F2P/P2P: it materialises a clean
|
| 100 |
+
checkout, applies `test_patch`, runs pytest, applies `patch`, runs pytest again,
|
| 101 |
+
and classifies every test by observed outcome. It refuses to write a row if the
|
| 102 |
+
gold patch leaves any failure behind, if F2P is empty, or if a test disappears.
|
| 103 |
+
|
| 104 |
+
## Budget
|
| 105 |
+
|
| 106 |
+
Measured for `max_tokens = 2048`, `context = 240000`:
|
| 107 |
+
|
| 108 |
+
- prompt per instance: ~2.0k–2.6k tokens (problem statement + snapshot)
|
| 109 |
+
- gold patch: 11–41 added lines, ~200–700 tokens — comfortably inside 2048
|
| 110 |
+
- whole dataset: ~16k tokens, so all four rows fit in one context if you want
|
| 111 |
+
|
| 112 |
+
## Deliberate deviations from SWE-bench
|
| 113 |
+
|
| 114 |
+
- **Synthetic repos.** These are four purpose-built Python packages, not real
|
| 115 |
+
GitHub projects, so there is no `repo` clone step and no upstream history.
|
| 116 |
+
- **Snapshot in the row** instead of a Docker image. `base_commit` is a genuine
|
| 117 |
+
commit sha of that snapshot, so `git apply` and `git diff` behave normally.
|
| 118 |
+
- **No install step.** Packages are importable from the repo root, so
|
| 119 |
+
`dockerless.install` is empty and pytest runs with the repo as rootdir.
|
| 120 |
+
- **Python only**, single language, to keep the harness to `python + git + pytest`.
|
SWE-DockerLess.jsonl
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"instance_id": "swedl__miniconf-1", "repo": "swedl/miniconf", "base_commit": "f7b5b3a4cfec71e8e6db94c76ae5c4e2ea509216", "patch": "diff --git a/miniconf/config.py b/miniconf/config.py\nindex 8751912..6bdf695 100644\n--- a/miniconf/config.py\n+++ b/miniconf/config.py\n@@ -1,12 +1,8 @@\n \"\"\"The Config object: layered key/value storage with ${...} interpolation.\"\"\"\n \n-import re\n-\n-from .errors import MissingKeyError\n+from .errors import CircularReferenceError, MissingKeyError\n from .loader import load_layers\n \n-REFERENCE = re.compile(r\"\\$\\{([A-Za-z_][A-Za-z0-9_.-]*)\\}\")\n-\n \n class Config:\n \"\"\"A small configuration mapping whose string values may reference keys.\n@@ -51,17 +47,46 @@ class Config:\n \"\"\"Return every key resolved to a plain value.\"\"\"\n return {key: self._resolve(key) for key in self._raw}\n \n- def _resolve(self, key):\n+ def _resolve(self, key, stack=()):\n+ if key in stack:\n+ cycle = stack[stack.index(key):] + (key,)\n+ raise CircularReferenceError(cycle)\n+ if key not in self._raw:\n+ raise MissingKeyError(key, stack[-1] if stack else None)\n value = self._raw[key]\n if not isinstance(value, str):\n return value\n- return self._interpolate(value, key)\n-\n- def _interpolate(self, text, origin):\n- def replace(match):\n- name = match.group(1)\n- if name not in self._raw:\n- raise MissingKeyError(name, origin)\n- return str(self._resolve(name))\n-\n- return REFERENCE.sub(replace, text)\n+ return self._render(value, stack + (key,))\n+\n+ def _render(self, text, stack):\n+ \"\"\"Resolve references in ``text`` in a single left-to-right pass.\"\"\"\n+ out = []\n+ index = 0\n+ length = len(text)\n+ while index < length:\n+ char = text[index]\n+ if char != \"$\":\n+ out.append(char)\n+ index += 1\n+ continue\n+ if text.startswith(\"$$\", index):\n+ out.append(\"$\")\n+ index += 2\n+ continue\n+ if text.startswith(\"${\", index):\n+ end = text.find(\"}\", index + 2)\n+ if end == -1:\n+ out.append(char)\n+ index += 1\n+ continue\n+ name = text[index + 2:end]\n+ if not name:\n+ out.append(char)\n+ index += 1\n+ continue\n+ out.append(str(self._resolve(name, stack)))\n+ index = end + 1\n+ continue\n+ out.append(char)\n+ index += 1\n+ return \"\".join(out)\n", "test_patch": "diff --git a/tests/test_interpolation_edge_cases.py b/tests/test_interpolation_edge_cases.py\nnew file mode 100644\nindex 0000000..6289238\n--- /dev/null\n+++ b/tests/test_interpolation_edge_cases.py\n@@ -0,0 +1,64 @@\n+import pytest\n+\n+from miniconf import CircularReferenceError, Config, MissingKeyError\n+\n+\n+def test_dollar_escape():\n+ cfg = Config({\"price\": \"$${amount}\"})\n+ assert cfg[\"price\"] == \"${amount}\"\n+\n+\n+def test_escape_next_to_reference():\n+ cfg = Config({\"name\": \"wren\", \"msg\": \"$${name} is ${name}\"})\n+ assert cfg[\"msg\"] == \"${name} is wren\"\n+\n+\n+def test_self_reference():\n+ cfg = Config({\"a\": \"x${a}\"})\n+ with pytest.raises(CircularReferenceError):\n+ cfg[\"a\"]\n+\n+\n+def test_as_dict_raises_on_cycle():\n+ cfg = Config({\"a\": \"${b}\", \"b\": \"${a}\"})\n+ with pytest.raises(CircularReferenceError):\n+ cfg.as_dict()\n+\n+\n+def test_cycle_not_involving_requested_key():\n+ cfg = Config({\"a\": \"fine\", \"b\": \"${c}\", \"c\": \"${b}\"})\n+ assert cfg[\"a\"] == \"fine\"\n+ with pytest.raises(CircularReferenceError):\n+ cfg[\"b\"]\n+\n+\n+def test_two_cycle_chain_starts_at_requested_key():\n+ cfg = Config({\"a\": \"${b}\", \"b\": \"${a}\"})\n+ with pytest.raises(CircularReferenceError) as excinfo:\n+ cfg[\"a\"]\n+ assert str(excinfo.value) == \"circular reference: a -> b -> a\"\n+\n+\n+def test_self_reference_message():\n+ cfg = Config({\"loop\": \"${loop}\"})\n+ with pytest.raises(CircularReferenceError) as excinfo:\n+ cfg[\"loop\"]\n+ assert str(excinfo.value) == \"circular reference: loop -> loop\"\n+\n+\n+def test_escaped_text_is_not_reinterpolated_through_a_reference():\n+ cfg = Config({\"inner\": \"$${leak}\", \"outer\": \"${inner}\"})\n+ assert cfg[\"outer\"] == \"${leak}\"\n+\n+\n+def test_missing_key_inside_reference_chain():\n+ cfg = Config({\"a\": \"${b}\", \"b\": \"${nope}\"})\n+ with pytest.raises(MissingKeyError) as excinfo:\n+ cfg[\"a\"]\n+ assert str(excinfo.value) == \"missing key 'nope' referenced from 'b'\"\n+\n+\n+def test_escape_does_not_hide_a_real_cycle():\n+ cfg = Config({\"a\": \"$${b}\", \"b\": \"${a}\"})\n+ assert cfg[\"a\"] == \"${b}\"\n+ assert cfg[\"b\"] == \"${b}\"\n", "problem_statement": "Config interpolation blows up on reference cycles and expands escaped text.\n\nWe use miniconf for service configs that are layered from a shared file plus a\nper-environment overlay. Two problems, both reproducible on main:\n\n1. A cycle anywhere in the mapping crashes with `RecursionError` instead of the\n documented error. `Config({'a': '${b}', 'b': '${a}'})['a']` recurses until the\n interpreter gives up, and the traceback is a thousand frames of `_interpolate`.\n The README says this must raise `CircularReferenceError` with a message naming\n the chain, and we rely on that message in our error reporting.\n\n2. `$$` does not escape. Our TLS config has `cert_subject: \"$${cluster}\"`, meaning\n the literal text `${cluster}` to be handed to a downstream templating tool, but\n miniconf substitutes the cluster name into it. The README's escape rule is not\n implemented at all.\n\nBoth bugs live in `Config._resolve` / `Config._interpolate` in\n`miniconf/config.py`. `README.md` section \"Interpolation contract\" is the spec we\nwant honoured, including the exact wording of the cycle message and the rule that\nsubstituted text is never rescanned for references. A cycle that is unrelated to\nthe key being read must not affect that read, and `as_dict()` must surface cycles\nlike any other read.\n\nThe existing suite passes; the escape and cycle behaviour was never covered.", "hints_text": "", "created_at": "2026-10-30", "version": "1.0", "environment_setup_commit": "f7b5b3a4cfec71e8e6db94c76ae5c4e2ea509216", "FAIL_TO_PASS": "[\"tests/test_interpolation_edge_cases.py::test_as_dict_raises_on_cycle\", \"tests/test_interpolation_edge_cases.py::test_cycle_not_involving_requested_key\", \"tests/test_interpolation_edge_cases.py::test_dollar_escape\", \"tests/test_interpolation_edge_cases.py::test_escape_does_not_hide_a_real_cycle\", \"tests/test_interpolation_edge_cases.py::test_escape_next_to_reference\", \"tests/test_interpolation_edge_cases.py::test_escaped_text_is_not_reinterpolated_through_a_reference\", \"tests/test_interpolation_edge_cases.py::test_self_reference\", \"tests/test_interpolation_edge_cases.py::test_self_reference_message\", \"tests/test_interpolation_edge_cases.py::test_two_cycle_chain_starts_at_requested_key\"]", "PASS_TO_PASS": "[\"tests/test_basic.py::test_as_dict_resolves_all\", \"tests/test_basic.py::test_chained_reference\", \"tests/test_basic.py::test_config_rejects_non_dict\", \"tests/test_basic.py::test_from_layers_override\", \"tests/test_basic.py::test_get_returns_default_for_unknown_key\", \"tests/test_basic.py::test_getitem_raises_keyerror_for_unknown_key\", \"tests/test_basic.py::test_load_layers_rejects_non_dict\", \"tests/test_basic.py::test_missing_key_message\", \"tests/test_basic.py::test_missing_key_raises\", \"tests/test_basic.py::test_non_string_passthrough\", \"tests/test_basic.py::test_plain_values\", \"tests/test_basic.py::test_raw_does_not_resolve\", \"tests/test_basic.py::test_simple_reference\", \"tests/test_basic.py::test_two_references_in_one_value\", \"tests/test_interpolation_edge_cases.py::test_missing_key_inside_reference_chain\"]", "dockerless": {"language": "python", "python_requires": ">=3.9", "install": [], "run_from": ".", "test_cmd": "python -m pytest -v --tb=no -p no:cacheprovider {tests}", "test_files": ["tests/test_basic.py", "tests/test_interpolation_edge_cases.py"], "files": {"miniconf/__init__.py": "\"\"\"miniconf - a tiny layered configuration library with string interpolation.\"\"\"\n\nfrom .config import Config\nfrom .errors import (\n CircularReferenceError,\n ConfigError,\n InterpolationError,\n MissingKeyError,\n)\nfrom .loader import load_layers\n\n__version__ = \"0.4.1\"\n\n__all__ = [\n \"CircularReferenceError\",\n \"Config\",\n \"ConfigError\",\n \"InterpolationError\",\n \"MissingKeyError\",\n \"load_layers\",\n]\n", "miniconf/config.py": "\"\"\"The Config object: layered key/value storage with ${...} interpolation.\"\"\"\n\nimport re\n\nfrom .errors import MissingKeyError\nfrom .loader import load_layers\n\nREFERENCE = re.compile(r\"\\$\\{([A-Za-z_][A-Za-z0-9_.-]*)\\}\")\n\n\nclass Config:\n \"\"\"A small configuration mapping whose string values may reference keys.\n\n See README.md for the interpolation contract.\n \"\"\"\n\n def __init__(self, data):\n if not isinstance(data, dict):\n raise TypeError(\"Config data must be a dict\")\n self._raw = dict(data)\n\n @classmethod\n def from_layers(cls, *layers):\n \"\"\"Build a Config from one or more dict layers, later layers winning.\"\"\"\n return cls(load_layers(*layers))\n\n def raw(self, key, default=None):\n \"\"\"Return the unresolved value stored under ``key``.\"\"\"\n return self._raw.get(key, default)\n\n def keys(self):\n return list(self._raw)\n\n def get(self, key, default=None):\n if key not in self._raw:\n return default\n return self._resolve(key)\n\n def __getitem__(self, key):\n if key not in self._raw:\n raise KeyError(key)\n return self._resolve(key)\n\n def __contains__(self, key):\n return key in self._raw\n\n def __len__(self):\n return len(self._raw)\n\n def as_dict(self):\n \"\"\"Return every key resolved to a plain value.\"\"\"\n return {key: self._resolve(key) for key in self._raw}\n\n def _resolve(self, key):\n value = self._raw[key]\n if not isinstance(value, str):\n return value\n return self._interpolate(value, key)\n\n def _interpolate(self, text, origin):\n def replace(match):\n name = match.group(1)\n if name not in self._raw:\n raise MissingKeyError(name, origin)\n return str(self._resolve(name))\n\n return REFERENCE.sub(replace, text)\n", "miniconf/errors.py": "\"\"\"Exception hierarchy for miniconf.\"\"\"\n\n\nclass ConfigError(Exception):\n \"\"\"Base class for every error raised by miniconf.\"\"\"\n\n\nclass InterpolationError(ConfigError):\n \"\"\"Raised when a ``${...}`` reference cannot be resolved.\"\"\"\n\n\nclass MissingKeyError(InterpolationError):\n \"\"\"Raised when a reference points at a key that does not exist.\"\"\"\n\n def __init__(self, key, referenced_from=None):\n self.key = key\n self.referenced_from = referenced_from\n if referenced_from is None:\n message = \"missing key: %r\" % (key,)\n else:\n message = \"missing key %r referenced from %r\" % (key, referenced_from)\n super().__init__(message)\n\n\nclass CircularReferenceError(InterpolationError):\n \"\"\"Raised when interpolation would loop forever.\"\"\"\n\n def __init__(self, cycle):\n self.cycle = list(cycle)\n super().__init__(\"circular reference: \" + \" -> \".join(self.cycle))\n", "miniconf/loader.py": "\"\"\"Layer merging for miniconf configuration dicts.\"\"\"\n\nfrom .errors import ConfigError\n\n\ndef load_layers(*layers):\n \"\"\"Merge configuration layers left to right; later layers win.\"\"\"\n merged = {}\n for layer in layers:\n if not isinstance(layer, dict):\n raise ConfigError(\"layer must be a dict, got %r\" % (type(layer).__name__,))\n for key, value in layer.items():\n merged[key] = value\n return merged\n", "README.md": "# miniconf\n\nA tiny layered configuration library. Values may reference other keys with\n`${key}` and are resolved lazily on access.\n\n## Interpolation contract\n\nThis is the documented behaviour of `Config`; `tests/` covers the common cases.\n\n1. **References.** A string value may contain any number of `${key}` references.\n The reference is replaced by the *fully resolved* value of `key`, stringified\n with `str()`. References may chain: `a` -> `b` -> `c`.\n2. **Single pass.** Substituted text is never scanned again for references. If\n `inner` resolves to the literal text `${leak}`, then `\"${inner}\"` resolves to\n `${leak}` and not to the value of `leak`.\n3. **Escaping.** `$$` is an escape that produces a single literal `$`. It never\n starts a reference, so `\"$${amount}\"` resolves to the five characters\n `${amount}` — the reference is *not* expanded. `\"$${name} is ${name}\"` with\n `name=wren` resolves to `\"${name} is wren\"`.\n4. **Cycles.** A reference cycle, including a self reference such as\n `a = \"x${a}\"`, must raise `CircularReferenceError` instead of recursing. The\n message is `\"circular reference: \"` followed by the key chain joined with\n `\" -> \"`, starting at the key the caller asked for and ending with that same\n key repeated. Asking for `a` where `a -> b -> a` raises\n `\"circular reference: a -> b -> a\"`; asking for `a` where `a -> a` raises\n `\"circular reference: a -> a\"`.\n5. **Missing keys.** A reference to an unknown key raises `MissingKeyError`.\n6. **Non-string values** are returned as-is and are never scanned.\n\n`Config.as_dict()` resolves every key, so any cycle or missing reference in the\nmapping surfaces there too. A cycle that is unrelated to the key being read must\nnot affect that read: with `a=\"fine\"`, `b=\"${c}\"`, `c=\"${b}\"`, reading `a` still\nreturns `\"fine\"` while reading `b` raises.\n\n## Usage\n\n```python\nfrom miniconf import Config\n\ncfg = Config.from_layers(\n {\"host\": \"localhost\", \"port\": 8080},\n {\"url\": \"http://${host}:${port}/\"},\n)\ncfg[\"url\"] # 'http://localhost:8080/'\ncfg.as_dict() # {'host': 'localhost', 'port': 8080, 'url': 'http://localhost:8080/'}\n```\n\n## Tests\n\n```\npython -m pytest -q\n```\n", "tests/test_basic.py": "import pytest\n\nfrom miniconf import Config, ConfigError, MissingKeyError, load_layers\n\n\ndef test_plain_values():\n cfg = Config({\"name\": \"app\", \"port\": 8080})\n assert cfg[\"name\"] == \"app\"\n assert cfg[\"port\"] == 8080\n assert len(cfg) == 2\n\n\ndef test_simple_reference():\n cfg = Config({\"host\": \"localhost\", \"url\": \"http://${host}/\"})\n assert cfg[\"url\"] == \"http://localhost/\"\n\n\ndef test_two_references_in_one_value():\n cfg = Config({\"host\": \"db\", \"port\": 5432, \"dsn\": \"${host}:${port}\"})\n assert cfg[\"dsn\"] == \"db:5432\"\n\n\ndef test_chained_reference():\n cfg = Config({\"a\": \"x\", \"b\": \"${a}-y\", \"c\": \"${b}-z\"})\n assert cfg[\"c\"] == \"x-y-z\"\n\n\ndef test_missing_key_raises():\n cfg = Config({\"a\": \"${nope}\"})\n with pytest.raises(MissingKeyError):\n cfg[\"a\"]\n\n\ndef test_missing_key_message():\n cfg = Config({\"a\": \"${nope}\"})\n with pytest.raises(MissingKeyError) as excinfo:\n cfg[\"a\"]\n assert str(excinfo.value) == \"missing key 'nope' referenced from 'a'\"\n\n\ndef test_non_string_passthrough():\n cfg = Config({\"n\": 3, \"flag\": True, \"nested\": {\"x\": 1}})\n assert cfg[\"n\"] == 3\n assert cfg[\"flag\"] is True\n assert cfg[\"nested\"] == {\"x\": 1}\n\n\ndef test_get_returns_default_for_unknown_key():\n cfg = Config({\"a\": \"1\"})\n assert cfg.get(\"zzz\") is None\n assert cfg.get(\"zzz\", \"dflt\") == \"dflt\"\n\n\ndef test_getitem_raises_keyerror_for_unknown_key():\n cfg = Config({\"a\": \"1\"})\n with pytest.raises(KeyError):\n cfg[\"zzz\"]\n\n\ndef test_raw_does_not_resolve():\n cfg = Config({\"a\": \"1\", \"b\": \"${a}\"})\n assert cfg.raw(\"b\") == \"${a}\"\n\n\ndef test_from_layers_override():\n cfg = Config.from_layers({\"a\": \"1\", \"b\": \"2\"}, {\"b\": \"3\"})\n assert cfg.as_dict() == {\"a\": \"1\", \"b\": \"3\"}\n\n\ndef test_load_layers_rejects_non_dict():\n with pytest.raises(ConfigError):\n load_layers({\"a\": 1}, [\"not\", \"a\", \"dict\"])\n\n\ndef test_as_dict_resolves_all():\n cfg = Config({\"a\": \"1\", \"b\": \"${a}2\"})\n assert cfg.as_dict() == {\"a\": \"1\", \"b\": \"12\"}\n\n\ndef test_config_rejects_non_dict():\n with pytest.raises(TypeError):\n Config([(\"a\", 1)])\n"}}}
|
| 2 |
+
{"instance_id": "swedl__ttlcache-1", "repo": "swedl/ttlcache", "base_commit": "7c1d3594fec009ac3cf7ef48506ec6d3b0d56132", "patch": "diff --git a/ttlcache/cache.py b/ttlcache/cache.py\nindex fda1724..4c36ad4 100644\n--- a/ttlcache/cache.py\n+++ b/ttlcache/cache.py\n@@ -28,8 +28,15 @@ class TTLCache:\n def _expired(deadline, now):\n return now >= deadline\n \n+ def _purge(self, now):\n+ \"\"\"Drop every expired entry so it can never be counted or evicted.\"\"\"\n+ stale = [key for key, (_, deadline) in self._data.items() if self._expired(deadline, now)]\n+ for key in stale:\n+ del self._data[key]\n+\n def __len__(self):\n- return len(self._data)\n+ now = self._now()\n+ return sum(1 for _, deadline in self._data.values() if not self._expired(deadline, now))\n \n def __contains__(self, key):\n if key not in self._data:\n@@ -39,6 +46,9 @@ class TTLCache:\n \n def __setitem__(self, key, value):\n now = self._now()\n+ self._purge(now)\n+ if key in self._data:\n+ del self._data[key]\n self._data[key] = (value, now + self.ttl)\n while len(self._data) > self.maxsize:\n self._data.popitem(last=False)\n@@ -48,6 +58,7 @@ class TTLCache:\n if self._expired(deadline, self._now()):\n del self._data[key]\n raise KeyError(key)\n+ self._data.move_to_end(key)\n return value\n \n def get(self, key, default=None):\n", "test_patch": "diff --git a/tests/test_lru_semantics.py b/tests/test_lru_semantics.py\nnew file mode 100644\nindex 0000000..1563011\n--- /dev/null\n+++ b/tests/test_lru_semantics.py\n@@ -0,0 +1,55 @@\n+from ttlcache import FakeClock, TTLCache\n+\n+\n+def test_get_refreshes_recency():\n+ clock = FakeClock()\n+ cache = TTLCache(maxsize=2, ttl=10.0, clock=clock)\n+ cache[\"a\"] = 1\n+ cache[\"b\"] = 2\n+ assert cache[\"a\"] == 1\n+ cache[\"c\"] = 3\n+ assert \"a\" in cache\n+ assert \"b\" not in cache\n+\n+\n+def test_overwrite_refreshes_recency():\n+ clock = FakeClock()\n+ cache = TTLCache(maxsize=2, ttl=10.0, clock=clock)\n+ cache[\"a\"] = 1\n+ cache[\"b\"] = 2\n+ cache[\"a\"] = 9\n+ cache[\"c\"] = 3\n+ assert cache[\"a\"] == 9\n+ assert \"b\" not in cache\n+\n+\n+def test_setdefault_refreshes_recency():\n+ clock = FakeClock()\n+ cache = TTLCache(maxsize=2, ttl=10.0, clock=clock)\n+ cache[\"a\"] = 1\n+ cache[\"b\"] = 2\n+ assert cache.setdefault(\"a\", lambda: 99) == 1\n+ cache[\"c\"] = 3\n+ assert \"a\" in cache\n+ assert \"b\" not in cache\n+\n+\n+def test_len_ignores_expired_entries():\n+ clock = FakeClock()\n+ cache = TTLCache(maxsize=4, ttl=10.0, clock=clock)\n+ cache[\"a\"] = 1\n+ clock.advance(5)\n+ cache[\"b\"] = 2\n+ clock.advance(6)\n+ assert len(cache) == 1\n+ assert \"a\" not in cache\n+ assert \"b\" in cache\n+\n+\n+def test_items_reports_least_recently_used_first():\n+ clock = FakeClock()\n+ cache = TTLCache(maxsize=3, ttl=10.0, clock=clock)\n+ cache[\"a\"] = 1\n+ cache[\"b\"] = 2\n+ assert cache[\"a\"] == 1\n+ assert list(cache.items()) == [(\"b\", 2), (\"a\", 1)]\n", "problem_statement": "LRU eviction ignores reads, and len() counts entries that already expired.\n\nOur request cache is a `TTLCache(maxsize=N, ttl=...)` and it is evicting the wrong\nkeys under load:\n\n1. Reading a key does not protect it. With `maxsize=2`, writing `a`, `b`, reading\n `a`, then writing `c` drops `a` and keeps `b` -- the exact opposite of LRU. The\n same happens when a key is overwritten: the new write keeps the old position\n instead of becoming the most recently used entry.\n\n2. `len(cache)` returns the number of stored rows, including rows whose TTL already\n elapsed. Our metrics dashboard reports a cache that never shrinks even though\n every read says the key is gone.\n\n3. `items()` order does not reflect access order for the same reason.\n\n`README.md` documents the intended contract (sections 2-5): reads and overwrites\nmark an entry most-recently-used, expired entries are purged before eviction\nhappens, and `len()` counts live entries only. `ttlcache/cache.py` is the file to\nfix; `ttlcache/clock.py` already gives you a deterministic `FakeClock`.\n\nBe careful that expiry still deletes the entry on read and that `get()` keeps\nreturning its default for missing keys.", "hints_text": "", "created_at": "2026-10-30", "version": "1.0", "environment_setup_commit": "7c1d3594fec009ac3cf7ef48506ec6d3b0d56132", "FAIL_TO_PASS": "[\"tests/test_lru_semantics.py::test_get_refreshes_recency\", \"tests/test_lru_semantics.py::test_items_reports_least_recently_used_first\", \"tests/test_lru_semantics.py::test_len_ignores_expired_entries\", \"tests/test_lru_semantics.py::test_overwrite_refreshes_recency\", \"tests/test_lru_semantics.py::test_setdefault_refreshes_recency\"]", "PASS_TO_PASS": "[\"tests/test_cache.py::test_entry_expires\", \"tests/test_cache.py::test_entry_survives_before_deadline\", \"tests/test_cache.py::test_eviction_order_on_insert\", \"tests/test_cache.py::test_getitem_raises_for_missing_key\", \"tests/test_cache.py::test_invalid_arguments\", \"tests/test_cache.py::test_items_skips_expired\", \"tests/test_cache.py::test_missing_key_returns_default\", \"tests/test_cache.py::test_overwrite_updates_value\", \"tests/test_cache.py::test_pop_removes_entry\", \"tests/test_cache.py::test_repr_reports_live_count\", \"tests/test_cache.py::test_set_and_get\", \"tests/test_cache.py::test_setdefault_uses_factory_once\"]", "dockerless": {"language": "python", "python_requires": ">=3.9", "install": [], "run_from": ".", "test_cmd": "python -m pytest -v --tb=no -p no:cacheprovider {tests}", "test_files": ["tests/test_cache.py", "tests/test_lru_semantics.py"], "files": {"README.md": "# ttlcache\n\nA bounded cache with per-entry time-to-live and LRU eviction. The cache never\nreads the system clock directly; pass a `clock` object exposing `now()` (see\n`ttlcache.clock.FakeClock`) so expiry is deterministic under test.\n\n## Contract\n\n1. **Expiry.** An entry written at time `t` is live while `now < t + ttl` and\n expired once `now >= t + ttl`. Reading an expired key deletes it and raises\n `KeyError` (`get` returns its default instead). `key in cache` is `False` for\n expired entries.\n2. **Recency.** `__getitem__`, `get` and `setdefault` mark the entry as most\n recently used. Writing an existing key also moves it to the most-recently-used\n end *and* restarts its TTL.\n3. **Eviction.** When a write would exceed `maxsize`, expired entries are dropped\n first; if the cache is still full, the least recently used live entry is\n dropped.\n4. **`len(cache)`** counts live entries only. Expired-but-not-yet-collected\n entries must not inflate the count.\n5. **`items()`** yields live pairs in LRU order (least recently used first).\n\n## Usage\n\n```python\nfrom ttlcache import FakeClock, TTLCache\n\nclock = FakeClock()\ncache = TTLCache(maxsize=2, ttl=10.0, clock=clock)\ncache[\"a\"] = 1\ncache[\"b\"] = 2\ncache[\"a\"] # 'a' is now most recently used\ncache[\"c\"] = 3 # evicts 'b', not 'a'\n```\n\n## Tests\n\n```\npython -m pytest -q\n```\n", "tests/test_cache.py": "import pytest\n\nfrom ttlcache import FakeClock, TTLCache\n\n\ndef make(clock, maxsize=3, ttl=10.0):\n return TTLCache(maxsize=maxsize, ttl=ttl, clock=clock)\n\n\ndef test_set_and_get():\n clock = FakeClock()\n cache = make(clock)\n cache[\"a\"] = 1\n assert cache[\"a\"] == 1\n assert cache.get(\"a\") == 1\n\n\ndef test_missing_key_returns_default():\n cache = make(FakeClock())\n assert cache.get(\"nope\") is None\n assert cache.get(\"nope\", 7) == 7\n\n\ndef test_getitem_raises_for_missing_key():\n cache = make(FakeClock())\n with pytest.raises(KeyError):\n cache[\"nope\"]\n\n\ndef test_entry_expires():\n clock = FakeClock()\n cache = make(clock)\n cache[\"a\"] = 1\n clock.advance(10)\n assert \"a\" not in cache\n assert cache.get(\"a\") is None\n with pytest.raises(KeyError):\n cache[\"a\"]\n\n\ndef test_entry_survives_before_deadline():\n clock = FakeClock()\n cache = make(clock)\n cache[\"a\"] = 1\n clock.advance(9.5)\n assert cache[\"a\"] == 1\n\n\ndef test_eviction_order_on_insert():\n clock = FakeClock()\n cache = make(clock, maxsize=2)\n cache[\"a\"] = 1\n cache[\"b\"] = 2\n cache[\"c\"] = 3\n assert \"a\" not in cache\n assert \"b\" in cache and \"c\" in cache\n\n\ndef test_overwrite_updates_value():\n clock = FakeClock()\n cache = make(clock)\n cache[\"a\"] = 1\n cache[\"a\"] = 2\n assert cache[\"a\"] == 2\n assert len(cache) == 1\n\n\ndef test_setdefault_uses_factory_once():\n clock = FakeClock()\n cache = make(clock)\n calls = []\n\n def factory():\n calls.append(1)\n return \"v\"\n\n assert cache.setdefault(\"k\", factory) == \"v\"\n assert cache.setdefault(\"k\", factory) == \"v\"\n assert len(calls) == 1\n\n\ndef test_pop_removes_entry():\n clock = FakeClock()\n cache = make(clock)\n cache[\"a\"] = 1\n assert cache.pop(\"a\") == 1\n assert cache.pop(\"a\", \"gone\") == \"gone\"\n\n\ndef test_items_skips_expired():\n clock = FakeClock()\n cache = make(clock)\n cache[\"a\"] = 1\n clock.advance(4)\n cache[\"b\"] = 2\n clock.advance(7)\n assert list(cache.items()) == [(\"b\", 2)]\n\n\ndef test_invalid_arguments():\n with pytest.raises(ValueError):\n TTLCache(maxsize=0)\n with pytest.raises(ValueError):\n TTLCache(ttl=0)\n\n\ndef test_repr_reports_live_count():\n clock = FakeClock()\n cache = make(clock)\n cache[\"a\"] = 1\n assert \"live=1\" in repr(cache)\n", "ttlcache/__init__.py": "\"\"\"ttlcache - a bounded TTL + LRU cache.\"\"\"\n\nfrom .cache import TTLCache\nfrom .clock import FakeClock, SystemClock\n\n__version__ = \"0.3.0\"\n\n__all__ = [\"FakeClock\", \"SystemClock\", \"TTLCache\"]\n", "ttlcache/cache.py": "\"\"\"A bounded cache with per-entry time-to-live and LRU eviction.\"\"\"\n\nfrom collections import OrderedDict\n\n\nclass TTLCache:\n \"\"\"Fixed-capacity cache where every entry expires ``ttl`` seconds after write.\n\n Entries are kept in least-recently-used order. When the cache is full the\n least recently used entry is dropped first. See README.md for the full\n contract.\n \"\"\"\n\n def __init__(self, maxsize=128, ttl=60.0, clock=None):\n if maxsize <= 0:\n raise ValueError(\"maxsize must be positive\")\n if ttl <= 0:\n raise ValueError(\"ttl must be positive\")\n self.maxsize = maxsize\n self.ttl = float(ttl)\n self._clock = clock\n self._data = OrderedDict()\n\n def _now(self):\n return self._clock.now() if self._clock is not None else 0.0\n\n @staticmethod\n def _expired(deadline, now):\n return now >= deadline\n\n def __len__(self):\n return len(self._data)\n\n def __contains__(self, key):\n if key not in self._data:\n return False\n _, deadline = self._data[key]\n return not self._expired(deadline, self._now())\n\n def __setitem__(self, key, value):\n now = self._now()\n self._data[key] = (value, now + self.ttl)\n while len(self._data) > self.maxsize:\n self._data.popitem(last=False)\n\n def __getitem__(self, key):\n value, deadline = self._data[key]\n if self._expired(deadline, self._now()):\n del self._data[key]\n raise KeyError(key)\n return value\n\n def get(self, key, default=None):\n try:\n return self[key]\n except KeyError:\n return default\n\n def setdefault(self, key, factory):\n \"\"\"Return the live value for ``key``, computing and storing it if absent.\"\"\"\n if key in self:\n return self[key]\n value = factory()\n self[key] = value\n return value\n\n def pop(self, key, default=None):\n if key not in self._data:\n return default\n value, _ = self._data.pop(key)\n return value\n\n def items(self):\n \"\"\"Iterate over live ``(key, value)`` pairs in LRU order.\"\"\"\n now = self._now()\n return [(k, v) for k, (v, d) in self._data.items() if not self._expired(d, now)]\n\n def __repr__(self):\n return \"TTLCache(maxsize=%d, ttl=%r, live=%d)\" % (\n self.maxsize,\n self.ttl,\n len(self),\n )\n", "ttlcache/clock.py": "\"\"\"Clocks used by ttlcache.\n\nThe cache never calls ``time`` directly: it asks a clock object for the current\ntime, which makes expiry deterministic in tests.\n\"\"\"\n\nimport time\n\n\nclass SystemClock:\n \"\"\"Monotonic wall clock for production use.\"\"\"\n\n def now(self):\n return time.monotonic()\n\n\nclass FakeClock:\n \"\"\"A clock that only moves when you tell it to.\"\"\"\n\n def __init__(self, start=0.0):\n self._now = float(start)\n\n def now(self):\n return self._now\n\n def advance(self, seconds):\n self._now += float(seconds)\n return self._now\n"}}}
|
| 3 |
+
{"instance_id": "swedl__cronlite-1", "repo": "swedl/cronlite", "base_commit": "cb8cd4f111a02d9358774debe834b530bcbfeb94", "patch": "diff --git a/cronlite/fields.py b/cronlite/fields.py\nindex 4945d5e..b3c0eb6 100644\n--- a/cronlite/fields.py\n+++ b/cronlite/fields.py\n@@ -1,13 +1,19 @@\n \"\"\"Parsing of individual cron fields into sets of integers.\"\"\"\n \n \n+def _sunday_alias(value):\n+ \"\"\"Collapse the cron Sunday alias 7 onto 0.\"\"\"\n+ return 0 if value == 7 else value\n+\n+\n class FieldSpec:\n \"\"\"Describes one of the five cron fields.\"\"\"\n \n- def __init__(self, name, minimum, maximum):\n+ def __init__(self, name, minimum, maximum, normalize=None):\n self.name = name\n self.minimum = minimum\n self.maximum = maximum\n+ self.normalize = normalize\n \n def _check(self, value):\n if value < self.minimum or value > self.maximum:\n@@ -23,6 +29,8 @@ class FieldSpec:\n values |= self._parse_part(part)\n if not values:\n raise ValueError(\"empty %s field\" % self.name)\n+ if self.normalize is not None:\n+ values = {self.normalize(value) for value in values}\n return values\n \n def _parse_part(self, part):\n@@ -42,7 +50,7 @@ class FieldSpec:\n start, stop = self._parse_bounds(base)\n else:\n start = self._check(int(base))\n- stop = start\n+ stop = self.maximum\n return set(range(start, stop + 1, step))\n if \"-\" in part:\n start, stop = self._parse_bounds(part)\n@@ -66,7 +74,7 @@ MINUTE = FieldSpec(\"minute\", 0, 59)\n HOUR = FieldSpec(\"hour\", 0, 23)\n DAY_OF_MONTH = FieldSpec(\"day_of_month\", 1, 31)\n MONTH = FieldSpec(\"month\", 1, 12)\n-DAY_OF_WEEK = FieldSpec(\"day_of_week\", 0, 6)\n+DAY_OF_WEEK = FieldSpec(\"day_of_week\", 0, 7, normalize=_sunday_alias)\n \n \n def parse_field(spec, text):\n", "test_patch": "diff --git a/tests/test_step_and_sunday.py b/tests/test_step_and_sunday.py\nnew file mode 100644\nindex 0000000..7529ebc\n--- /dev/null\n+++ b/tests/test_step_and_sunday.py\n@@ -0,0 +1,43 @@\n+from datetime import datetime\n+\n+from cronlite import CronExpression, next_after\n+from cronlite.fields import DAY_OF_WEEK, MINUTE, parse_field\n+\n+\n+def test_single_value_with_step_runs_to_end_of_range():\n+ assert parse_field(MINUTE, \"5/15\") == {5, 20, 35, 50}\n+\n+\n+def test_single_value_with_step_stops_at_field_maximum():\n+ assert parse_field(MINUTE, \"58/5\") == {58}\n+\n+\n+def test_next_after_with_stepped_start():\n+ expr = CronExpression(\"5/20 * * * *\")\n+ assert next_after(expr, datetime(2026, 3, 4, 0, 6)) == datetime(2026, 3, 4, 0, 25)\n+\n+\n+def test_sunday_may_be_written_as_seven():\n+ assert parse_field(DAY_OF_WEEK, \"7\") == {0}\n+\n+\n+def test_zero_and_seven_are_equivalent():\n+ assert parse_field(DAY_OF_WEEK, \"0\") == parse_field(DAY_OF_WEEK, \"7\")\n+\n+\n+def test_weekend_range_ending_in_seven():\n+ assert parse_field(DAY_OF_WEEK, \"6-7\") == {6, 0}\n+\n+\n+def test_seven_inside_a_list():\n+ assert parse_field(DAY_OF_WEEK, \"1,7\") == {1, 0}\n+\n+\n+def test_next_after_sunday_seven():\n+ expr = CronExpression(\"0 12 * * 7\")\n+ assert next_after(expr, datetime(2026, 3, 4, 0, 0)) == datetime(2026, 3, 8, 12, 0)\n+\n+\n+def test_weekend_range_next_run():\n+ expr = CronExpression(\"0 9 * * 6-7\")\n+ assert next_after(expr, datetime(2026, 3, 6, 0, 0)) == datetime(2026, 3, 7, 9, 0)\n", "problem_statement": "`5/15` in the minute field only fires once an hour, and day-of-week 7 is rejected.\n\nTwo field-parsing bugs, both in `cronlite/fields.py`:\n\n1. A single value followed by a step is treated as a one-element set. `5/15`\n parses to `{5}` instead of `{5, 20, 35, 50}`, so a job scheduled at\n `5/20 * * * *` runs at 00:05, 01:05, 02:05 instead of every twenty minutes\n starting at :05. `*/20` and `1-10/3` are correct -- only the `<value>/<step>`\n form is wrong, because the upper bound collapses to the start value.\n\n2. Sunday cannot be written as 7. `0 12 * * 7` raises\n `ValueError: day_of_week value out of range: 7` at construction time, and so do\n `7`, `1,7` and `6-7`. Standard cron accepts 0-7 for that field with both 0 and 7\n meaning Sunday. The README's \"Day of week\" section asks for the 7 alias to be\n accepted everywhere a value is accepted and collapsed onto 0, so that a parsed\n day-of-week set is always expressed on the 0-6 scale (`6-7` -> `{6, 0}`).\n\nEverything else about the field syntax must keep working, including the existing\nrange checks and the descending-range rejection.", "hints_text": "", "created_at": "2026-10-30", "version": "1.0", "environment_setup_commit": "cb8cd4f111a02d9358774debe834b530bcbfeb94", "FAIL_TO_PASS": "[\"tests/test_step_and_sunday.py::test_next_after_sunday_seven\", \"tests/test_step_and_sunday.py::test_next_after_with_stepped_start\", \"tests/test_step_and_sunday.py::test_seven_inside_a_list\", \"tests/test_step_and_sunday.py::test_single_value_with_step_runs_to_end_of_range\", \"tests/test_step_and_sunday.py::test_sunday_may_be_written_as_seven\", \"tests/test_step_and_sunday.py::test_weekend_range_ending_in_seven\", \"tests/test_step_and_sunday.py::test_weekend_range_next_run\", \"tests/test_step_and_sunday.py::test_zero_and_seven_are_equivalent\"]", "PASS_TO_PASS": "[\"tests/test_basic.py::test_bad_step_rejected\", \"tests/test_basic.py::test_cron_weekday_mapping\", \"tests/test_basic.py::test_descending_range_rejected\", \"tests/test_basic.py::test_invalid_field_count\", \"tests/test_basic.py::test_list_field\", \"tests/test_basic.py::test_list_with_range\", \"tests/test_basic.py::test_matches_every_minute\", \"tests/test_basic.py::test_matches_rejects_wrong_minute\", \"tests/test_basic.py::test_next_after_accepts_string\", \"tests/test_basic.py::test_next_after_basic\", \"tests/test_basic.py::test_next_after_is_strict\", \"tests/test_basic.py::test_next_after_rolls_to_next_day\", \"tests/test_basic.py::test_out_of_range_rejected\", \"tests/test_basic.py::test_range_field\", \"tests/test_basic.py::test_range_step_field\", \"tests/test_basic.py::test_single_value_field\", \"tests/test_basic.py::test_star_field\", \"tests/test_basic.py::test_star_step_field\", \"tests/test_basic.py::test_str_roundtrip\", \"tests/test_basic.py::test_weekday_field_within_zero_to_six\", \"tests/test_step_and_sunday.py::test_single_value_with_step_stops_at_field_maximum\"]", "dockerless": {"language": "python", "python_requires": ">=3.9", "install": [], "run_from": ".", "test_cmd": "python -m pytest -v --tb=no -p no:cacheprovider {tests}", "test_files": ["tests/test_basic.py", "tests/test_step_and_sunday.py"], "files": {"cronlite/__init__.py": "\"\"\"cronlite - a dependency-free five-field cron expression evaluator.\"\"\"\n\nfrom .expr import CronExpression, cron_weekday, next_after\nfrom .fields import (\n DAY_OF_MONTH,\n DAY_OF_WEEK,\n HOUR,\n MINUTE,\n MONTH,\n FieldSpec,\n parse_field,\n)\n\n__version__ = \"0.2.3\"\n\n__all__ = [\n \"CronExpression\",\n \"DAY_OF_MONTH\",\n \"DAY_OF_WEEK\",\n \"FieldSpec\",\n \"HOUR\",\n \"MINUTE\",\n \"MONTH\",\n \"cron_weekday\",\n \"next_after\",\n \"parse_field\",\n]\n", "cronlite/expr.py": "\"\"\"A five-field cron expression evaluator.\"\"\"\n\nfrom datetime import timedelta\n\nfrom .fields import DAY_OF_MONTH, DAY_OF_WEEK, HOUR, MINUTE, MONTH, parse_field\n\nSEARCH_LIMIT_DAYS = 366\n\n\ndef cron_weekday(moment):\n \"\"\"Return the cron weekday of ``moment``: 0 = Sunday, 6 = Saturday.\"\"\"\n return (moment.weekday() + 1) % 7\n\n\nclass CronExpression:\n \"\"\"A parsed ``minute hour day-of-month month day-of-week`` expression.\"\"\"\n\n def __init__(self, expression):\n parts = expression.split()\n if len(parts) != 5:\n raise ValueError(\"expected 5 fields, got %d\" % (len(parts),))\n self.expression = \" \".join(parts)\n self.minutes = parse_field(MINUTE, parts[0])\n self.hours = parse_field(HOUR, parts[1])\n self.days = parse_field(DAY_OF_MONTH, parts[2])\n self.months = parse_field(MONTH, parts[3])\n self.weekdays = parse_field(DAY_OF_WEEK, parts[4])\n\n def matches(self, moment):\n return (\n moment.minute in self.minutes\n and moment.hour in self.hours\n and moment.day in self.days\n and moment.month in self.months\n and cron_weekday(moment) in self.weekdays\n )\n\n def next_after(self, moment):\n \"\"\"Return the first matching minute strictly after ``moment``.\"\"\"\n current = moment.replace(second=0, microsecond=0) + timedelta(minutes=1)\n limit = current + timedelta(days=SEARCH_LIMIT_DAYS)\n while current <= limit:\n if self.matches(current):\n return current\n current += timedelta(minutes=1)\n raise ValueError(\"no matching time within %d days\" % SEARCH_LIMIT_DAYS)\n\n def __str__(self):\n return self.expression\n\n def __repr__(self):\n return \"CronExpression(%r)\" % (self.expression,)\n\n\ndef next_after(expression, moment):\n \"\"\"Convenience wrapper: ``next_after(\"*/5 * * * *\", moment)``.\"\"\"\n if not isinstance(expression, CronExpression):\n expression = CronExpression(expression)\n return expression.next_after(moment)\n", "cronlite/fields.py": "\"\"\"Parsing of individual cron fields into sets of integers.\"\"\"\n\n\nclass FieldSpec:\n \"\"\"Describes one of the five cron fields.\"\"\"\n\n def __init__(self, name, minimum, maximum):\n self.name = name\n self.minimum = minimum\n self.maximum = maximum\n\n def _check(self, value):\n if value < self.minimum or value > self.maximum:\n raise ValueError(\"%s value out of range: %d\" % (self.name, value))\n return value\n\n def parse(self, text):\n values = set()\n for part in text.split(\",\"):\n part = part.strip()\n if not part:\n raise ValueError(\"empty %s component\" % self.name)\n values |= self._parse_part(part)\n if not values:\n raise ValueError(\"empty %s field\" % self.name)\n return values\n\n def _parse_part(self, part):\n if part == \"*\":\n return set(range(self.minimum, self.maximum + 1))\n if \"/\" in part:\n base, _, step_text = part.partition(\"/\")\n try:\n step = int(step_text)\n except ValueError:\n raise ValueError(\"invalid step in %s field: %r\" % (self.name, part))\n if step <= 0:\n raise ValueError(\"step must be positive in %s field\" % self.name)\n if base == \"*\":\n start, stop = self.minimum, self.maximum\n elif \"-\" in base:\n start, stop = self._parse_bounds(base)\n else:\n start = self._check(int(base))\n stop = start\n return set(range(start, stop + 1, step))\n if \"-\" in part:\n start, stop = self._parse_bounds(part)\n return set(range(start, stop + 1))\n return {self._check(int(part))}\n\n def _parse_bounds(self, text):\n low, _, high = text.partition(\"-\")\n try:\n start, stop = int(low), int(high)\n except ValueError:\n raise ValueError(\"invalid range in %s field: %r\" % (self.name, text))\n self._check(start)\n self._check(stop)\n if stop < start:\n raise ValueError(\"descending range in %s field: %r\" % (self.name, text))\n return start, stop\n\n\nMINUTE = FieldSpec(\"minute\", 0, 59)\nHOUR = FieldSpec(\"hour\", 0, 23)\nDAY_OF_MONTH = FieldSpec(\"day_of_month\", 1, 31)\nMONTH = FieldSpec(\"month\", 1, 12)\nDAY_OF_WEEK = FieldSpec(\"day_of_week\", 0, 6)\n\n\ndef parse_field(spec, text):\n \"\"\"Parse ``text`` against ``spec`` and return the matching values.\"\"\"\n return spec.parse(text)\n", "README.md": "# cronlite\n\nA dependency-free evaluator for standard five-field cron expressions:\n\n```\nminute hour day-of-month month day-of-week\n```\n\n## Field syntax\n\n| Form | Meaning |\n| --- | --- |\n| `*` | every value in the field range |\n| `7` | a single value |\n| `1-5` | an inclusive range |\n| `*/15` | every 15th value across the whole range |\n| `1-10/3` | every 3rd value inside the range (`1, 4, 7, 10`) |\n| `5/15` | every 15th value starting at 5 and **running to the end of the range** (`5, 20, 35, 50`) |\n| `0,15,30` | a list; any of the above forms may be combined in a list |\n\nField ranges: minute `0-59`, hour `0-23`, day-of-month `1-31`, month `1-12`.\n\n## Day of week\n\nThe day-of-week field is `0-7` where **both `0` and `7` mean Sunday** and\n`1-6` mean Monday to Saturday. `7` must be accepted anywhere a value is\naccepted, including inside ranges such as `6-7` (Saturday and Sunday), and it\nmust behave exactly like `0`. Internally a parsed day-of-week set is expressed\non the `0-6` scale, so `parse_field(DAY_OF_WEEK, \"7\") == {0}` and\n`parse_field(DAY_OF_WEEK, \"6-7\") == {6, 0}`.\n\n## Matching\n\n`CronExpression.matches(moment)` is true when all five fields match. The\nday-of-month and day-of-week fields are combined with AND (not the traditional\nOR). `next_after(moment)` returns the first matching minute strictly after\n`moment`, scanning forward up to 366 days.\n\n## Tests\n\n```\npython -m pytest -q\n```\n", "tests/test_basic.py": "from datetime import datetime\n\nimport pytest\n\nfrom cronlite import CronExpression, cron_weekday, next_after\nfrom cronlite.fields import DAY_OF_WEEK, HOUR, MINUTE, parse_field\n\n\ndef test_star_field():\n assert parse_field(MINUTE, \"*\") == set(range(60))\n\n\ndef test_single_value_field():\n assert parse_field(MINUTE, \"17\") == {17}\n\n\ndef test_list_field():\n assert parse_field(MINUTE, \"0,15,30,45\") == {0, 15, 30, 45}\n\n\ndef test_list_with_range():\n assert parse_field(MINUTE, \"0,10-12\") == {0, 10, 11, 12}\n\n\ndef test_range_field():\n assert parse_field(HOUR, \"9-11\") == {9, 10, 11}\n\n\ndef test_star_step_field():\n assert parse_field(MINUTE, \"*/15\") == {0, 15, 30, 45}\n\n\ndef test_range_step_field():\n assert parse_field(MINUTE, \"1-10/3\") == {1, 4, 7, 10}\n\n\ndef test_out_of_range_rejected():\n with pytest.raises(ValueError):\n parse_field(MINUTE, \"60\")\n with pytest.raises(ValueError):\n parse_field(HOUR, \"24\")\n\n\ndef test_bad_step_rejected():\n with pytest.raises(ValueError):\n parse_field(MINUTE, \"*/0\")\n\n\ndef test_descending_range_rejected():\n with pytest.raises(ValueError):\n parse_field(MINUTE, \"10-1\")\n\n\ndef test_weekday_field_within_zero_to_six():\n assert parse_field(DAY_OF_WEEK, \"1-5\") == {1, 2, 3, 4, 5}\n\n\ndef test_cron_weekday_mapping():\n assert cron_weekday(datetime(2026, 3, 4)) == 3\n assert cron_weekday(datetime(2026, 3, 8)) == 0\n\n\ndef test_matches_every_minute():\n expr = CronExpression(\"* * * * *\")\n assert expr.matches(datetime(2026, 3, 4, 5, 6))\n\n\ndef test_matches_rejects_wrong_minute():\n expr = CronExpression(\"30 9 * * *\")\n assert not expr.matches(datetime(2026, 3, 4, 9, 31))\n\n\ndef test_next_after_basic():\n expr = CronExpression(\"30 9 * * *\")\n assert next_after(expr, datetime(2026, 3, 4, 5, 6)) == datetime(2026, 3, 4, 9, 30)\n\n\ndef test_next_after_is_strict():\n expr = CronExpression(\"30 9 * * *\")\n assert next_after(expr, datetime(2026, 3, 4, 9, 30)) == datetime(2026, 3, 5, 9, 30)\n\n\ndef test_next_after_rolls_to_next_day():\n expr = CronExpression(\"0 0 * * *\")\n assert next_after(expr, datetime(2026, 3, 4, 0, 0)) == datetime(2026, 3, 5, 0, 0)\n\n\ndef test_next_after_accepts_string():\n assert next_after(\"*/30 * * * *\", datetime(2026, 3, 4, 5, 6)) == datetime(2026, 3, 4, 5, 30)\n\n\ndef test_invalid_field_count():\n with pytest.raises(ValueError):\n CronExpression(\"* * * *\")\n\n\ndef test_str_roundtrip():\n assert str(CronExpression(\" 0 12 * * * \")) == \"0 12 * * *\"\n"}}}
|
| 4 |
+
{"instance_id": "swedl__semverlite-1", "repo": "swedl/semverlite", "base_commit": "248819bc3fd540992754e15ab48fcba834ff11d4", "patch": "diff --git a/semverlite/range.py b/semverlite/range.py\nindex abe2ece..ad1fc9e 100644\n--- a/semverlite/range.py\n+++ b/semverlite/range.py\n@@ -40,7 +40,16 @@ class VersionRange:\n raise ValueError(\"empty range\")\n self.clauses = [_parse_clause(part) for part in parts]\n \n+ def _allows_prerelease(self, version):\n+ \"\"\"A prerelease needs an explicit opt-in on the same release triple.\"\"\"\n+ for _, target in self.clauses:\n+ if target.prerelease and target.core == version.core:\n+ return True\n+ return False\n+\n def allows(self, version):\n+ if version.prerelease and not self._allows_prerelease(version):\n+ return False\n return all(self._allows_clause(op, target, version) for op, target in self.clauses)\n \n @staticmethod\ndiff --git a/semverlite/version.py b/semverlite/version.py\nindex bbb779c..4d498ed 100644\n--- a/semverlite/version.py\n+++ b/semverlite/version.py\n@@ -9,6 +9,18 @@ _VERSION_RE = re.compile(\n )\n \n \n+def _identifier_key(identifier):\n+ \"\"\"Sort key for one prerelease identifier.\n+\n+ Numeric identifiers sort numerically and below alphanumeric ones; the third\n+ element keeps keys comparable when the first two tie.\n+ \"\"\"\n+ text = str(identifier)\n+ if text.isdigit():\n+ return (0, int(text), \"\")\n+ return (1, 0, text)\n+\n+\n class Version:\n \"\"\"A semantic version: major.minor.patch[-prerelease][+build].\"\"\"\n \n@@ -43,8 +55,8 @@ class Version:\n \"\"\"Sort key implementing semver precedence (build metadata ignored).\"\"\"\n core = self.core\n if not self.prerelease:\n- return core + (1, ())\n- return core + (0, tuple(str(part) for part in self.prerelease))\n+ return core + ((1,),)\n+ return core + ((0, tuple(_identifier_key(part) for part in self.prerelease)),)\n \n def __str__(self):\n text = \"%d.%d.%d\" % self.core\n", "test_patch": "diff --git a/tests/test_prerelease_precedence.py b/tests/test_prerelease_precedence.py\nnew file mode 100644\nindex 0000000..1ba29fb\n--- /dev/null\n+++ b/tests/test_prerelease_precedence.py\n@@ -0,0 +1,55 @@\n+import pytest\n+\n+from semverlite import Version, VersionRange\n+\n+\n+def v(text):\n+ return Version.parse(text)\n+\n+\n+def test_numeric_identifiers_compare_numerically():\n+ assert v(\"1.0.0-alpha.2\") < v(\"1.0.0-alpha.10\")\n+\n+\n+def test_numeric_identifier_is_lower_than_alphanumeric():\n+ assert v(\"1.0.0-1\") < v(\"1.0.0-alpha\")\n+\n+\n+def test_total_order_matches_semver_precedence():\n+ chain = [\n+ \"1.0.0-alpha\",\n+ \"1.0.0-alpha.1\",\n+ \"1.0.0-alpha.beta\",\n+ \"1.0.0-beta\",\n+ \"1.0.0-beta.2\",\n+ \"1.0.0-beta.11\",\n+ \"1.0.0-rc.1\",\n+ \"1.0.0\",\n+ ]\n+ versions = [v(text) for text in chain]\n+ assert versions == sorted(versions)\n+ assert len(set(versions)) == len(versions)\n+\n+\n+def test_prerelease_does_not_satisfy_release_lower_bound():\n+ assert not VersionRange(\">=1.0.0\").allows(v(\"1.0.1-alpha\"))\n+\n+\n+def test_prerelease_does_not_satisfy_caret_range():\n+ assert not VersionRange(\"^1.2.3\").allows(v(\"1.3.0-beta\"))\n+\n+\n+def test_tilde_range_excludes_prerelease():\n+ assert VersionRange(\"~1.2.3\").allows(v(\"1.2.4\"))\n+ assert not VersionRange(\"~1.2.3\").allows(v(\"1.2.4-rc.1\"))\n+\n+\n+def test_range_allows_prerelease_when_same_triple_is_named():\n+ assert VersionRange(\">=1.0.0-alpha\").allows(v(\"1.0.0-beta\"))\n+ assert not VersionRange(\">=1.0.0-alpha\").allows(v(\"1.0.1-beta\"))\n+\n+\n+def test_prerelease_ordering_is_stable_across_comparisons():\n+ assert v(\"2.0.0-beta.9\") > v(\"2.0.0-beta.8\")\n+ assert v(\"2.0.0-beta.10\") > v(\"2.0.0-beta.9\")\n+ assert not v(\"2.0.0-beta.10\") < v(\"2.0.0-beta.9\")\n", "problem_statement": "Prerelease ordering is lexicographic, and prereleases leak into release ranges.\n\nTwo spec violations against semver 2.0.0:\n\n1. Prerelease identifiers are compared as strings. `1.0.0-alpha.2` sorts *after*\n `1.0.0-alpha.10` because `\"2\" > \"1\"`, and `1.0.0-beta.2` sorts after\n `1.0.0-beta.11`. Sorting a release train produces a wrong order. Numeric\n identifiers must compare numerically, a numeric identifier must sort below an\n alphanumeric one, and a shorter prerelease must sort below a longer one that\n shares its prefix.\n\n2. A prerelease satisfies ranges that never opted into prereleases.\n `VersionRange('>=1.0.0').allows(Version.parse('1.0.1-alpha'))` returns True, and\n so does `~1.2.3` with `1.2.4-rc.1`. That makes our upgrade tool offer release\n candidates as if they were final releases. A prerelease may only satisfy a range\n when some clause names a prerelease on the same `major.minor.patch` triple, so\n `>=1.0.0-alpha` does allow `1.0.0-beta` but not `1.0.1-beta`.\n\n`semverlite/version.py` holds precedence, `semverlite/range.py` holds range\nmatching; `README.md` states both rules precisely. Build metadata must stay\nignored for precedence, and equality/hashing must stay consistent with the new\nordering.", "hints_text": "", "created_at": "2026-10-30", "version": "1.0", "environment_setup_commit": "248819bc3fd540992754e15ab48fcba834ff11d4", "FAIL_TO_PASS": "[\"tests/test_prerelease_precedence.py::test_numeric_identifiers_compare_numerically\", \"tests/test_prerelease_precedence.py::test_prerelease_does_not_satisfy_caret_range\", \"tests/test_prerelease_precedence.py::test_prerelease_does_not_satisfy_release_lower_bound\", \"tests/test_prerelease_precedence.py::test_prerelease_ordering_is_stable_across_comparisons\", \"tests/test_prerelease_precedence.py::test_range_allows_prerelease_when_same_triple_is_named\", \"tests/test_prerelease_precedence.py::test_tilde_range_excludes_prerelease\", \"tests/test_prerelease_precedence.py::test_total_order_matches_semver_precedence\"]", "PASS_TO_PASS": "[\"tests/test_prerelease_precedence.py::test_numeric_identifier_is_lower_than_alphanumeric\", \"tests/test_version.py::test_caret_range\", \"tests/test_version.py::test_equality_and_hash\", \"tests/test_version.py::test_equality_ignores_build_metadata\", \"tests/test_version.py::test_exact_clause\", \"tests/test_version.py::test_inequality\", \"tests/test_version.py::test_invalid_clause_rejected\", \"tests/test_version.py::test_invalid_versions_rejected\", \"tests/test_version.py::test_ordering_by_major\", \"tests/test_version.py::test_ordering_by_minor\", \"tests/test_version.py::test_ordering_by_patch\", \"tests/test_version.py::test_parse_basic\", \"tests/test_version.py::test_parse_prerelease_and_build\", \"tests/test_version.py::test_range_lower_and_upper_bound\", \"tests/test_version.py::test_range_membership_operator\", \"tests/test_version.py::test_release_beats_prerelease\", \"tests/test_version.py::test_shorter_prerelease_is_lower\", \"tests/test_version.py::test_str_roundtrip\", \"tests/test_version.py::test_tilde_range\"]", "dockerless": {"language": "python", "python_requires": ">=3.9", "install": [], "run_from": ".", "test_cmd": "python -m pytest -v --tb=no -p no:cacheprovider {tests}", "test_files": ["tests/test_prerelease_precedence.py", "tests/test_version.py"], "files": {"README.md": "# semverlite\n\nSemantic version parsing, precedence and range matching, following\n[semver 2.0.0](https://semver.org/).\n\n## Precedence\n\n`Version` objects sort by semver precedence:\n\n1. Compare `major`, then `minor`, then `patch` numerically.\n2. A version **with** a prerelease is lower than the same version without one\n (`1.0.0-alpha < 1.0.0`).\n3. Prerelease identifiers are compared left to right. A **numeric** identifier\n always compares numerically, so `1.0.0-alpha.2 < 1.0.0-alpha.10`. An\n alphanumeric identifier compares by ASCII order and is always **greater**\n than a numeric identifier of the same position (`1.0.0-1 < 1.0.0-alpha`).\n A longer prerelease wins when all shared identifiers are equal\n (`1.0.0-alpha < 1.0.0-alpha.1`).\n4. Build metadata is ignored for precedence, so `1.0.0+a == 1.0.0+b`.\n\n## Ranges\n\n`VersionRange(text)` accepts a comma-separated conjunction of clauses using\n`=`, `==`, `>`, `>=`, `<`, `<=`, `^` and `~`. `^1.2.3` means\n`>=1.2.3, <2.0.0`; `~1.2.3` means `>=1.2.3, <1.3.0`.\n\n**Prereleases are excluded from a range unless the range itself opts in.** A\nversion carrying a prerelease satisfies the range only when some clause names a\nprerelease with the *same* `major.minor.patch` triple. So `>=1.0.0` does not\nallow `1.0.1-alpha`, while `>=1.0.0-alpha` does allow `1.0.0-beta`.\n\n## Usage\n\n```python\nfrom semverlite import Version, VersionRange\n\nVersion.parse(\"1.2.3-alpha.1\") < Version.parse(\"1.2.3\")\nVersion.parse(\"1.2.3\") in VersionRange(\">=1.0.0, <2.0.0\")\n```\n\n## Tests\n\n```\npython -m pytest -q\n```\n", "semverlite/__init__.py": "\"\"\"semverlite - semantic version parsing, precedence and ranges.\"\"\"\n\nfrom .range import VersionRange\nfrom .version import Version\n\n__version__ = \"0.5.2\"\n\n__all__ = [\"Version\", \"VersionRange\"]\n", "semverlite/range.py": "\"\"\"Version range expressions such as ``\">=1.2.3, <2.0.0\"``.\"\"\"\n\nimport re\n\nfrom .version import Version\n\n_CLAUSE_RE = re.compile(r\"^(>=|<=|==|=|>|<|\\^|~)?\\s*(\\S+)$\")\n\n\ndef _parse_clause(text):\n match = _CLAUSE_RE.match(text)\n if match is None:\n raise ValueError(\"invalid range clause: %r\" % (text,))\n operator = match.group(1) or \"==\"\n return operator, Version.parse(match.group(2))\n\n\ndef _caret_upper(target):\n \"\"\"Upper bound of ``^target``: the next release that breaks the left-most\n non-zero component.\"\"\"\n if target.major > 0:\n return Version(target.major + 1, 0, 0)\n if target.minor > 0:\n return Version(0, target.minor + 1, 0)\n return Version(0, 0, target.patch + 1)\n\n\ndef _tilde_upper(target):\n \"\"\"Upper bound of ``~target``: the next minor release.\"\"\"\n return Version(target.major, target.minor + 1, 0)\n\n\nclass VersionRange:\n \"\"\"A comma-separated list of clauses that must all hold.\"\"\"\n\n def __init__(self, text):\n self.text = text.strip()\n parts = [part.strip() for part in self.text.split(\",\") if part.strip()]\n if not parts:\n raise ValueError(\"empty range\")\n self.clauses = [_parse_clause(part) for part in parts]\n\n def allows(self, version):\n return all(self._allows_clause(op, target, version) for op, target in self.clauses)\n\n @staticmethod\n def _allows_clause(operator, target, version):\n if operator in (\"=\", \"==\"):\n return version == target\n if operator == \">\":\n return version > target\n if operator == \">=\":\n return version >= target\n if operator == \"<\":\n return version < target\n if operator == \"<=\":\n return version <= target\n if operator == \"^\":\n return target <= version < _caret_upper(target)\n if operator == \"~\":\n return target <= version < _tilde_upper(target)\n raise ValueError(\"unknown operator: %r\" % (operator,))\n\n def __contains__(self, version):\n return self.allows(version)\n\n def __str__(self):\n return self.text\n\n def __repr__(self):\n return \"VersionRange(%r)\" % (self.text,)\n", "semverlite/version.py": "\"\"\"Semantic version parsing and precedence.\"\"\"\n\nimport re\n\n_VERSION_RE = re.compile(\n r\"^(?P<major>0|[1-9]\\d*)\\.(?P<minor>0|[1-9]\\d*)\\.(?P<patch>0|[1-9]\\d*)\"\n r\"(?:-(?P<pre>[0-9A-Za-z.-]+))?\"\n r\"(?:\\+(?P<build>[0-9A-Za-z.-]+))?$\"\n)\n\n\nclass Version:\n \"\"\"A semantic version: major.minor.patch[-prerelease][+build].\"\"\"\n\n def __init__(self, major, minor, patch, prerelease=(), build=()):\n self.major = int(major)\n self.minor = int(minor)\n self.patch = int(patch)\n self.prerelease = tuple(prerelease)\n self.build = tuple(build)\n\n @classmethod\n def parse(cls, text):\n match = _VERSION_RE.match(text.strip())\n if match is None:\n raise ValueError(\"invalid version: %r\" % (text,))\n prerelease = tuple(match.group(\"pre\").split(\".\")) if match.group(\"pre\") else ()\n build = tuple(match.group(\"build\").split(\".\")) if match.group(\"build\") else ()\n return cls(\n match.group(\"major\"),\n match.group(\"minor\"),\n match.group(\"patch\"),\n prerelease,\n build,\n )\n\n @property\n def core(self):\n \"\"\"The (major, minor, patch) triple.\"\"\"\n return (self.major, self.minor, self.patch)\n\n def precedence_key(self):\n \"\"\"Sort key implementing semver precedence (build metadata ignored).\"\"\"\n core = self.core\n if not self.prerelease:\n return core + (1, ())\n return core + (0, tuple(str(part) for part in self.prerelease))\n\n def __str__(self):\n text = \"%d.%d.%d\" % self.core\n if self.prerelease:\n text += \"-\" + \".\".join(str(part) for part in self.prerelease)\n if self.build:\n text += \"+\" + \".\".join(str(part) for part in self.build)\n return text\n\n def __repr__(self):\n return \"Version(%r)\" % (str(self),)\n\n def __eq__(self, other):\n if not isinstance(other, Version):\n return NotImplemented\n return self.precedence_key() == other.precedence_key()\n\n def __ne__(self, other):\n result = self.__eq__(other)\n if result is NotImplemented:\n return result\n return not result\n\n def __lt__(self, other):\n return self.precedence_key() < other.precedence_key()\n\n def __le__(self, other):\n return self.precedence_key() <= other.precedence_key()\n\n def __gt__(self, other):\n return self.precedence_key() > other.precedence_key()\n\n def __ge__(self, other):\n return self.precedence_key() >= other.precedence_key()\n\n def __hash__(self):\n return hash(self.precedence_key())\n", "tests/test_version.py": "import pytest\n\nfrom semverlite import Version, VersionRange\n\n\ndef v(text):\n return Version.parse(text)\n\n\ndef test_parse_basic():\n version = v(\"1.2.3\")\n assert version.major == 1\n assert version.minor == 2\n assert version.patch == 3\n assert version.prerelease == ()\n assert version.build == ()\n\n\ndef test_parse_prerelease_and_build():\n version = v(\"1.2.3-rc.1+build.5\")\n assert version.prerelease == (\"rc\", \"1\")\n assert version.build == (\"build\", \"5\")\n assert version.core == (1, 2, 3)\n\n\ndef test_str_roundtrip():\n for text in (\"0.0.1\", \"1.2.3-alpha\", \"1.2.3-alpha.1+build.5\"):\n assert str(v(text)) == text\n\n\ndef test_invalid_versions_rejected():\n for text in (\"1.2\", \"1.2.3.4\", \"v1.2.3\", \"1.2.x\"):\n with pytest.raises(ValueError):\n Version.parse(text)\n\n\ndef test_ordering_by_patch():\n assert v(\"1.2.3\") < v(\"1.2.4\")\n\n\ndef test_ordering_by_minor():\n assert v(\"1.2.9\") < v(\"1.3.0\")\n\n\ndef test_ordering_by_major():\n assert v(\"1.9.9\") < v(\"2.0.0\")\n\n\ndef test_release_beats_prerelease():\n assert v(\"1.0.0-alpha\") < v(\"1.0.0\")\n assert v(\"1.0.0\") > v(\"1.0.0-rc.9\")\n\n\ndef test_shorter_prerelease_is_lower():\n assert v(\"1.0.0-alpha\") < v(\"1.0.0-alpha.1\")\n\n\ndef test_equality_ignores_build_metadata():\n assert v(\"1.0.0+a\") == v(\"1.0.0+b\")\n assert v(\"1.0.0-alpha+x\") == v(\"1.0.0-alpha+y\")\n\n\ndef test_equality_and_hash():\n assert v(\"1.2.3\") == v(\"1.2.3\")\n assert len({v(\"1.2.3\"), v(\"1.2.3\")}) == 1\n\n\ndef test_inequality():\n assert v(\"1.2.3\") != v(\"1.2.4\")\n\n\ndef test_range_lower_and_upper_bound():\n version_range = VersionRange(\">=1.0.0, <2.0.0\")\n assert version_range.allows(v(\"1.0.0\"))\n assert version_range.allows(v(\"1.5.0\"))\n assert not version_range.allows(v(\"2.0.0\"))\n assert not version_range.allows(v(\"0.9.9\"))\n\n\ndef test_range_membership_operator():\n assert v(\"1.5.0\") in VersionRange(\">=1.0.0, <2.0.0\")\n assert v(\"2.5.0\") not in VersionRange(\">=1.0.0, <2.0.0\")\n\n\ndef test_caret_range():\n version_range = VersionRange(\"^1.2.3\")\n assert version_range.allows(v(\"1.2.3\"))\n assert version_range.allows(v(\"1.9.9\"))\n assert not version_range.allows(v(\"2.0.0\"))\n assert not version_range.allows(v(\"1.2.2\"))\n\n\ndef test_tilde_range():\n version_range = VersionRange(\"~1.2.3\")\n assert version_range.allows(v(\"1.2.9\"))\n assert not version_range.allows(v(\"1.3.0\"))\n\n\ndef test_exact_clause():\n assert VersionRange(\"==1.2.3\").allows(v(\"1.2.3\"))\n assert not VersionRange(\"==1.2.3\").allows(v(\"1.2.4\"))\n\n\ndef test_invalid_clause_rejected():\n with pytest.raises(ValueError):\n VersionRange(\">=nope\")\n"}}}
|
build_dataset.py
ADDED
|
@@ -0,0 +1,355 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Build SWE-DockerLess.jsonl from the repos/ fixed/ newtests/ trees.
|
| 2 |
+
|
| 3 |
+
For every instance this script really does the work:
|
| 4 |
+
|
| 5 |
+
1. git-init the base repo and record the commit sha (base_commit);
|
| 6 |
+
2. overlay the gold fix, capture ``git diff`` as the patch, restore;
|
| 7 |
+
3. overlay the new tests, capture ``git diff --cached`` as the test_patch;
|
| 8 |
+
4. materialise a clean checkout, apply test_patch, run pytest -> "before";
|
| 9 |
+
5. apply patch, run pytest -> "after";
|
| 10 |
+
6. derive FAIL_TO_PASS (fail before, pass after) and PASS_TO_PASS
|
| 11 |
+
(pass before, pass after) from the observed outcomes, not from guesses.
|
| 12 |
+
|
| 13 |
+
Run with: python build_dataset.py
|
| 14 |
+
"""
|
| 15 |
+
|
| 16 |
+
import json
|
| 17 |
+
import os
|
| 18 |
+
import re
|
| 19 |
+
import shutil
|
| 20 |
+
import subprocess
|
| 21 |
+
import sys
|
| 22 |
+
import tempfile
|
| 23 |
+
from pathlib import Path
|
| 24 |
+
|
| 25 |
+
ROOT = Path(__file__).resolve().parent
|
| 26 |
+
REPOS = ROOT / "repos"
|
| 27 |
+
FIXED = ROOT / "fixed"
|
| 28 |
+
NEWTESTS = ROOT / "newtests"
|
| 29 |
+
OUT = ROOT / "SWE-DockerLess.jsonl"
|
| 30 |
+
|
| 31 |
+
STAMP = "2026-10-30T12:00:00+0000"
|
| 32 |
+
GIT_ENV = {
|
| 33 |
+
"GIT_AUTHOR_NAME": "swedl",
|
| 34 |
+
"GIT_AUTHOR_EMAIL": "swedl@example.invalid",
|
| 35 |
+
"GIT_COMMITTER_NAME": "swedl",
|
| 36 |
+
"GIT_COMMITTER_EMAIL": "swedl@example.invalid",
|
| 37 |
+
"GIT_AUTHOR_DATE": STAMP,
|
| 38 |
+
"GIT_COMMITTER_DATE": STAMP,
|
| 39 |
+
"GIT_CONFIG_NOSYSTEM": "1",
|
| 40 |
+
}
|
| 41 |
+
GIT_FLAGS = [
|
| 42 |
+
"-c", "core.autocrlf=false",
|
| 43 |
+
"-c", "core.safecrlf=false",
|
| 44 |
+
"-c", "user.name=swedl",
|
| 45 |
+
"-c", "user.email=swedl@example.invalid",
|
| 46 |
+
"-c", "commit.gpgsign=false",
|
| 47 |
+
]
|
| 48 |
+
|
| 49 |
+
PYTEST_CMD = "python -m pytest -v --tb=no -p no:cacheprovider"
|
| 50 |
+
RESULT_RE = re.compile(
|
| 51 |
+
r"^(tests/[^\s:]+::[^\s]+)\s+(PASSED|FAILED|ERROR|SKIPPED|XFAIL|XPASS)"
|
| 52 |
+
)
|
| 53 |
+
|
| 54 |
+
INSTANCES = [
|
| 55 |
+
{
|
| 56 |
+
"instance_id": "swedl__miniconf-1",
|
| 57 |
+
"repo": "swedl/miniconf",
|
| 58 |
+
"name": "miniconf",
|
| 59 |
+
"problem_statement": (
|
| 60 |
+
"Config interpolation blows up on reference cycles and expands escaped text.\n\n"
|
| 61 |
+
"We use miniconf for service configs that are layered from a shared file plus a\n"
|
| 62 |
+
"per-environment overlay. Two problems, both reproducible on main:\n\n"
|
| 63 |
+
"1. A cycle anywhere in the mapping crashes with `RecursionError` instead of the\n"
|
| 64 |
+
" documented error. `Config({'a': '${b}', 'b': '${a}'})['a']` recurses until the\n"
|
| 65 |
+
" interpreter gives up, and the traceback is a thousand frames of `_interpolate`.\n"
|
| 66 |
+
" The README says this must raise `CircularReferenceError` with a message naming\n"
|
| 67 |
+
" the chain, and we rely on that message in our error reporting.\n\n"
|
| 68 |
+
"2. `$$` does not escape. Our TLS config has `cert_subject: \"$${cluster}\"`, meaning\n"
|
| 69 |
+
" the literal text `${cluster}` to be handed to a downstream templating tool, but\n"
|
| 70 |
+
" miniconf substitutes the cluster name into it. The README's escape rule is not\n"
|
| 71 |
+
" implemented at all.\n\n"
|
| 72 |
+
"Both bugs live in `Config._resolve` / `Config._interpolate` in\n"
|
| 73 |
+
"`miniconf/config.py`. `README.md` section \"Interpolation contract\" is the spec we\n"
|
| 74 |
+
"want honoured, including the exact wording of the cycle message and the rule that\n"
|
| 75 |
+
"substituted text is never rescanned for references. A cycle that is unrelated to\n"
|
| 76 |
+
"the key being read must not affect that read, and `as_dict()` must surface cycles\n"
|
| 77 |
+
"like any other read.\n\n"
|
| 78 |
+
"The existing suite passes; the escape and cycle behaviour was never covered."
|
| 79 |
+
),
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"instance_id": "swedl__ttlcache-1",
|
| 83 |
+
"name": "ttlcache",
|
| 84 |
+
"repo": "swedl/ttlcache",
|
| 85 |
+
"problem_statement": (
|
| 86 |
+
"LRU eviction ignores reads, and len() counts entries that already expired.\n\n"
|
| 87 |
+
"Our request cache is a `TTLCache(maxsize=N, ttl=...)` and it is evicting the wrong\n"
|
| 88 |
+
"keys under load:\n\n"
|
| 89 |
+
"1. Reading a key does not protect it. With `maxsize=2`, writing `a`, `b`, reading\n"
|
| 90 |
+
" `a`, then writing `c` drops `a` and keeps `b` -- the exact opposite of LRU. The\n"
|
| 91 |
+
" same happens when a key is overwritten: the new write keeps the old position\n"
|
| 92 |
+
" instead of becoming the most recently used entry.\n\n"
|
| 93 |
+
"2. `len(cache)` returns the number of stored rows, including rows whose TTL already\n"
|
| 94 |
+
" elapsed. Our metrics dashboard reports a cache that never shrinks even though\n"
|
| 95 |
+
" every read says the key is gone.\n\n"
|
| 96 |
+
"3. `items()` order does not reflect access order for the same reason.\n\n"
|
| 97 |
+
"`README.md` documents the intended contract (sections 2-5): reads and overwrites\n"
|
| 98 |
+
"mark an entry most-recently-used, expired entries are purged before eviction\n"
|
| 99 |
+
"happens, and `len()` counts live entries only. `ttlcache/cache.py` is the file to\n"
|
| 100 |
+
"fix; `ttlcache/clock.py` already gives you a deterministic `FakeClock`.\n\n"
|
| 101 |
+
"Be careful that expiry still deletes the entry on read and that `get()` keeps\n"
|
| 102 |
+
"returning its default for missing keys."
|
| 103 |
+
),
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"instance_id": "swedl__cronlite-1",
|
| 107 |
+
"name": "cronlite",
|
| 108 |
+
"repo": "swedl/cronlite",
|
| 109 |
+
"problem_statement": (
|
| 110 |
+
"`5/15` in the minute field only fires once an hour, and day-of-week 7 is rejected.\n\n"
|
| 111 |
+
"Two field-parsing bugs, both in `cronlite/fields.py`:\n\n"
|
| 112 |
+
"1. A single value followed by a step is treated as a one-element set. `5/15`\n"
|
| 113 |
+
" parses to `{5}` instead of `{5, 20, 35, 50}`, so a job scheduled at\n"
|
| 114 |
+
" `5/20 * * * *` runs at 00:05, 01:05, 02:05 instead of every twenty minutes\n"
|
| 115 |
+
" starting at :05. `*/20` and `1-10/3` are correct -- only the `<value>/<step>`\n"
|
| 116 |
+
" form is wrong, because the upper bound collapses to the start value.\n\n"
|
| 117 |
+
"2. Sunday cannot be written as 7. `0 12 * * 7` raises\n"
|
| 118 |
+
" `ValueError: day_of_week value out of range: 7` at construction time, and so do\n"
|
| 119 |
+
" `7`, `1,7` and `6-7`. Standard cron accepts 0-7 for that field with both 0 and 7\n"
|
| 120 |
+
" meaning Sunday. The README's \"Day of week\" section asks for the 7 alias to be\n"
|
| 121 |
+
" accepted everywhere a value is accepted and collapsed onto 0, so that a parsed\n"
|
| 122 |
+
" day-of-week set is always expressed on the 0-6 scale (`6-7` -> `{6, 0}`).\n\n"
|
| 123 |
+
"Everything else about the field syntax must keep working, including the existing\n"
|
| 124 |
+
"range checks and the descending-range rejection."
|
| 125 |
+
),
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"instance_id": "swedl__semverlite-1",
|
| 129 |
+
"name": "semverlite",
|
| 130 |
+
"repo": "swedl/semverlite",
|
| 131 |
+
"problem_statement": (
|
| 132 |
+
"Prerelease ordering is lexicographic, and prereleases leak into release ranges.\n\n"
|
| 133 |
+
"Two spec violations against semver 2.0.0:\n\n"
|
| 134 |
+
"1. Prerelease identifiers are compared as strings. `1.0.0-alpha.2` sorts *after*\n"
|
| 135 |
+
" `1.0.0-alpha.10` because `\"2\" > \"1\"`, and `1.0.0-beta.2` sorts after\n"
|
| 136 |
+
" `1.0.0-beta.11`. Sorting a release train produces a wrong order. Numeric\n"
|
| 137 |
+
" identifiers must compare numerically, a numeric identifier must sort below an\n"
|
| 138 |
+
" alphanumeric one, and a shorter prerelease must sort below a longer one that\n"
|
| 139 |
+
" shares its prefix.\n\n"
|
| 140 |
+
"2. A prerelease satisfies ranges that never opted into prereleases.\n"
|
| 141 |
+
" `VersionRange('>=1.0.0').allows(Version.parse('1.0.1-alpha'))` returns True, and\n"
|
| 142 |
+
" so does `~1.2.3` with `1.2.4-rc.1`. That makes our upgrade tool offer release\n"
|
| 143 |
+
" candidates as if they were final releases. A prerelease may only satisfy a range\n"
|
| 144 |
+
" when some clause names a prerelease on the same `major.minor.patch` triple, so\n"
|
| 145 |
+
" `>=1.0.0-alpha` does allow `1.0.0-beta` but not `1.0.1-beta`.\n\n"
|
| 146 |
+
"`semverlite/version.py` holds precedence, `semverlite/range.py` holds range\n"
|
| 147 |
+
"matching; `README.md` states both rules precisely. Build metadata must stay\n"
|
| 148 |
+
"ignored for precedence, and equality/hashing must stay consistent with the new\n"
|
| 149 |
+
"ordering."
|
| 150 |
+
),
|
| 151 |
+
},
|
| 152 |
+
]
|
| 153 |
+
|
| 154 |
+
|
| 155 |
+
def run(cmd, cwd, check=True):
|
| 156 |
+
env = os.environ.copy()
|
| 157 |
+
env.update(GIT_ENV)
|
| 158 |
+
proc = subprocess.run(
|
| 159 |
+
cmd, cwd=str(cwd), capture_output=True, text=True, encoding="utf-8", env=env
|
| 160 |
+
)
|
| 161 |
+
if check and proc.returncode != 0:
|
| 162 |
+
raise RuntimeError(
|
| 163 |
+
"command failed (%d): %s\n--- stdout ---\n%s\n--- stderr ---\n%s"
|
| 164 |
+
% (proc.returncode, " ".join(cmd), proc.stdout, proc.stderr)
|
| 165 |
+
)
|
| 166 |
+
return proc
|
| 167 |
+
|
| 168 |
+
|
| 169 |
+
def git(args, cwd, check=True):
|
| 170 |
+
return run(["git"] + GIT_FLAGS + args, cwd, check=check)
|
| 171 |
+
|
| 172 |
+
|
| 173 |
+
def overlay(src_root, dst_root):
|
| 174 |
+
"""Copy every file under src_root into dst_root, preserving relative paths."""
|
| 175 |
+
for path in sorted(src_root.rglob("*")):
|
| 176 |
+
if path.is_dir() or ".git" in path.parts:
|
| 177 |
+
continue
|
| 178 |
+
target = dst_root / path.relative_to(src_root)
|
| 179 |
+
target.parent.mkdir(parents=True, exist_ok=True)
|
| 180 |
+
shutil.copyfile(path, target)
|
| 181 |
+
|
| 182 |
+
|
| 183 |
+
def read_tree(root):
|
| 184 |
+
files = {}
|
| 185 |
+
for path in sorted(root.rglob("*")):
|
| 186 |
+
if path.is_dir() or ".git" in path.parts:
|
| 187 |
+
continue
|
| 188 |
+
rel = path.relative_to(root).as_posix()
|
| 189 |
+
files[rel] = path.read_text(encoding="utf-8")
|
| 190 |
+
return files
|
| 191 |
+
|
| 192 |
+
|
| 193 |
+
def write_tree(files, dest):
|
| 194 |
+
for rel, content in files.items():
|
| 195 |
+
path = dest / rel
|
| 196 |
+
path.parent.mkdir(parents=True, exist_ok=True)
|
| 197 |
+
with open(path, "w", encoding="utf-8", newline="\n") as handle:
|
| 198 |
+
handle.write(content)
|
| 199 |
+
|
| 200 |
+
|
| 201 |
+
def parse_outcomes(stdout):
|
| 202 |
+
outcomes = {}
|
| 203 |
+
for raw in stdout.splitlines():
|
| 204 |
+
match = RESULT_RE.match(raw.strip())
|
| 205 |
+
if match:
|
| 206 |
+
outcomes[match.group(1)] = match.group(2)
|
| 207 |
+
return outcomes
|
| 208 |
+
|
| 209 |
+
|
| 210 |
+
def run_tests(dest, test_files):
|
| 211 |
+
proc = run(
|
| 212 |
+
[sys.executable, "-m", "pytest", "-v", "--tb=no", "-p", "no:cacheprovider"]
|
| 213 |
+
+ list(test_files),
|
| 214 |
+
dest,
|
| 215 |
+
check=False,
|
| 216 |
+
)
|
| 217 |
+
return parse_outcomes(proc.stdout), proc
|
| 218 |
+
|
| 219 |
+
|
| 220 |
+
def build_instance(spec):
|
| 221 |
+
name = spec["name"]
|
| 222 |
+
repo_dir = REPOS / name
|
| 223 |
+
|
| 224 |
+
git_dir = repo_dir / ".git"
|
| 225 |
+
if git_dir.exists():
|
| 226 |
+
shutil.rmtree(git_dir)
|
| 227 |
+
git(["init", "-q", "-b", "main"], repo_dir)
|
| 228 |
+
git(["add", "-A"], repo_dir)
|
| 229 |
+
git(["commit", "-q", "-m", "initial import"], repo_dir)
|
| 230 |
+
base_commit = git(["rev-parse", "HEAD"], repo_dir).stdout.strip()
|
| 231 |
+
|
| 232 |
+
# --- gold patch -------------------------------------------------------
|
| 233 |
+
overlay(FIXED / name, repo_dir)
|
| 234 |
+
patch = git(["diff", "--no-color", "--no-ext-diff"], repo_dir).stdout
|
| 235 |
+
git(["checkout", "--", "."], repo_dir)
|
| 236 |
+
if not patch.strip():
|
| 237 |
+
raise RuntimeError("%s: gold patch is empty" % name)
|
| 238 |
+
|
| 239 |
+
# --- test patch -------------------------------------------------------
|
| 240 |
+
overlay(NEWTESTS / name, repo_dir)
|
| 241 |
+
git(["add", "-A"], repo_dir)
|
| 242 |
+
test_patch = git(["diff", "--cached", "--no-color", "--no-ext-diff"], repo_dir).stdout
|
| 243 |
+
git(["reset", "-q", "--hard", "HEAD"], repo_dir)
|
| 244 |
+
git(["clean", "-qfd"], repo_dir)
|
| 245 |
+
if not test_patch.strip():
|
| 246 |
+
raise RuntimeError("%s: test patch is empty" % name)
|
| 247 |
+
|
| 248 |
+
base_files = read_tree(repo_dir)
|
| 249 |
+
test_files = sorted(
|
| 250 |
+
rel for rel in base_files if rel.startswith("tests/") and rel.endswith(".py")
|
| 251 |
+
)
|
| 252 |
+
test_files += sorted(
|
| 253 |
+
rel
|
| 254 |
+
for rel in read_tree(NEWTESTS / name)
|
| 255 |
+
if rel.startswith("tests/") and rel.endswith(".py")
|
| 256 |
+
)
|
| 257 |
+
test_files = sorted(set(test_files))
|
| 258 |
+
|
| 259 |
+
# --- verify by really running pytest ----------------------------------
|
| 260 |
+
with tempfile.TemporaryDirectory(prefix="swedl-") as tmp:
|
| 261 |
+
work = Path(tmp) / "work"
|
| 262 |
+
work.mkdir()
|
| 263 |
+
write_tree(base_files, work)
|
| 264 |
+
git(["init", "-q", "-b", "main"], work)
|
| 265 |
+
git(["add", "-A"], work)
|
| 266 |
+
git(["commit", "-q", "-m", "base"], work)
|
| 267 |
+
|
| 268 |
+
test_patch_file = Path(tmp) / "test.patch"
|
| 269 |
+
test_patch_file.write_text(test_patch, encoding="utf-8", newline="\n")
|
| 270 |
+
git(["apply", "--check", str(test_patch_file)], work)
|
| 271 |
+
git(["apply", str(test_patch_file)], work)
|
| 272 |
+
|
| 273 |
+
before, before_proc = run_tests(work, test_files)
|
| 274 |
+
|
| 275 |
+
patch_file = Path(tmp) / "gold.patch"
|
| 276 |
+
patch_file.write_text(patch, encoding="utf-8", newline="\n")
|
| 277 |
+
git(["apply", str(patch_file)], work)
|
| 278 |
+
after, after_proc = run_tests(work, test_files)
|
| 279 |
+
|
| 280 |
+
if not before:
|
| 281 |
+
raise RuntimeError(
|
| 282 |
+
"%s: no test outcomes collected\n%s\n%s"
|
| 283 |
+
% (name, before_proc.stdout[-3000:], before_proc.stderr[-2000:])
|
| 284 |
+
)
|
| 285 |
+
missing = set(before) - set(after)
|
| 286 |
+
if missing:
|
| 287 |
+
raise RuntimeError("%s: tests disappeared after patch: %s" % (name, sorted(missing)))
|
| 288 |
+
|
| 289 |
+
f2p = sorted(
|
| 290 |
+
test
|
| 291 |
+
for test, outcome in before.items()
|
| 292 |
+
if outcome in ("FAILED", "ERROR") and after.get(test) == "PASSED"
|
| 293 |
+
)
|
| 294 |
+
p2p = sorted(
|
| 295 |
+
test
|
| 296 |
+
for test, outcome in before.items()
|
| 297 |
+
if outcome == "PASSED" and after.get(test) == "PASSED"
|
| 298 |
+
)
|
| 299 |
+
broken = sorted(test for test, outcome in after.items() if outcome != "PASSED")
|
| 300 |
+
if broken:
|
| 301 |
+
raise RuntimeError("%s: gold patch leaves failures: %s" % (name, broken))
|
| 302 |
+
if not f2p:
|
| 303 |
+
raise RuntimeError("%s: no FAIL_TO_PASS tests" % name)
|
| 304 |
+
if not p2p:
|
| 305 |
+
raise RuntimeError("%s: no PASS_TO_PASS tests" % name)
|
| 306 |
+
|
| 307 |
+
row = {
|
| 308 |
+
"instance_id": spec["instance_id"],
|
| 309 |
+
"repo": spec["repo"],
|
| 310 |
+
"base_commit": base_commit,
|
| 311 |
+
"patch": patch,
|
| 312 |
+
"test_patch": test_patch,
|
| 313 |
+
"problem_statement": spec["problem_statement"],
|
| 314 |
+
"hints_text": "",
|
| 315 |
+
"created_at": "2026-10-30",
|
| 316 |
+
"version": "1.0",
|
| 317 |
+
"environment_setup_commit": base_commit,
|
| 318 |
+
"FAIL_TO_PASS": json.dumps(f2p),
|
| 319 |
+
"PASS_TO_PASS": json.dumps(p2p),
|
| 320 |
+
"dockerless": {
|
| 321 |
+
"language": "python",
|
| 322 |
+
"python_requires": ">=3.9",
|
| 323 |
+
"install": [],
|
| 324 |
+
"run_from": ".",
|
| 325 |
+
"test_cmd": PYTEST_CMD + " {tests}",
|
| 326 |
+
"test_files": test_files,
|
| 327 |
+
"files": base_files,
|
| 328 |
+
},
|
| 329 |
+
}
|
| 330 |
+
return row, f2p, p2p, len(patch.splitlines())
|
| 331 |
+
|
| 332 |
+
|
| 333 |
+
def main():
|
| 334 |
+
rows = []
|
| 335 |
+
print("%-22s %5s %5s %7s %8s" % ("instance", "F2P", "P2P", "patch", "tests"))
|
| 336 |
+
for spec in INSTANCES:
|
| 337 |
+
row, f2p, p2p, patch_lines = build_instance(spec)
|
| 338 |
+
rows.append(row)
|
| 339 |
+
print(
|
| 340 |
+
"%-22s %5d %5d %7d %8d"
|
| 341 |
+
% (row["instance_id"], len(f2p), len(p2p), patch_lines, len(row["dockerless"]["test_files"]))
|
| 342 |
+
)
|
| 343 |
+
for test in f2p:
|
| 344 |
+
print(" F2P %s" % test)
|
| 345 |
+
|
| 346 |
+
with open(OUT, "w", encoding="utf-8", newline="\n") as handle:
|
| 347 |
+
for row in rows:
|
| 348 |
+
handle.write(json.dumps(row, ensure_ascii=False) + "\n")
|
| 349 |
+
|
| 350 |
+
size = OUT.stat().st_size
|
| 351 |
+
print("\nwrote %s (%d rows, %.1f KiB)" % (OUT.name, len(rows), size / 1024.0))
|
| 352 |
+
|
| 353 |
+
|
| 354 |
+
if __name__ == "__main__":
|
| 355 |
+
main()
|
fixed/cronlite/cronlite/fields.py
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Parsing of individual cron fields into sets of integers."""
|
| 2 |
+
|
| 3 |
+
|
| 4 |
+
def _sunday_alias(value):
|
| 5 |
+
"""Collapse the cron Sunday alias 7 onto 0."""
|
| 6 |
+
return 0 if value == 7 else value
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
class FieldSpec:
|
| 10 |
+
"""Describes one of the five cron fields."""
|
| 11 |
+
|
| 12 |
+
def __init__(self, name, minimum, maximum, normalize=None):
|
| 13 |
+
self.name = name
|
| 14 |
+
self.minimum = minimum
|
| 15 |
+
self.maximum = maximum
|
| 16 |
+
self.normalize = normalize
|
| 17 |
+
|
| 18 |
+
def _check(self, value):
|
| 19 |
+
if value < self.minimum or value > self.maximum:
|
| 20 |
+
raise ValueError("%s value out of range: %d" % (self.name, value))
|
| 21 |
+
return value
|
| 22 |
+
|
| 23 |
+
def parse(self, text):
|
| 24 |
+
values = set()
|
| 25 |
+
for part in text.split(","):
|
| 26 |
+
part = part.strip()
|
| 27 |
+
if not part:
|
| 28 |
+
raise ValueError("empty %s component" % self.name)
|
| 29 |
+
values |= self._parse_part(part)
|
| 30 |
+
if not values:
|
| 31 |
+
raise ValueError("empty %s field" % self.name)
|
| 32 |
+
if self.normalize is not None:
|
| 33 |
+
values = {self.normalize(value) for value in values}
|
| 34 |
+
return values
|
| 35 |
+
|
| 36 |
+
def _parse_part(self, part):
|
| 37 |
+
if part == "*":
|
| 38 |
+
return set(range(self.minimum, self.maximum + 1))
|
| 39 |
+
if "/" in part:
|
| 40 |
+
base, _, step_text = part.partition("/")
|
| 41 |
+
try:
|
| 42 |
+
step = int(step_text)
|
| 43 |
+
except ValueError:
|
| 44 |
+
raise ValueError("invalid step in %s field: %r" % (self.name, part))
|
| 45 |
+
if step <= 0:
|
| 46 |
+
raise ValueError("step must be positive in %s field" % self.name)
|
| 47 |
+
if base == "*":
|
| 48 |
+
start, stop = self.minimum, self.maximum
|
| 49 |
+
elif "-" in base:
|
| 50 |
+
start, stop = self._parse_bounds(base)
|
| 51 |
+
else:
|
| 52 |
+
start = self._check(int(base))
|
| 53 |
+
stop = self.maximum
|
| 54 |
+
return set(range(start, stop + 1, step))
|
| 55 |
+
if "-" in part:
|
| 56 |
+
start, stop = self._parse_bounds(part)
|
| 57 |
+
return set(range(start, stop + 1))
|
| 58 |
+
return {self._check(int(part))}
|
| 59 |
+
|
| 60 |
+
def _parse_bounds(self, text):
|
| 61 |
+
low, _, high = text.partition("-")
|
| 62 |
+
try:
|
| 63 |
+
start, stop = int(low), int(high)
|
| 64 |
+
except ValueError:
|
| 65 |
+
raise ValueError("invalid range in %s field: %r" % (self.name, text))
|
| 66 |
+
self._check(start)
|
| 67 |
+
self._check(stop)
|
| 68 |
+
if stop < start:
|
| 69 |
+
raise ValueError("descending range in %s field: %r" % (self.name, text))
|
| 70 |
+
return start, stop
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
MINUTE = FieldSpec("minute", 0, 59)
|
| 74 |
+
HOUR = FieldSpec("hour", 0, 23)
|
| 75 |
+
DAY_OF_MONTH = FieldSpec("day_of_month", 1, 31)
|
| 76 |
+
MONTH = FieldSpec("month", 1, 12)
|
| 77 |
+
DAY_OF_WEEK = FieldSpec("day_of_week", 0, 7, normalize=_sunday_alias)
|
| 78 |
+
|
| 79 |
+
|
| 80 |
+
def parse_field(spec, text):
|
| 81 |
+
"""Parse ``text`` against ``spec`` and return the matching values."""
|
| 82 |
+
return spec.parse(text)
|
fixed/miniconf/miniconf/config.py
ADDED
|
@@ -0,0 +1,92 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""The Config object: layered key/value storage with ${...} interpolation."""
|
| 2 |
+
|
| 3 |
+
from .errors import CircularReferenceError, MissingKeyError
|
| 4 |
+
from .loader import load_layers
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
class Config:
|
| 8 |
+
"""A small configuration mapping whose string values may reference keys.
|
| 9 |
+
|
| 10 |
+
See README.md for the interpolation contract.
|
| 11 |
+
"""
|
| 12 |
+
|
| 13 |
+
def __init__(self, data):
|
| 14 |
+
if not isinstance(data, dict):
|
| 15 |
+
raise TypeError("Config data must be a dict")
|
| 16 |
+
self._raw = dict(data)
|
| 17 |
+
|
| 18 |
+
@classmethod
|
| 19 |
+
def from_layers(cls, *layers):
|
| 20 |
+
"""Build a Config from one or more dict layers, later layers winning."""
|
| 21 |
+
return cls(load_layers(*layers))
|
| 22 |
+
|
| 23 |
+
def raw(self, key, default=None):
|
| 24 |
+
"""Return the unresolved value stored under ``key``."""
|
| 25 |
+
return self._raw.get(key, default)
|
| 26 |
+
|
| 27 |
+
def keys(self):
|
| 28 |
+
return list(self._raw)
|
| 29 |
+
|
| 30 |
+
def get(self, key, default=None):
|
| 31 |
+
if key not in self._raw:
|
| 32 |
+
return default
|
| 33 |
+
return self._resolve(key)
|
| 34 |
+
|
| 35 |
+
def __getitem__(self, key):
|
| 36 |
+
if key not in self._raw:
|
| 37 |
+
raise KeyError(key)
|
| 38 |
+
return self._resolve(key)
|
| 39 |
+
|
| 40 |
+
def __contains__(self, key):
|
| 41 |
+
return key in self._raw
|
| 42 |
+
|
| 43 |
+
def __len__(self):
|
| 44 |
+
return len(self._raw)
|
| 45 |
+
|
| 46 |
+
def as_dict(self):
|
| 47 |
+
"""Return every key resolved to a plain value."""
|
| 48 |
+
return {key: self._resolve(key) for key in self._raw}
|
| 49 |
+
|
| 50 |
+
def _resolve(self, key, stack=()):
|
| 51 |
+
if key in stack:
|
| 52 |
+
cycle = stack[stack.index(key):] + (key,)
|
| 53 |
+
raise CircularReferenceError(cycle)
|
| 54 |
+
if key not in self._raw:
|
| 55 |
+
raise MissingKeyError(key, stack[-1] if stack else None)
|
| 56 |
+
value = self._raw[key]
|
| 57 |
+
if not isinstance(value, str):
|
| 58 |
+
return value
|
| 59 |
+
return self._render(value, stack + (key,))
|
| 60 |
+
|
| 61 |
+
def _render(self, text, stack):
|
| 62 |
+
"""Resolve references in ``text`` in a single left-to-right pass."""
|
| 63 |
+
out = []
|
| 64 |
+
index = 0
|
| 65 |
+
length = len(text)
|
| 66 |
+
while index < length:
|
| 67 |
+
char = text[index]
|
| 68 |
+
if char != "$":
|
| 69 |
+
out.append(char)
|
| 70 |
+
index += 1
|
| 71 |
+
continue
|
| 72 |
+
if text.startswith("$$", index):
|
| 73 |
+
out.append("$")
|
| 74 |
+
index += 2
|
| 75 |
+
continue
|
| 76 |
+
if text.startswith("${", index):
|
| 77 |
+
end = text.find("}", index + 2)
|
| 78 |
+
if end == -1:
|
| 79 |
+
out.append(char)
|
| 80 |
+
index += 1
|
| 81 |
+
continue
|
| 82 |
+
name = text[index + 2:end]
|
| 83 |
+
if not name:
|
| 84 |
+
out.append(char)
|
| 85 |
+
index += 1
|
| 86 |
+
continue
|
| 87 |
+
out.append(str(self._resolve(name, stack)))
|
| 88 |
+
index = end + 1
|
| 89 |
+
continue
|
| 90 |
+
out.append(char)
|
| 91 |
+
index += 1
|
| 92 |
+
return "".join(out)
|
fixed/semverlite/semverlite/range.py
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Version range expressions such as ``">=1.2.3, <2.0.0"``."""
|
| 2 |
+
|
| 3 |
+
import re
|
| 4 |
+
|
| 5 |
+
from .version import Version
|
| 6 |
+
|
| 7 |
+
_CLAUSE_RE = re.compile(r"^(>=|<=|==|=|>|<|\^|~)?\s*(\S+)$")
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def _parse_clause(text):
|
| 11 |
+
match = _CLAUSE_RE.match(text)
|
| 12 |
+
if match is None:
|
| 13 |
+
raise ValueError("invalid range clause: %r" % (text,))
|
| 14 |
+
operator = match.group(1) or "=="
|
| 15 |
+
return operator, Version.parse(match.group(2))
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
def _caret_upper(target):
|
| 19 |
+
"""Upper bound of ``^target``: the next release that breaks the left-most
|
| 20 |
+
non-zero component."""
|
| 21 |
+
if target.major > 0:
|
| 22 |
+
return Version(target.major + 1, 0, 0)
|
| 23 |
+
if target.minor > 0:
|
| 24 |
+
return Version(0, target.minor + 1, 0)
|
| 25 |
+
return Version(0, 0, target.patch + 1)
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
def _tilde_upper(target):
|
| 29 |
+
"""Upper bound of ``~target``: the next minor release."""
|
| 30 |
+
return Version(target.major, target.minor + 1, 0)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
class VersionRange:
|
| 34 |
+
"""A comma-separated list of clauses that must all hold."""
|
| 35 |
+
|
| 36 |
+
def __init__(self, text):
|
| 37 |
+
self.text = text.strip()
|
| 38 |
+
parts = [part.strip() for part in self.text.split(",") if part.strip()]
|
| 39 |
+
if not parts:
|
| 40 |
+
raise ValueError("empty range")
|
| 41 |
+
self.clauses = [_parse_clause(part) for part in parts]
|
| 42 |
+
|
| 43 |
+
def _allows_prerelease(self, version):
|
| 44 |
+
"""A prerelease needs an explicit opt-in on the same release triple."""
|
| 45 |
+
for _, target in self.clauses:
|
| 46 |
+
if target.prerelease and target.core == version.core:
|
| 47 |
+
return True
|
| 48 |
+
return False
|
| 49 |
+
|
| 50 |
+
def allows(self, version):
|
| 51 |
+
if version.prerelease and not self._allows_prerelease(version):
|
| 52 |
+
return False
|
| 53 |
+
return all(self._allows_clause(op, target, version) for op, target in self.clauses)
|
| 54 |
+
|
| 55 |
+
@staticmethod
|
| 56 |
+
def _allows_clause(operator, target, version):
|
| 57 |
+
if operator in ("=", "=="):
|
| 58 |
+
return version == target
|
| 59 |
+
if operator == ">":
|
| 60 |
+
return version > target
|
| 61 |
+
if operator == ">=":
|
| 62 |
+
return version >= target
|
| 63 |
+
if operator == "<":
|
| 64 |
+
return version < target
|
| 65 |
+
if operator == "<=":
|
| 66 |
+
return version <= target
|
| 67 |
+
if operator == "^":
|
| 68 |
+
return target <= version < _caret_upper(target)
|
| 69 |
+
if operator == "~":
|
| 70 |
+
return target <= version < _tilde_upper(target)
|
| 71 |
+
raise ValueError("unknown operator: %r" % (operator,))
|
| 72 |
+
|
| 73 |
+
def __contains__(self, version):
|
| 74 |
+
return self.allows(version)
|
| 75 |
+
|
| 76 |
+
def __str__(self):
|
| 77 |
+
return self.text
|
| 78 |
+
|
| 79 |
+
def __repr__(self):
|
| 80 |
+
return "VersionRange(%r)" % (self.text,)
|
fixed/semverlite/semverlite/version.py
ADDED
|
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Semantic version parsing and precedence."""
|
| 2 |
+
|
| 3 |
+
import re
|
| 4 |
+
|
| 5 |
+
_VERSION_RE = re.compile(
|
| 6 |
+
r"^(?P<major>0|[1-9]\d*)\.(?P<minor>0|[1-9]\d*)\.(?P<patch>0|[1-9]\d*)"
|
| 7 |
+
r"(?:-(?P<pre>[0-9A-Za-z.-]+))?"
|
| 8 |
+
r"(?:\+(?P<build>[0-9A-Za-z.-]+))?$"
|
| 9 |
+
)
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
def _identifier_key(identifier):
|
| 13 |
+
"""Sort key for one prerelease identifier.
|
| 14 |
+
|
| 15 |
+
Numeric identifiers sort numerically and below alphanumeric ones; the third
|
| 16 |
+
element keeps keys comparable when the first two tie.
|
| 17 |
+
"""
|
| 18 |
+
text = str(identifier)
|
| 19 |
+
if text.isdigit():
|
| 20 |
+
return (0, int(text), "")
|
| 21 |
+
return (1, 0, text)
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
class Version:
|
| 25 |
+
"""A semantic version: major.minor.patch[-prerelease][+build]."""
|
| 26 |
+
|
| 27 |
+
def __init__(self, major, minor, patch, prerelease=(), build=()):
|
| 28 |
+
self.major = int(major)
|
| 29 |
+
self.minor = int(minor)
|
| 30 |
+
self.patch = int(patch)
|
| 31 |
+
self.prerelease = tuple(prerelease)
|
| 32 |
+
self.build = tuple(build)
|
| 33 |
+
|
| 34 |
+
@classmethod
|
| 35 |
+
def parse(cls, text):
|
| 36 |
+
match = _VERSION_RE.match(text.strip())
|
| 37 |
+
if match is None:
|
| 38 |
+
raise ValueError("invalid version: %r" % (text,))
|
| 39 |
+
prerelease = tuple(match.group("pre").split(".")) if match.group("pre") else ()
|
| 40 |
+
build = tuple(match.group("build").split(".")) if match.group("build") else ()
|
| 41 |
+
return cls(
|
| 42 |
+
match.group("major"),
|
| 43 |
+
match.group("minor"),
|
| 44 |
+
match.group("patch"),
|
| 45 |
+
prerelease,
|
| 46 |
+
build,
|
| 47 |
+
)
|
| 48 |
+
|
| 49 |
+
@property
|
| 50 |
+
def core(self):
|
| 51 |
+
"""The (major, minor, patch) triple."""
|
| 52 |
+
return (self.major, self.minor, self.patch)
|
| 53 |
+
|
| 54 |
+
def precedence_key(self):
|
| 55 |
+
"""Sort key implementing semver precedence (build metadata ignored)."""
|
| 56 |
+
core = self.core
|
| 57 |
+
if not self.prerelease:
|
| 58 |
+
return core + ((1,),)
|
| 59 |
+
return core + ((0, tuple(_identifier_key(part) for part in self.prerelease)),)
|
| 60 |
+
|
| 61 |
+
def __str__(self):
|
| 62 |
+
text = "%d.%d.%d" % self.core
|
| 63 |
+
if self.prerelease:
|
| 64 |
+
text += "-" + ".".join(str(part) for part in self.prerelease)
|
| 65 |
+
if self.build:
|
| 66 |
+
text += "+" + ".".join(str(part) for part in self.build)
|
| 67 |
+
return text
|
| 68 |
+
|
| 69 |
+
def __repr__(self):
|
| 70 |
+
return "Version(%r)" % (str(self),)
|
| 71 |
+
|
| 72 |
+
def __eq__(self, other):
|
| 73 |
+
if not isinstance(other, Version):
|
| 74 |
+
return NotImplemented
|
| 75 |
+
return self.precedence_key() == other.precedence_key()
|
| 76 |
+
|
| 77 |
+
def __ne__(self, other):
|
| 78 |
+
result = self.__eq__(other)
|
| 79 |
+
if result is NotImplemented:
|
| 80 |
+
return result
|
| 81 |
+
return not result
|
| 82 |
+
|
| 83 |
+
def __lt__(self, other):
|
| 84 |
+
return self.precedence_key() < other.precedence_key()
|
| 85 |
+
|
| 86 |
+
def __le__(self, other):
|
| 87 |
+
return self.precedence_key() <= other.precedence_key()
|
| 88 |
+
|
| 89 |
+
def __gt__(self, other):
|
| 90 |
+
return self.precedence_key() > other.precedence_key()
|
| 91 |
+
|
| 92 |
+
def __ge__(self, other):
|
| 93 |
+
return self.precedence_key() >= other.precedence_key()
|
| 94 |
+
|
| 95 |
+
def __hash__(self):
|
| 96 |
+
return hash(self.precedence_key())
|
fixed/ttlcache/ttlcache/cache.py
ADDED
|
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""A bounded cache with per-entry time-to-live and LRU eviction."""
|
| 2 |
+
|
| 3 |
+
from collections import OrderedDict
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
class TTLCache:
|
| 7 |
+
"""Fixed-capacity cache where every entry expires ``ttl`` seconds after write.
|
| 8 |
+
|
| 9 |
+
Entries are kept in least-recently-used order. When the cache is full the
|
| 10 |
+
least recently used entry is dropped first. See README.md for the full
|
| 11 |
+
contract.
|
| 12 |
+
"""
|
| 13 |
+
|
| 14 |
+
def __init__(self, maxsize=128, ttl=60.0, clock=None):
|
| 15 |
+
if maxsize <= 0:
|
| 16 |
+
raise ValueError("maxsize must be positive")
|
| 17 |
+
if ttl <= 0:
|
| 18 |
+
raise ValueError("ttl must be positive")
|
| 19 |
+
self.maxsize = maxsize
|
| 20 |
+
self.ttl = float(ttl)
|
| 21 |
+
self._clock = clock
|
| 22 |
+
self._data = OrderedDict()
|
| 23 |
+
|
| 24 |
+
def _now(self):
|
| 25 |
+
return self._clock.now() if self._clock is not None else 0.0
|
| 26 |
+
|
| 27 |
+
@staticmethod
|
| 28 |
+
def _expired(deadline, now):
|
| 29 |
+
return now >= deadline
|
| 30 |
+
|
| 31 |
+
def _purge(self, now):
|
| 32 |
+
"""Drop every expired entry so it can never be counted or evicted."""
|
| 33 |
+
stale = [key for key, (_, deadline) in self._data.items() if self._expired(deadline, now)]
|
| 34 |
+
for key in stale:
|
| 35 |
+
del self._data[key]
|
| 36 |
+
|
| 37 |
+
def __len__(self):
|
| 38 |
+
now = self._now()
|
| 39 |
+
return sum(1 for _, deadline in self._data.values() if not self._expired(deadline, now))
|
| 40 |
+
|
| 41 |
+
def __contains__(self, key):
|
| 42 |
+
if key not in self._data:
|
| 43 |
+
return False
|
| 44 |
+
_, deadline = self._data[key]
|
| 45 |
+
return not self._expired(deadline, self._now())
|
| 46 |
+
|
| 47 |
+
def __setitem__(self, key, value):
|
| 48 |
+
now = self._now()
|
| 49 |
+
self._purge(now)
|
| 50 |
+
if key in self._data:
|
| 51 |
+
del self._data[key]
|
| 52 |
+
self._data[key] = (value, now + self.ttl)
|
| 53 |
+
while len(self._data) > self.maxsize:
|
| 54 |
+
self._data.popitem(last=False)
|
| 55 |
+
|
| 56 |
+
def __getitem__(self, key):
|
| 57 |
+
value, deadline = self._data[key]
|
| 58 |
+
if self._expired(deadline, self._now()):
|
| 59 |
+
del self._data[key]
|
| 60 |
+
raise KeyError(key)
|
| 61 |
+
self._data.move_to_end(key)
|
| 62 |
+
return value
|
| 63 |
+
|
| 64 |
+
def get(self, key, default=None):
|
| 65 |
+
try:
|
| 66 |
+
return self[key]
|
| 67 |
+
except KeyError:
|
| 68 |
+
return default
|
| 69 |
+
|
| 70 |
+
def setdefault(self, key, factory):
|
| 71 |
+
"""Return the live value for ``key``, computing and storing it if absent."""
|
| 72 |
+
if key in self:
|
| 73 |
+
return self[key]
|
| 74 |
+
value = factory()
|
| 75 |
+
self[key] = value
|
| 76 |
+
return value
|
| 77 |
+
|
| 78 |
+
def pop(self, key, default=None):
|
| 79 |
+
if key not in self._data:
|
| 80 |
+
return default
|
| 81 |
+
value, _ = self._data.pop(key)
|
| 82 |
+
return value
|
| 83 |
+
|
| 84 |
+
def items(self):
|
| 85 |
+
"""Iterate over live ``(key, value)`` pairs in LRU order."""
|
| 86 |
+
now = self._now()
|
| 87 |
+
return [(k, v) for k, (v, d) in self._data.items() if not self._expired(d, now)]
|
| 88 |
+
|
| 89 |
+
def __repr__(self):
|
| 90 |
+
return "TTLCache(maxsize=%d, ttl=%r, live=%d)" % (
|
| 91 |
+
self.maxsize,
|
| 92 |
+
self.ttl,
|
| 93 |
+
len(self),
|
| 94 |
+
)
|
newtests/cronlite/tests/test_step_and_sunday.py
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from datetime import datetime
|
| 2 |
+
|
| 3 |
+
from cronlite import CronExpression, next_after
|
| 4 |
+
from cronlite.fields import DAY_OF_WEEK, MINUTE, parse_field
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
def test_single_value_with_step_runs_to_end_of_range():
|
| 8 |
+
assert parse_field(MINUTE, "5/15") == {5, 20, 35, 50}
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
def test_single_value_with_step_stops_at_field_maximum():
|
| 12 |
+
assert parse_field(MINUTE, "58/5") == {58}
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
def test_next_after_with_stepped_start():
|
| 16 |
+
expr = CronExpression("5/20 * * * *")
|
| 17 |
+
assert next_after(expr, datetime(2026, 3, 4, 0, 6)) == datetime(2026, 3, 4, 0, 25)
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
def test_sunday_may_be_written_as_seven():
|
| 21 |
+
assert parse_field(DAY_OF_WEEK, "7") == {0}
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
def test_zero_and_seven_are_equivalent():
|
| 25 |
+
assert parse_field(DAY_OF_WEEK, "0") == parse_field(DAY_OF_WEEK, "7")
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
def test_weekend_range_ending_in_seven():
|
| 29 |
+
assert parse_field(DAY_OF_WEEK, "6-7") == {6, 0}
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
def test_seven_inside_a_list():
|
| 33 |
+
assert parse_field(DAY_OF_WEEK, "1,7") == {1, 0}
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
def test_next_after_sunday_seven():
|
| 37 |
+
expr = CronExpression("0 12 * * 7")
|
| 38 |
+
assert next_after(expr, datetime(2026, 3, 4, 0, 0)) == datetime(2026, 3, 8, 12, 0)
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
def test_weekend_range_next_run():
|
| 42 |
+
expr = CronExpression("0 9 * * 6-7")
|
| 43 |
+
assert next_after(expr, datetime(2026, 3, 6, 0, 0)) == datetime(2026, 3, 7, 9, 0)
|
newtests/miniconf/tests/test_interpolation_edge_cases.py
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import pytest
|
| 2 |
+
|
| 3 |
+
from miniconf import CircularReferenceError, Config, MissingKeyError
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
def test_dollar_escape():
|
| 7 |
+
cfg = Config({"price": "$${amount}"})
|
| 8 |
+
assert cfg["price"] == "${amount}"
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
def test_escape_next_to_reference():
|
| 12 |
+
cfg = Config({"name": "wren", "msg": "$${name} is ${name}"})
|
| 13 |
+
assert cfg["msg"] == "${name} is wren"
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
def test_self_reference():
|
| 17 |
+
cfg = Config({"a": "x${a}"})
|
| 18 |
+
with pytest.raises(CircularReferenceError):
|
| 19 |
+
cfg["a"]
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
def test_as_dict_raises_on_cycle():
|
| 23 |
+
cfg = Config({"a": "${b}", "b": "${a}"})
|
| 24 |
+
with pytest.raises(CircularReferenceError):
|
| 25 |
+
cfg.as_dict()
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
def test_cycle_not_involving_requested_key():
|
| 29 |
+
cfg = Config({"a": "fine", "b": "${c}", "c": "${b}"})
|
| 30 |
+
assert cfg["a"] == "fine"
|
| 31 |
+
with pytest.raises(CircularReferenceError):
|
| 32 |
+
cfg["b"]
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
def test_two_cycle_chain_starts_at_requested_key():
|
| 36 |
+
cfg = Config({"a": "${b}", "b": "${a}"})
|
| 37 |
+
with pytest.raises(CircularReferenceError) as excinfo:
|
| 38 |
+
cfg["a"]
|
| 39 |
+
assert str(excinfo.value) == "circular reference: a -> b -> a"
|
| 40 |
+
|
| 41 |
+
|
| 42 |
+
def test_self_reference_message():
|
| 43 |
+
cfg = Config({"loop": "${loop}"})
|
| 44 |
+
with pytest.raises(CircularReferenceError) as excinfo:
|
| 45 |
+
cfg["loop"]
|
| 46 |
+
assert str(excinfo.value) == "circular reference: loop -> loop"
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
def test_escaped_text_is_not_reinterpolated_through_a_reference():
|
| 50 |
+
cfg = Config({"inner": "$${leak}", "outer": "${inner}"})
|
| 51 |
+
assert cfg["outer"] == "${leak}"
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
def test_missing_key_inside_reference_chain():
|
| 55 |
+
cfg = Config({"a": "${b}", "b": "${nope}"})
|
| 56 |
+
with pytest.raises(MissingKeyError) as excinfo:
|
| 57 |
+
cfg["a"]
|
| 58 |
+
assert str(excinfo.value) == "missing key 'nope' referenced from 'b'"
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
def test_escape_does_not_hide_a_real_cycle():
|
| 62 |
+
cfg = Config({"a": "$${b}", "b": "${a}"})
|
| 63 |
+
assert cfg["a"] == "${b}"
|
| 64 |
+
assert cfg["b"] == "${b}"
|
newtests/semverlite/tests/test_prerelease_precedence.py
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import pytest
|
| 2 |
+
|
| 3 |
+
from semverlite import Version, VersionRange
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
def v(text):
|
| 7 |
+
return Version.parse(text)
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def test_numeric_identifiers_compare_numerically():
|
| 11 |
+
assert v("1.0.0-alpha.2") < v("1.0.0-alpha.10")
|
| 12 |
+
|
| 13 |
+
|
| 14 |
+
def test_numeric_identifier_is_lower_than_alphanumeric():
|
| 15 |
+
assert v("1.0.0-1") < v("1.0.0-alpha")
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
def test_total_order_matches_semver_precedence():
|
| 19 |
+
chain = [
|
| 20 |
+
"1.0.0-alpha",
|
| 21 |
+
"1.0.0-alpha.1",
|
| 22 |
+
"1.0.0-alpha.beta",
|
| 23 |
+
"1.0.0-beta",
|
| 24 |
+
"1.0.0-beta.2",
|
| 25 |
+
"1.0.0-beta.11",
|
| 26 |
+
"1.0.0-rc.1",
|
| 27 |
+
"1.0.0",
|
| 28 |
+
]
|
| 29 |
+
versions = [v(text) for text in chain]
|
| 30 |
+
assert versions == sorted(versions)
|
| 31 |
+
assert len(set(versions)) == len(versions)
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
def test_prerelease_does_not_satisfy_release_lower_bound():
|
| 35 |
+
assert not VersionRange(">=1.0.0").allows(v("1.0.1-alpha"))
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
def test_prerelease_does_not_satisfy_caret_range():
|
| 39 |
+
assert not VersionRange("^1.2.3").allows(v("1.3.0-beta"))
|
| 40 |
+
|
| 41 |
+
|
| 42 |
+
def test_tilde_range_excludes_prerelease():
|
| 43 |
+
assert VersionRange("~1.2.3").allows(v("1.2.4"))
|
| 44 |
+
assert not VersionRange("~1.2.3").allows(v("1.2.4-rc.1"))
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
def test_range_allows_prerelease_when_same_triple_is_named():
|
| 48 |
+
assert VersionRange(">=1.0.0-alpha").allows(v("1.0.0-beta"))
|
| 49 |
+
assert not VersionRange(">=1.0.0-alpha").allows(v("1.0.1-beta"))
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
def test_prerelease_ordering_is_stable_across_comparisons():
|
| 53 |
+
assert v("2.0.0-beta.9") > v("2.0.0-beta.8")
|
| 54 |
+
assert v("2.0.0-beta.10") > v("2.0.0-beta.9")
|
| 55 |
+
assert not v("2.0.0-beta.10") < v("2.0.0-beta.9")
|
newtests/ttlcache/tests/test_lru_semantics.py
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from ttlcache import FakeClock, TTLCache
|
| 2 |
+
|
| 3 |
+
|
| 4 |
+
def test_get_refreshes_recency():
|
| 5 |
+
clock = FakeClock()
|
| 6 |
+
cache = TTLCache(maxsize=2, ttl=10.0, clock=clock)
|
| 7 |
+
cache["a"] = 1
|
| 8 |
+
cache["b"] = 2
|
| 9 |
+
assert cache["a"] == 1
|
| 10 |
+
cache["c"] = 3
|
| 11 |
+
assert "a" in cache
|
| 12 |
+
assert "b" not in cache
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
def test_overwrite_refreshes_recency():
|
| 16 |
+
clock = FakeClock()
|
| 17 |
+
cache = TTLCache(maxsize=2, ttl=10.0, clock=clock)
|
| 18 |
+
cache["a"] = 1
|
| 19 |
+
cache["b"] = 2
|
| 20 |
+
cache["a"] = 9
|
| 21 |
+
cache["c"] = 3
|
| 22 |
+
assert cache["a"] == 9
|
| 23 |
+
assert "b" not in cache
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
def test_setdefault_refreshes_recency():
|
| 27 |
+
clock = FakeClock()
|
| 28 |
+
cache = TTLCache(maxsize=2, ttl=10.0, clock=clock)
|
| 29 |
+
cache["a"] = 1
|
| 30 |
+
cache["b"] = 2
|
| 31 |
+
assert cache.setdefault("a", lambda: 99) == 1
|
| 32 |
+
cache["c"] = 3
|
| 33 |
+
assert "a" in cache
|
| 34 |
+
assert "b" not in cache
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
def test_len_ignores_expired_entries():
|
| 38 |
+
clock = FakeClock()
|
| 39 |
+
cache = TTLCache(maxsize=4, ttl=10.0, clock=clock)
|
| 40 |
+
cache["a"] = 1
|
| 41 |
+
clock.advance(5)
|
| 42 |
+
cache["b"] = 2
|
| 43 |
+
clock.advance(6)
|
| 44 |
+
assert len(cache) == 1
|
| 45 |
+
assert "a" not in cache
|
| 46 |
+
assert "b" in cache
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
def test_items_reports_least_recently_used_first():
|
| 50 |
+
clock = FakeClock()
|
| 51 |
+
cache = TTLCache(maxsize=3, ttl=10.0, clock=clock)
|
| 52 |
+
cache["a"] = 1
|
| 53 |
+
cache["b"] = 2
|
| 54 |
+
assert cache["a"] == 1
|
| 55 |
+
assert list(cache.items()) == [("b", 2), ("a", 1)]
|
predictions/cross.jsonl
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"instance_id": "swedl__miniconf-1", "model_name_or_path": "control-cross", "model_patch": "diff --git a/ttlcache/cache.py b/ttlcache/cache.py\nindex fda1724..4c36ad4 100644\n--- a/ttlcache/cache.py\n+++ b/ttlcache/cache.py\n@@ -28,8 +28,15 @@ class TTLCache:\n def _expired(deadline, now):\n return now >= deadline\n \n+ def _purge(self, now):\n+ \"\"\"Drop every expired entry so it can never be counted or evicted.\"\"\"\n+ stale = [key for key, (_, deadline) in self._data.items() if self._expired(deadline, now)]\n+ for key in stale:\n+ del self._data[key]\n+\n def __len__(self):\n- return len(self._data)\n+ now = self._now()\n+ return sum(1 for _, deadline in self._data.values() if not self._expired(deadline, now))\n \n def __contains__(self, key):\n if key not in self._data:\n@@ -39,6 +46,9 @@ class TTLCache:\n \n def __setitem__(self, key, value):\n now = self._now()\n+ self._purge(now)\n+ if key in self._data:\n+ del self._data[key]\n self._data[key] = (value, now + self.ttl)\n while len(self._data) > self.maxsize:\n self._data.popitem(last=False)\n@@ -48,6 +58,7 @@ class TTLCache:\n if self._expired(deadline, self._now()):\n del self._data[key]\n raise KeyError(key)\n+ self._data.move_to_end(key)\n return value\n \n def get(self, key, default=None):\n"}
|
| 2 |
+
{"instance_id": "swedl__ttlcache-1", "model_name_or_path": "control-cross", "model_patch": "diff --git a/cronlite/fields.py b/cronlite/fields.py\nindex 4945d5e..b3c0eb6 100644\n--- a/cronlite/fields.py\n+++ b/cronlite/fields.py\n@@ -1,13 +1,19 @@\n \"\"\"Parsing of individual cron fields into sets of integers.\"\"\"\n \n \n+def _sunday_alias(value):\n+ \"\"\"Collapse the cron Sunday alias 7 onto 0.\"\"\"\n+ return 0 if value == 7 else value\n+\n+\n class FieldSpec:\n \"\"\"Describes one of the five cron fields.\"\"\"\n \n- def __init__(self, name, minimum, maximum):\n+ def __init__(self, name, minimum, maximum, normalize=None):\n self.name = name\n self.minimum = minimum\n self.maximum = maximum\n+ self.normalize = normalize\n \n def _check(self, value):\n if value < self.minimum or value > self.maximum:\n@@ -23,6 +29,8 @@ class FieldSpec:\n values |= self._parse_part(part)\n if not values:\n raise ValueError(\"empty %s field\" % self.name)\n+ if self.normalize is not None:\n+ values = {self.normalize(value) for value in values}\n return values\n \n def _parse_part(self, part):\n@@ -42,7 +50,7 @@ class FieldSpec:\n start, stop = self._parse_bounds(base)\n else:\n start = self._check(int(base))\n- stop = start\n+ stop = self.maximum\n return set(range(start, stop + 1, step))\n if \"-\" in part:\n start, stop = self._parse_bounds(part)\n@@ -66,7 +74,7 @@ MINUTE = FieldSpec(\"minute\", 0, 59)\n HOUR = FieldSpec(\"hour\", 0, 23)\n DAY_OF_MONTH = FieldSpec(\"day_of_month\", 1, 31)\n MONTH = FieldSpec(\"month\", 1, 12)\n-DAY_OF_WEEK = FieldSpec(\"day_of_week\", 0, 6)\n+DAY_OF_WEEK = FieldSpec(\"day_of_week\", 0, 7, normalize=_sunday_alias)\n \n \n def parse_field(spec, text):\n"}
|
| 3 |
+
{"instance_id": "swedl__cronlite-1", "model_name_or_path": "control-cross", "model_patch": "diff --git a/semverlite/range.py b/semverlite/range.py\nindex abe2ece..ad1fc9e 100644\n--- a/semverlite/range.py\n+++ b/semverlite/range.py\n@@ -40,7 +40,16 @@ class VersionRange:\n raise ValueError(\"empty range\")\n self.clauses = [_parse_clause(part) for part in parts]\n \n+ def _allows_prerelease(self, version):\n+ \"\"\"A prerelease needs an explicit opt-in on the same release triple.\"\"\"\n+ for _, target in self.clauses:\n+ if target.prerelease and target.core == version.core:\n+ return True\n+ return False\n+\n def allows(self, version):\n+ if version.prerelease and not self._allows_prerelease(version):\n+ return False\n return all(self._allows_clause(op, target, version) for op, target in self.clauses)\n \n @staticmethod\ndiff --git a/semverlite/version.py b/semverlite/version.py\nindex bbb779c..4d498ed 100644\n--- a/semverlite/version.py\n+++ b/semverlite/version.py\n@@ -9,6 +9,18 @@ _VERSION_RE = re.compile(\n )\n \n \n+def _identifier_key(identifier):\n+ \"\"\"Sort key for one prerelease identifier.\n+\n+ Numeric identifiers sort numerically and below alphanumeric ones; the third\n+ element keeps keys comparable when the first two tie.\n+ \"\"\"\n+ text = str(identifier)\n+ if text.isdigit():\n+ return (0, int(text), \"\")\n+ return (1, 0, text)\n+\n+\n class Version:\n \"\"\"A semantic version: major.minor.patch[-prerelease][+build].\"\"\"\n \n@@ -43,8 +55,8 @@ class Version:\n \"\"\"Sort key implementing semver precedence (build metadata ignored).\"\"\"\n core = self.core\n if not self.prerelease:\n- return core + (1, ())\n- return core + (0, tuple(str(part) for part in self.prerelease))\n+ return core + ((1,),)\n+ return core + ((0, tuple(_identifier_key(part) for part in self.prerelease)),)\n \n def __str__(self):\n text = \"%d.%d.%d\" % self.core\n"}
|
| 4 |
+
{"instance_id": "swedl__semverlite-1", "model_name_or_path": "control-cross", "model_patch": "diff --git a/miniconf/config.py b/miniconf/config.py\nindex 8751912..6bdf695 100644\n--- a/miniconf/config.py\n+++ b/miniconf/config.py\n@@ -1,12 +1,8 @@\n \"\"\"The Config object: layered key/value storage with ${...} interpolation.\"\"\"\n \n-import re\n-\n-from .errors import MissingKeyError\n+from .errors import CircularReferenceError, MissingKeyError\n from .loader import load_layers\n \n-REFERENCE = re.compile(r\"\\$\\{([A-Za-z_][A-Za-z0-9_.-]*)\\}\")\n-\n \n class Config:\n \"\"\"A small configuration mapping whose string values may reference keys.\n@@ -51,17 +47,46 @@ class Config:\n \"\"\"Return every key resolved to a plain value.\"\"\"\n return {key: self._resolve(key) for key in self._raw}\n \n- def _resolve(self, key):\n+ def _resolve(self, key, stack=()):\n+ if key in stack:\n+ cycle = stack[stack.index(key):] + (key,)\n+ raise CircularReferenceError(cycle)\n+ if key not in self._raw:\n+ raise MissingKeyError(key, stack[-1] if stack else None)\n value = self._raw[key]\n if not isinstance(value, str):\n return value\n- return self._interpolate(value, key)\n-\n- def _interpolate(self, text, origin):\n- def replace(match):\n- name = match.group(1)\n- if name not in self._raw:\n- raise MissingKeyError(name, origin)\n- return str(self._resolve(name))\n-\n- return REFERENCE.sub(replace, text)\n+ return self._render(value, stack + (key,))\n+\n+ def _render(self, text, stack):\n+ \"\"\"Resolve references in ``text`` in a single left-to-right pass.\"\"\"\n+ out = []\n+ index = 0\n+ length = len(text)\n+ while index < length:\n+ char = text[index]\n+ if char != \"$\":\n+ out.append(char)\n+ index += 1\n+ continue\n+ if text.startswith(\"$$\", index):\n+ out.append(\"$\")\n+ index += 2\n+ continue\n+ if text.startswith(\"${\", index):\n+ end = text.find(\"}\", index + 2)\n+ if end == -1:\n+ out.append(char)\n+ index += 1\n+ continue\n+ name = text[index + 2:end]\n+ if not name:\n+ out.append(char)\n+ index += 1\n+ continue\n+ out.append(str(self._resolve(name, stack)))\n+ index = end + 1\n+ continue\n+ out.append(char)\n+ index += 1\n+ return \"\".join(out)\n"}
|
predictions/empty.jsonl
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"instance_id": "swedl__miniconf-1", "model_name_or_path": "control-empty", "model_patch": ""}
|
| 2 |
+
{"instance_id": "swedl__ttlcache-1", "model_name_or_path": "control-empty", "model_patch": ""}
|
| 3 |
+
{"instance_id": "swedl__cronlite-1", "model_name_or_path": "control-empty", "model_patch": ""}
|
| 4 |
+
{"instance_id": "swedl__semverlite-1", "model_name_or_path": "control-empty", "model_patch": ""}
|
predictions/partial.jsonl
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{"instance_id": "swedl__miniconf-1", "model_name_or_path": "control-partial", "model_patch": "diff --git a/miniconf/config.py b/miniconf/config.py\nindex 8751912..952e616 100644\n--- a/miniconf/config.py\n+++ b/miniconf/config.py\n@@ -1,12 +1,8 @@\n \"\"\"The Config object: layered key/value storage with ${...} interpolation.\"\"\"\n \n-import re\n-\n-from .errors import MissingKeyError\n+from .errors import CircularReferenceError, MissingKeyError\n from .loader import load_layers\n \n-REFERENCE = re.compile(r\"\\$\\{([A-Za-z_][A-Za-z0-9_.-]*)\\}\")\n-\n \n class Config:\n \"\"\"A small configuration mapping whose string values may reference keys.\n@@ -51,17 +47,42 @@ class Config:\n \"\"\"Return every key resolved to a plain value.\"\"\"\n return {key: self._resolve(key) for key in self._raw}\n \n- def _resolve(self, key):\n+ def _resolve(self, key, stack=()):\n+ if key in stack:\n+ cycle = stack[stack.index(key):] + (key,)\n+ raise CircularReferenceError(cycle)\n+ if key not in self._raw:\n+ raise MissingKeyError(key, stack[-1] if stack else None)\n value = self._raw[key]\n if not isinstance(value, str):\n return value\n- return self._interpolate(value, key)\n-\n- def _interpolate(self, text, origin):\n- def replace(match):\n- name = match.group(1)\n- if name not in self._raw:\n- raise MissingKeyError(name, origin)\n- return str(self._resolve(name))\n-\n- return REFERENCE.sub(replace, text)\n+ return self._render(value, stack + (key,))\n+\n+ def _render(self, text, stack):\n+ \"\"\"Resolve references in ``text`` in a single left-to-right pass.\"\"\"\n+ out = []\n+ index = 0\n+ length = len(text)\n+ while index < length:\n+ char = text[index]\n+ if char != \"$\":\n+ out.append(char)\n+ index += 1\n+ continue\n+ if text.startswith(\"${\", index):\n+ end = text.find(\"}\", index + 2)\n+ if end == -1:\n+ out.append(char)\n+ index += 1\n+ continue\n+ name = text[index + 2:end]\n+ if not name:\n+ out.append(char)\n+ index += 1\n+ continue\n+ out.append(str(self._resolve(name, stack)))\n+ index = end + 1\n+ continue\n+ out.append(char)\n+ index += 1\n+ return \"\".join(out)\n"}
|
| 2 |
+
{"instance_id": "swedl__ttlcache-1", "model_name_or_path": "control-partial", "model_patch": "diff --git a/ttlcache/cache.py b/ttlcache/cache.py\nindex fda1724..4088655 100644\n--- a/ttlcache/cache.py\n+++ b/ttlcache/cache.py\n@@ -28,8 +28,15 @@ class TTLCache:\n def _expired(deadline, now):\n return now >= deadline\n \n+ def _purge(self, now):\n+ \"\"\"Drop every expired entry so it can never be counted or evicted.\"\"\"\n+ stale = [key for key, (_, deadline) in self._data.items() if self._expired(deadline, now)]\n+ for key in stale:\n+ del self._data[key]\n+\n def __len__(self):\n- return len(self._data)\n+ now = self._now()\n+ return sum(1 for _, deadline in self._data.values() if not self._expired(deadline, now))\n \n def __contains__(self, key):\n if key not in self._data:\n@@ -39,6 +46,9 @@ class TTLCache:\n \n def __setitem__(self, key, value):\n now = self._now()\n+ self._purge(now)\n+ if key in self._data:\n+ del self._data[key]\n self._data[key] = (value, now + self.ttl)\n while len(self._data) > self.maxsize:\n self._data.popitem(last=False)\n"}
|
| 3 |
+
{"instance_id": "swedl__cronlite-1", "model_name_or_path": "control-partial", "model_patch": "diff --git a/cronlite/fields.py b/cronlite/fields.py\nindex 4945d5e..6af09a9 100644\n--- a/cronlite/fields.py\n+++ b/cronlite/fields.py\n@@ -1,13 +1,19 @@\n \"\"\"Parsing of individual cron fields into sets of integers.\"\"\"\n \n \n+def _sunday_alias(value):\n+ \"\"\"Collapse the cron Sunday alias 7 onto 0.\"\"\"\n+ return 0 if value == 7 else value\n+\n+\n class FieldSpec:\n \"\"\"Describes one of the five cron fields.\"\"\"\n \n- def __init__(self, name, minimum, maximum):\n+ def __init__(self, name, minimum, maximum, normalize=None):\n self.name = name\n self.minimum = minimum\n self.maximum = maximum\n+ self.normalize = normalize\n \n def _check(self, value):\n if value < self.minimum or value > self.maximum:\n@@ -23,6 +29,8 @@ class FieldSpec:\n values |= self._parse_part(part)\n if not values:\n raise ValueError(\"empty %s field\" % self.name)\n+ if self.normalize is not None:\n+ values = {self.normalize(value) for value in values}\n return values\n \n def _parse_part(self, part):\n@@ -66,7 +74,7 @@ MINUTE = FieldSpec(\"minute\", 0, 59)\n HOUR = FieldSpec(\"hour\", 0, 23)\n DAY_OF_MONTH = FieldSpec(\"day_of_month\", 1, 31)\n MONTH = FieldSpec(\"month\", 1, 12)\n-DAY_OF_WEEK = FieldSpec(\"day_of_week\", 0, 6)\n+DAY_OF_WEEK = FieldSpec(\"day_of_week\", 0, 7, normalize=_sunday_alias)\n \n \n def parse_field(spec, text):\n"}
|
| 4 |
+
{"instance_id": "swedl__semverlite-1", "model_name_or_path": "control-partial", "model_patch": "diff --git a/semverlite/version.py b/semverlite/version.py\nindex bbb779c..4d498ed 100644\n--- a/semverlite/version.py\n+++ b/semverlite/version.py\n@@ -9,6 +9,18 @@ _VERSION_RE = re.compile(\n )\n \n \n+def _identifier_key(identifier):\n+ \"\"\"Sort key for one prerelease identifier.\n+\n+ Numeric identifiers sort numerically and below alphanumeric ones; the third\n+ element keeps keys comparable when the first two tie.\n+ \"\"\"\n+ text = str(identifier)\n+ if text.isdigit():\n+ return (0, int(text), \"\")\n+ return (1, 0, text)\n+\n+\n class Version:\n \"\"\"A semantic version: major.minor.patch[-prerelease][+build].\"\"\"\n \n@@ -43,8 +55,8 @@ class Version:\n \"\"\"Sort key implementing semver precedence (build metadata ignored).\"\"\"\n core = self.core\n if not self.prerelease:\n- return core + (1, ())\n- return core + (0, tuple(str(part) for part in self.prerelease))\n+ return core + ((1,),)\n+ return core + ((0, tuple(_identifier_key(part) for part in self.prerelease)),)\n \n def __str__(self):\n text = \"%d.%d.%d\" % self.core\n"}
|
reports/cross.json
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"dataset": "SWE-DockerLess.jsonl",
|
| 3 |
+
"instances": 4,
|
| 4 |
+
"resolved": 0,
|
| 5 |
+
"resolve_rate": 0.0,
|
| 6 |
+
"seconds": 2.58,
|
| 7 |
+
"results": [
|
| 8 |
+
{
|
| 9 |
+
"instance_id": "swedl__miniconf-1",
|
| 10 |
+
"resolved": false,
|
| 11 |
+
"reason": "model patch failed to apply: ['error: ttlcache/cache.py: No such file or directory']",
|
| 12 |
+
"f2p_passed": 0,
|
| 13 |
+
"f2p_total": 9,
|
| 14 |
+
"p2p_passed": 0,
|
| 15 |
+
"p2p_total": 15
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"instance_id": "swedl__ttlcache-1",
|
| 19 |
+
"resolved": false,
|
| 20 |
+
"reason": "model patch failed to apply: ['error: cronlite/fields.py: No such file or directory']",
|
| 21 |
+
"f2p_passed": 0,
|
| 22 |
+
"f2p_total": 5,
|
| 23 |
+
"p2p_passed": 0,
|
| 24 |
+
"p2p_total": 12
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"instance_id": "swedl__cronlite-1",
|
| 28 |
+
"resolved": false,
|
| 29 |
+
"reason": "model patch failed to apply: ['error: semverlite/range.py: No such file or directory', 'error: semverlite/version.py: No such file or directory']",
|
| 30 |
+
"f2p_passed": 0,
|
| 31 |
+
"f2p_total": 8,
|
| 32 |
+
"p2p_passed": 0,
|
| 33 |
+
"p2p_total": 21
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"instance_id": "swedl__semverlite-1",
|
| 37 |
+
"resolved": false,
|
| 38 |
+
"reason": "model patch failed to apply: ['error: miniconf/config.py: No such file or directory']",
|
| 39 |
+
"f2p_passed": 0,
|
| 40 |
+
"f2p_total": 7,
|
| 41 |
+
"p2p_passed": 0,
|
| 42 |
+
"p2p_total": 19
|
| 43 |
+
}
|
| 44 |
+
]
|
| 45 |
+
}
|
reports/empty.json
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"dataset": "SWE-DockerLess.jsonl",
|
| 3 |
+
"instances": 4,
|
| 4 |
+
"resolved": 0,
|
| 5 |
+
"resolve_rate": 0.0,
|
| 6 |
+
"seconds": 2.49,
|
| 7 |
+
"results": [
|
| 8 |
+
{
|
| 9 |
+
"instance_id": "swedl__miniconf-1",
|
| 10 |
+
"resolved": false,
|
| 11 |
+
"reason": "model patch failed to apply: empty patch",
|
| 12 |
+
"f2p_passed": 0,
|
| 13 |
+
"f2p_total": 9,
|
| 14 |
+
"p2p_passed": 0,
|
| 15 |
+
"p2p_total": 15
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"instance_id": "swedl__ttlcache-1",
|
| 19 |
+
"resolved": false,
|
| 20 |
+
"reason": "model patch failed to apply: empty patch",
|
| 21 |
+
"f2p_passed": 0,
|
| 22 |
+
"f2p_total": 5,
|
| 23 |
+
"p2p_passed": 0,
|
| 24 |
+
"p2p_total": 12
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"instance_id": "swedl__cronlite-1",
|
| 28 |
+
"resolved": false,
|
| 29 |
+
"reason": "model patch failed to apply: empty patch",
|
| 30 |
+
"f2p_passed": 0,
|
| 31 |
+
"f2p_total": 8,
|
| 32 |
+
"p2p_passed": 0,
|
| 33 |
+
"p2p_total": 21
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"instance_id": "swedl__semverlite-1",
|
| 37 |
+
"resolved": false,
|
| 38 |
+
"reason": "model patch failed to apply: empty patch",
|
| 39 |
+
"f2p_passed": 0,
|
| 40 |
+
"f2p_total": 7,
|
| 41 |
+
"p2p_passed": 0,
|
| 42 |
+
"p2p_total": 19
|
| 43 |
+
}
|
| 44 |
+
]
|
| 45 |
+
}
|
reports/gold.json
ADDED
|
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"dataset": "SWE-DockerLess.jsonl",
|
| 3 |
+
"instances": 4,
|
| 4 |
+
"resolved": 4,
|
| 5 |
+
"resolve_rate": 1.0,
|
| 6 |
+
"seconds": 6.71,
|
| 7 |
+
"results": [
|
| 8 |
+
{
|
| 9 |
+
"instance_id": "swedl__miniconf-1",
|
| 10 |
+
"resolved": true,
|
| 11 |
+
"reason": "ok",
|
| 12 |
+
"f2p_passed": 9,
|
| 13 |
+
"f2p_total": 9,
|
| 14 |
+
"p2p_passed": 15,
|
| 15 |
+
"p2p_total": 15,
|
| 16 |
+
"failing": []
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"instance_id": "swedl__ttlcache-1",
|
| 20 |
+
"resolved": true,
|
| 21 |
+
"reason": "ok",
|
| 22 |
+
"f2p_passed": 5,
|
| 23 |
+
"f2p_total": 5,
|
| 24 |
+
"p2p_passed": 12,
|
| 25 |
+
"p2p_total": 12,
|
| 26 |
+
"failing": []
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"instance_id": "swedl__cronlite-1",
|
| 30 |
+
"resolved": true,
|
| 31 |
+
"reason": "ok",
|
| 32 |
+
"f2p_passed": 8,
|
| 33 |
+
"f2p_total": 8,
|
| 34 |
+
"p2p_passed": 21,
|
| 35 |
+
"p2p_total": 21,
|
| 36 |
+
"failing": []
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"instance_id": "swedl__semverlite-1",
|
| 40 |
+
"resolved": true,
|
| 41 |
+
"reason": "ok",
|
| 42 |
+
"f2p_passed": 7,
|
| 43 |
+
"f2p_total": 7,
|
| 44 |
+
"p2p_passed": 19,
|
| 45 |
+
"p2p_total": 19,
|
| 46 |
+
"failing": []
|
| 47 |
+
}
|
| 48 |
+
]
|
| 49 |
+
}
|
reports/partial.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"dataset": "SWE-DockerLess.jsonl",
|
| 3 |
+
"instances": 4,
|
| 4 |
+
"resolved": 0,
|
| 5 |
+
"resolve_rate": 0.0,
|
| 6 |
+
"seconds": 11.23,
|
| 7 |
+
"results": [
|
| 8 |
+
{
|
| 9 |
+
"instance_id": "swedl__miniconf-1",
|
| 10 |
+
"resolved": false,
|
| 11 |
+
"reason": "unmet tests: 4",
|
| 12 |
+
"f2p_passed": 5,
|
| 13 |
+
"f2p_total": 9,
|
| 14 |
+
"p2p_passed": 15,
|
| 15 |
+
"p2p_total": 15,
|
| 16 |
+
"failing": [
|
| 17 |
+
"tests/test_interpolation_edge_cases.py::test_dollar_escape",
|
| 18 |
+
"tests/test_interpolation_edge_cases.py::test_escape_does_not_hide_a_real_cycle",
|
| 19 |
+
"tests/test_interpolation_edge_cases.py::test_escape_next_to_reference",
|
| 20 |
+
"tests/test_interpolation_edge_cases.py::test_escaped_text_is_not_reinterpolated_through_a_reference"
|
| 21 |
+
]
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"instance_id": "swedl__ttlcache-1",
|
| 25 |
+
"resolved": false,
|
| 26 |
+
"reason": "unmet tests: 3",
|
| 27 |
+
"f2p_passed": 2,
|
| 28 |
+
"f2p_total": 5,
|
| 29 |
+
"p2p_passed": 12,
|
| 30 |
+
"p2p_total": 12,
|
| 31 |
+
"failing": [
|
| 32 |
+
"tests/test_lru_semantics.py::test_get_refreshes_recency",
|
| 33 |
+
"tests/test_lru_semantics.py::test_items_reports_least_recently_used_first",
|
| 34 |
+
"tests/test_lru_semantics.py::test_setdefault_refreshes_recency"
|
| 35 |
+
]
|
| 36 |
+
},
|
| 37 |
+
{
|
| 38 |
+
"instance_id": "swedl__cronlite-1",
|
| 39 |
+
"resolved": false,
|
| 40 |
+
"reason": "unmet tests: 2",
|
| 41 |
+
"f2p_passed": 6,
|
| 42 |
+
"f2p_total": 8,
|
| 43 |
+
"p2p_passed": 21,
|
| 44 |
+
"p2p_total": 21,
|
| 45 |
+
"failing": [
|
| 46 |
+
"tests/test_step_and_sunday.py::test_next_after_with_stepped_start",
|
| 47 |
+
"tests/test_step_and_sunday.py::test_single_value_with_step_runs_to_end_of_range"
|
| 48 |
+
]
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"instance_id": "swedl__semverlite-1",
|
| 52 |
+
"resolved": false,
|
| 53 |
+
"reason": "unmet tests: 4",
|
| 54 |
+
"f2p_passed": 3,
|
| 55 |
+
"f2p_total": 7,
|
| 56 |
+
"p2p_passed": 19,
|
| 57 |
+
"p2p_total": 19,
|
| 58 |
+
"failing": [
|
| 59 |
+
"tests/test_prerelease_precedence.py::test_prerelease_does_not_satisfy_caret_range",
|
| 60 |
+
"tests/test_prerelease_precedence.py::test_prerelease_does_not_satisfy_release_lower_bound",
|
| 61 |
+
"tests/test_prerelease_precedence.py::test_range_allows_prerelease_when_same_triple_is_named",
|
| 62 |
+
"tests/test_prerelease_precedence.py::test_tilde_range_excludes_prerelease"
|
| 63 |
+
]
|
| 64 |
+
}
|
| 65 |
+
]
|
| 66 |
+
}
|
repos/cronlite/README.md
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# cronlite
|
| 2 |
+
|
| 3 |
+
A dependency-free evaluator for standard five-field cron expressions:
|
| 4 |
+
|
| 5 |
+
```
|
| 6 |
+
minute hour day-of-month month day-of-week
|
| 7 |
+
```
|
| 8 |
+
|
| 9 |
+
## Field syntax
|
| 10 |
+
|
| 11 |
+
| Form | Meaning |
|
| 12 |
+
| --- | --- |
|
| 13 |
+
| `*` | every value in the field range |
|
| 14 |
+
| `7` | a single value |
|
| 15 |
+
| `1-5` | an inclusive range |
|
| 16 |
+
| `*/15` | every 15th value across the whole range |
|
| 17 |
+
| `1-10/3` | every 3rd value inside the range (`1, 4, 7, 10`) |
|
| 18 |
+
| `5/15` | every 15th value starting at 5 and **running to the end of the range** (`5, 20, 35, 50`) |
|
| 19 |
+
| `0,15,30` | a list; any of the above forms may be combined in a list |
|
| 20 |
+
|
| 21 |
+
Field ranges: minute `0-59`, hour `0-23`, day-of-month `1-31`, month `1-12`.
|
| 22 |
+
|
| 23 |
+
## Day of week
|
| 24 |
+
|
| 25 |
+
The day-of-week field is `0-7` where **both `0` and `7` mean Sunday** and
|
| 26 |
+
`1-6` mean Monday to Saturday. `7` must be accepted anywhere a value is
|
| 27 |
+
accepted, including inside ranges such as `6-7` (Saturday and Sunday), and it
|
| 28 |
+
must behave exactly like `0`. Internally a parsed day-of-week set is expressed
|
| 29 |
+
on the `0-6` scale, so `parse_field(DAY_OF_WEEK, "7") == {0}` and
|
| 30 |
+
`parse_field(DAY_OF_WEEK, "6-7") == {6, 0}`.
|
| 31 |
+
|
| 32 |
+
## Matching
|
| 33 |
+
|
| 34 |
+
`CronExpression.matches(moment)` is true when all five fields match. The
|
| 35 |
+
day-of-month and day-of-week fields are combined with AND (not the traditional
|
| 36 |
+
OR). `next_after(moment)` returns the first matching minute strictly after
|
| 37 |
+
`moment`, scanning forward up to 366 days.
|
| 38 |
+
|
| 39 |
+
## Tests
|
| 40 |
+
|
| 41 |
+
```
|
| 42 |
+
python -m pytest -q
|
| 43 |
+
```
|
repos/cronlite/cronlite/__init__.py
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""cronlite - a dependency-free five-field cron expression evaluator."""
|
| 2 |
+
|
| 3 |
+
from .expr import CronExpression, cron_weekday, next_after
|
| 4 |
+
from .fields import (
|
| 5 |
+
DAY_OF_MONTH,
|
| 6 |
+
DAY_OF_WEEK,
|
| 7 |
+
HOUR,
|
| 8 |
+
MINUTE,
|
| 9 |
+
MONTH,
|
| 10 |
+
FieldSpec,
|
| 11 |
+
parse_field,
|
| 12 |
+
)
|
| 13 |
+
|
| 14 |
+
__version__ = "0.2.3"
|
| 15 |
+
|
| 16 |
+
__all__ = [
|
| 17 |
+
"CronExpression",
|
| 18 |
+
"DAY_OF_MONTH",
|
| 19 |
+
"DAY_OF_WEEK",
|
| 20 |
+
"FieldSpec",
|
| 21 |
+
"HOUR",
|
| 22 |
+
"MINUTE",
|
| 23 |
+
"MONTH",
|
| 24 |
+
"cron_weekday",
|
| 25 |
+
"next_after",
|
| 26 |
+
"parse_field",
|
| 27 |
+
]
|
repos/cronlite/cronlite/expr.py
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""A five-field cron expression evaluator."""
|
| 2 |
+
|
| 3 |
+
from datetime import timedelta
|
| 4 |
+
|
| 5 |
+
from .fields import DAY_OF_MONTH, DAY_OF_WEEK, HOUR, MINUTE, MONTH, parse_field
|
| 6 |
+
|
| 7 |
+
SEARCH_LIMIT_DAYS = 366
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def cron_weekday(moment):
|
| 11 |
+
"""Return the cron weekday of ``moment``: 0 = Sunday, 6 = Saturday."""
|
| 12 |
+
return (moment.weekday() + 1) % 7
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
class CronExpression:
|
| 16 |
+
"""A parsed ``minute hour day-of-month month day-of-week`` expression."""
|
| 17 |
+
|
| 18 |
+
def __init__(self, expression):
|
| 19 |
+
parts = expression.split()
|
| 20 |
+
if len(parts) != 5:
|
| 21 |
+
raise ValueError("expected 5 fields, got %d" % (len(parts),))
|
| 22 |
+
self.expression = " ".join(parts)
|
| 23 |
+
self.minutes = parse_field(MINUTE, parts[0])
|
| 24 |
+
self.hours = parse_field(HOUR, parts[1])
|
| 25 |
+
self.days = parse_field(DAY_OF_MONTH, parts[2])
|
| 26 |
+
self.months = parse_field(MONTH, parts[3])
|
| 27 |
+
self.weekdays = parse_field(DAY_OF_WEEK, parts[4])
|
| 28 |
+
|
| 29 |
+
def matches(self, moment):
|
| 30 |
+
return (
|
| 31 |
+
moment.minute in self.minutes
|
| 32 |
+
and moment.hour in self.hours
|
| 33 |
+
and moment.day in self.days
|
| 34 |
+
and moment.month in self.months
|
| 35 |
+
and cron_weekday(moment) in self.weekdays
|
| 36 |
+
)
|
| 37 |
+
|
| 38 |
+
def next_after(self, moment):
|
| 39 |
+
"""Return the first matching minute strictly after ``moment``."""
|
| 40 |
+
current = moment.replace(second=0, microsecond=0) + timedelta(minutes=1)
|
| 41 |
+
limit = current + timedelta(days=SEARCH_LIMIT_DAYS)
|
| 42 |
+
while current <= limit:
|
| 43 |
+
if self.matches(current):
|
| 44 |
+
return current
|
| 45 |
+
current += timedelta(minutes=1)
|
| 46 |
+
raise ValueError("no matching time within %d days" % SEARCH_LIMIT_DAYS)
|
| 47 |
+
|
| 48 |
+
def __str__(self):
|
| 49 |
+
return self.expression
|
| 50 |
+
|
| 51 |
+
def __repr__(self):
|
| 52 |
+
return "CronExpression(%r)" % (self.expression,)
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
def next_after(expression, moment):
|
| 56 |
+
"""Convenience wrapper: ``next_after("*/5 * * * *", moment)``."""
|
| 57 |
+
if not isinstance(expression, CronExpression):
|
| 58 |
+
expression = CronExpression(expression)
|
| 59 |
+
return expression.next_after(moment)
|
repos/cronlite/cronlite/fields.py
ADDED
|
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Parsing of individual cron fields into sets of integers."""
|
| 2 |
+
|
| 3 |
+
|
| 4 |
+
class FieldSpec:
|
| 5 |
+
"""Describes one of the five cron fields."""
|
| 6 |
+
|
| 7 |
+
def __init__(self, name, minimum, maximum):
|
| 8 |
+
self.name = name
|
| 9 |
+
self.minimum = minimum
|
| 10 |
+
self.maximum = maximum
|
| 11 |
+
|
| 12 |
+
def _check(self, value):
|
| 13 |
+
if value < self.minimum or value > self.maximum:
|
| 14 |
+
raise ValueError("%s value out of range: %d" % (self.name, value))
|
| 15 |
+
return value
|
| 16 |
+
|
| 17 |
+
def parse(self, text):
|
| 18 |
+
values = set()
|
| 19 |
+
for part in text.split(","):
|
| 20 |
+
part = part.strip()
|
| 21 |
+
if not part:
|
| 22 |
+
raise ValueError("empty %s component" % self.name)
|
| 23 |
+
values |= self._parse_part(part)
|
| 24 |
+
if not values:
|
| 25 |
+
raise ValueError("empty %s field" % self.name)
|
| 26 |
+
return values
|
| 27 |
+
|
| 28 |
+
def _parse_part(self, part):
|
| 29 |
+
if part == "*":
|
| 30 |
+
return set(range(self.minimum, self.maximum + 1))
|
| 31 |
+
if "/" in part:
|
| 32 |
+
base, _, step_text = part.partition("/")
|
| 33 |
+
try:
|
| 34 |
+
step = int(step_text)
|
| 35 |
+
except ValueError:
|
| 36 |
+
raise ValueError("invalid step in %s field: %r" % (self.name, part))
|
| 37 |
+
if step <= 0:
|
| 38 |
+
raise ValueError("step must be positive in %s field" % self.name)
|
| 39 |
+
if base == "*":
|
| 40 |
+
start, stop = self.minimum, self.maximum
|
| 41 |
+
elif "-" in base:
|
| 42 |
+
start, stop = self._parse_bounds(base)
|
| 43 |
+
else:
|
| 44 |
+
start = self._check(int(base))
|
| 45 |
+
stop = start
|
| 46 |
+
return set(range(start, stop + 1, step))
|
| 47 |
+
if "-" in part:
|
| 48 |
+
start, stop = self._parse_bounds(part)
|
| 49 |
+
return set(range(start, stop + 1))
|
| 50 |
+
return {self._check(int(part))}
|
| 51 |
+
|
| 52 |
+
def _parse_bounds(self, text):
|
| 53 |
+
low, _, high = text.partition("-")
|
| 54 |
+
try:
|
| 55 |
+
start, stop = int(low), int(high)
|
| 56 |
+
except ValueError:
|
| 57 |
+
raise ValueError("invalid range in %s field: %r" % (self.name, text))
|
| 58 |
+
self._check(start)
|
| 59 |
+
self._check(stop)
|
| 60 |
+
if stop < start:
|
| 61 |
+
raise ValueError("descending range in %s field: %r" % (self.name, text))
|
| 62 |
+
return start, stop
|
| 63 |
+
|
| 64 |
+
|
| 65 |
+
MINUTE = FieldSpec("minute", 0, 59)
|
| 66 |
+
HOUR = FieldSpec("hour", 0, 23)
|
| 67 |
+
DAY_OF_MONTH = FieldSpec("day_of_month", 1, 31)
|
| 68 |
+
MONTH = FieldSpec("month", 1, 12)
|
| 69 |
+
DAY_OF_WEEK = FieldSpec("day_of_week", 0, 6)
|
| 70 |
+
|
| 71 |
+
|
| 72 |
+
def parse_field(spec, text):
|
| 73 |
+
"""Parse ``text`` against ``spec`` and return the matching values."""
|
| 74 |
+
return spec.parse(text)
|
repos/cronlite/tests/test_basic.py
ADDED
|
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from datetime import datetime
|
| 2 |
+
|
| 3 |
+
import pytest
|
| 4 |
+
|
| 5 |
+
from cronlite import CronExpression, cron_weekday, next_after
|
| 6 |
+
from cronlite.fields import DAY_OF_WEEK, HOUR, MINUTE, parse_field
|
| 7 |
+
|
| 8 |
+
|
| 9 |
+
def test_star_field():
|
| 10 |
+
assert parse_field(MINUTE, "*") == set(range(60))
|
| 11 |
+
|
| 12 |
+
|
| 13 |
+
def test_single_value_field():
|
| 14 |
+
assert parse_field(MINUTE, "17") == {17}
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
def test_list_field():
|
| 18 |
+
assert parse_field(MINUTE, "0,15,30,45") == {0, 15, 30, 45}
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
def test_list_with_range():
|
| 22 |
+
assert parse_field(MINUTE, "0,10-12") == {0, 10, 11, 12}
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
def test_range_field():
|
| 26 |
+
assert parse_field(HOUR, "9-11") == {9, 10, 11}
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
def test_star_step_field():
|
| 30 |
+
assert parse_field(MINUTE, "*/15") == {0, 15, 30, 45}
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
def test_range_step_field():
|
| 34 |
+
assert parse_field(MINUTE, "1-10/3") == {1, 4, 7, 10}
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
def test_out_of_range_rejected():
|
| 38 |
+
with pytest.raises(ValueError):
|
| 39 |
+
parse_field(MINUTE, "60")
|
| 40 |
+
with pytest.raises(ValueError):
|
| 41 |
+
parse_field(HOUR, "24")
|
| 42 |
+
|
| 43 |
+
|
| 44 |
+
def test_bad_step_rejected():
|
| 45 |
+
with pytest.raises(ValueError):
|
| 46 |
+
parse_field(MINUTE, "*/0")
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
def test_descending_range_rejected():
|
| 50 |
+
with pytest.raises(ValueError):
|
| 51 |
+
parse_field(MINUTE, "10-1")
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
def test_weekday_field_within_zero_to_six():
|
| 55 |
+
assert parse_field(DAY_OF_WEEK, "1-5") == {1, 2, 3, 4, 5}
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
def test_cron_weekday_mapping():
|
| 59 |
+
assert cron_weekday(datetime(2026, 3, 4)) == 3
|
| 60 |
+
assert cron_weekday(datetime(2026, 3, 8)) == 0
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
def test_matches_every_minute():
|
| 64 |
+
expr = CronExpression("* * * * *")
|
| 65 |
+
assert expr.matches(datetime(2026, 3, 4, 5, 6))
|
| 66 |
+
|
| 67 |
+
|
| 68 |
+
def test_matches_rejects_wrong_minute():
|
| 69 |
+
expr = CronExpression("30 9 * * *")
|
| 70 |
+
assert not expr.matches(datetime(2026, 3, 4, 9, 31))
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
def test_next_after_basic():
|
| 74 |
+
expr = CronExpression("30 9 * * *")
|
| 75 |
+
assert next_after(expr, datetime(2026, 3, 4, 5, 6)) == datetime(2026, 3, 4, 9, 30)
|
| 76 |
+
|
| 77 |
+
|
| 78 |
+
def test_next_after_is_strict():
|
| 79 |
+
expr = CronExpression("30 9 * * *")
|
| 80 |
+
assert next_after(expr, datetime(2026, 3, 4, 9, 30)) == datetime(2026, 3, 5, 9, 30)
|
| 81 |
+
|
| 82 |
+
|
| 83 |
+
def test_next_after_rolls_to_next_day():
|
| 84 |
+
expr = CronExpression("0 0 * * *")
|
| 85 |
+
assert next_after(expr, datetime(2026, 3, 4, 0, 0)) == datetime(2026, 3, 5, 0, 0)
|
| 86 |
+
|
| 87 |
+
|
| 88 |
+
def test_next_after_accepts_string():
|
| 89 |
+
assert next_after("*/30 * * * *", datetime(2026, 3, 4, 5, 6)) == datetime(2026, 3, 4, 5, 30)
|
| 90 |
+
|
| 91 |
+
|
| 92 |
+
def test_invalid_field_count():
|
| 93 |
+
with pytest.raises(ValueError):
|
| 94 |
+
CronExpression("* * * *")
|
| 95 |
+
|
| 96 |
+
|
| 97 |
+
def test_str_roundtrip():
|
| 98 |
+
assert str(CronExpression(" 0 12 * * * ")) == "0 12 * * *"
|
repos/miniconf/README.md
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# miniconf
|
| 2 |
+
|
| 3 |
+
A tiny layered configuration library. Values may reference other keys with
|
| 4 |
+
`${key}` and are resolved lazily on access.
|
| 5 |
+
|
| 6 |
+
## Interpolation contract
|
| 7 |
+
|
| 8 |
+
This is the documented behaviour of `Config`; `tests/` covers the common cases.
|
| 9 |
+
|
| 10 |
+
1. **References.** A string value may contain any number of `${key}` references.
|
| 11 |
+
The reference is replaced by the *fully resolved* value of `key`, stringified
|
| 12 |
+
with `str()`. References may chain: `a` -> `b` -> `c`.
|
| 13 |
+
2. **Single pass.** Substituted text is never scanned again for references. If
|
| 14 |
+
`inner` resolves to the literal text `${leak}`, then `"${inner}"` resolves to
|
| 15 |
+
`${leak}` and not to the value of `leak`.
|
| 16 |
+
3. **Escaping.** `$$` is an escape that produces a single literal `$`. It never
|
| 17 |
+
starts a reference, so `"$${amount}"` resolves to the five characters
|
| 18 |
+
`${amount}` — the reference is *not* expanded. `"$${name} is ${name}"` with
|
| 19 |
+
`name=wren` resolves to `"${name} is wren"`.
|
| 20 |
+
4. **Cycles.** A reference cycle, including a self reference such as
|
| 21 |
+
`a = "x${a}"`, must raise `CircularReferenceError` instead of recursing. The
|
| 22 |
+
message is `"circular reference: "` followed by the key chain joined with
|
| 23 |
+
`" -> "`, starting at the key the caller asked for and ending with that same
|
| 24 |
+
key repeated. Asking for `a` where `a -> b -> a` raises
|
| 25 |
+
`"circular reference: a -> b -> a"`; asking for `a` where `a -> a` raises
|
| 26 |
+
`"circular reference: a -> a"`.
|
| 27 |
+
5. **Missing keys.** A reference to an unknown key raises `MissingKeyError`.
|
| 28 |
+
6. **Non-string values** are returned as-is and are never scanned.
|
| 29 |
+
|
| 30 |
+
`Config.as_dict()` resolves every key, so any cycle or missing reference in the
|
| 31 |
+
mapping surfaces there too. A cycle that is unrelated to the key being read must
|
| 32 |
+
not affect that read: with `a="fine"`, `b="${c}"`, `c="${b}"`, reading `a` still
|
| 33 |
+
returns `"fine"` while reading `b` raises.
|
| 34 |
+
|
| 35 |
+
## Usage
|
| 36 |
+
|
| 37 |
+
```python
|
| 38 |
+
from miniconf import Config
|
| 39 |
+
|
| 40 |
+
cfg = Config.from_layers(
|
| 41 |
+
{"host": "localhost", "port": 8080},
|
| 42 |
+
{"url": "http://${host}:${port}/"},
|
| 43 |
+
)
|
| 44 |
+
cfg["url"] # 'http://localhost:8080/'
|
| 45 |
+
cfg.as_dict() # {'host': 'localhost', 'port': 8080, 'url': 'http://localhost:8080/'}
|
| 46 |
+
```
|
| 47 |
+
|
| 48 |
+
## Tests
|
| 49 |
+
|
| 50 |
+
```
|
| 51 |
+
python -m pytest -q
|
| 52 |
+
```
|
repos/miniconf/miniconf/__init__.py
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""miniconf - a tiny layered configuration library with string interpolation."""
|
| 2 |
+
|
| 3 |
+
from .config import Config
|
| 4 |
+
from .errors import (
|
| 5 |
+
CircularReferenceError,
|
| 6 |
+
ConfigError,
|
| 7 |
+
InterpolationError,
|
| 8 |
+
MissingKeyError,
|
| 9 |
+
)
|
| 10 |
+
from .loader import load_layers
|
| 11 |
+
|
| 12 |
+
__version__ = "0.4.1"
|
| 13 |
+
|
| 14 |
+
__all__ = [
|
| 15 |
+
"CircularReferenceError",
|
| 16 |
+
"Config",
|
| 17 |
+
"ConfigError",
|
| 18 |
+
"InterpolationError",
|
| 19 |
+
"MissingKeyError",
|
| 20 |
+
"load_layers",
|
| 21 |
+
]
|
repos/miniconf/miniconf/config.py
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""The Config object: layered key/value storage with ${...} interpolation."""
|
| 2 |
+
|
| 3 |
+
import re
|
| 4 |
+
|
| 5 |
+
from .errors import MissingKeyError
|
| 6 |
+
from .loader import load_layers
|
| 7 |
+
|
| 8 |
+
REFERENCE = re.compile(r"\$\{([A-Za-z_][A-Za-z0-9_.-]*)\}")
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
class Config:
|
| 12 |
+
"""A small configuration mapping whose string values may reference keys.
|
| 13 |
+
|
| 14 |
+
See README.md for the interpolation contract.
|
| 15 |
+
"""
|
| 16 |
+
|
| 17 |
+
def __init__(self, data):
|
| 18 |
+
if not isinstance(data, dict):
|
| 19 |
+
raise TypeError("Config data must be a dict")
|
| 20 |
+
self._raw = dict(data)
|
| 21 |
+
|
| 22 |
+
@classmethod
|
| 23 |
+
def from_layers(cls, *layers):
|
| 24 |
+
"""Build a Config from one or more dict layers, later layers winning."""
|
| 25 |
+
return cls(load_layers(*layers))
|
| 26 |
+
|
| 27 |
+
def raw(self, key, default=None):
|
| 28 |
+
"""Return the unresolved value stored under ``key``."""
|
| 29 |
+
return self._raw.get(key, default)
|
| 30 |
+
|
| 31 |
+
def keys(self):
|
| 32 |
+
return list(self._raw)
|
| 33 |
+
|
| 34 |
+
def get(self, key, default=None):
|
| 35 |
+
if key not in self._raw:
|
| 36 |
+
return default
|
| 37 |
+
return self._resolve(key)
|
| 38 |
+
|
| 39 |
+
def __getitem__(self, key):
|
| 40 |
+
if key not in self._raw:
|
| 41 |
+
raise KeyError(key)
|
| 42 |
+
return self._resolve(key)
|
| 43 |
+
|
| 44 |
+
def __contains__(self, key):
|
| 45 |
+
return key in self._raw
|
| 46 |
+
|
| 47 |
+
def __len__(self):
|
| 48 |
+
return len(self._raw)
|
| 49 |
+
|
| 50 |
+
def as_dict(self):
|
| 51 |
+
"""Return every key resolved to a plain value."""
|
| 52 |
+
return {key: self._resolve(key) for key in self._raw}
|
| 53 |
+
|
| 54 |
+
def _resolve(self, key):
|
| 55 |
+
value = self._raw[key]
|
| 56 |
+
if not isinstance(value, str):
|
| 57 |
+
return value
|
| 58 |
+
return self._interpolate(value, key)
|
| 59 |
+
|
| 60 |
+
def _interpolate(self, text, origin):
|
| 61 |
+
def replace(match):
|
| 62 |
+
name = match.group(1)
|
| 63 |
+
if name not in self._raw:
|
| 64 |
+
raise MissingKeyError(name, origin)
|
| 65 |
+
return str(self._resolve(name))
|
| 66 |
+
|
| 67 |
+
return REFERENCE.sub(replace, text)
|
repos/miniconf/miniconf/errors.py
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Exception hierarchy for miniconf."""
|
| 2 |
+
|
| 3 |
+
|
| 4 |
+
class ConfigError(Exception):
|
| 5 |
+
"""Base class for every error raised by miniconf."""
|
| 6 |
+
|
| 7 |
+
|
| 8 |
+
class InterpolationError(ConfigError):
|
| 9 |
+
"""Raised when a ``${...}`` reference cannot be resolved."""
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
class MissingKeyError(InterpolationError):
|
| 13 |
+
"""Raised when a reference points at a key that does not exist."""
|
| 14 |
+
|
| 15 |
+
def __init__(self, key, referenced_from=None):
|
| 16 |
+
self.key = key
|
| 17 |
+
self.referenced_from = referenced_from
|
| 18 |
+
if referenced_from is None:
|
| 19 |
+
message = "missing key: %r" % (key,)
|
| 20 |
+
else:
|
| 21 |
+
message = "missing key %r referenced from %r" % (key, referenced_from)
|
| 22 |
+
super().__init__(message)
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
class CircularReferenceError(InterpolationError):
|
| 26 |
+
"""Raised when interpolation would loop forever."""
|
| 27 |
+
|
| 28 |
+
def __init__(self, cycle):
|
| 29 |
+
self.cycle = list(cycle)
|
| 30 |
+
super().__init__("circular reference: " + " -> ".join(self.cycle))
|
repos/miniconf/miniconf/loader.py
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Layer merging for miniconf configuration dicts."""
|
| 2 |
+
|
| 3 |
+
from .errors import ConfigError
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
def load_layers(*layers):
|
| 7 |
+
"""Merge configuration layers left to right; later layers win."""
|
| 8 |
+
merged = {}
|
| 9 |
+
for layer in layers:
|
| 10 |
+
if not isinstance(layer, dict):
|
| 11 |
+
raise ConfigError("layer must be a dict, got %r" % (type(layer).__name__,))
|
| 12 |
+
for key, value in layer.items():
|
| 13 |
+
merged[key] = value
|
| 14 |
+
return merged
|
repos/miniconf/tests/test_basic.py
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import pytest
|
| 2 |
+
|
| 3 |
+
from miniconf import Config, ConfigError, MissingKeyError, load_layers
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
def test_plain_values():
|
| 7 |
+
cfg = Config({"name": "app", "port": 8080})
|
| 8 |
+
assert cfg["name"] == "app"
|
| 9 |
+
assert cfg["port"] == 8080
|
| 10 |
+
assert len(cfg) == 2
|
| 11 |
+
|
| 12 |
+
|
| 13 |
+
def test_simple_reference():
|
| 14 |
+
cfg = Config({"host": "localhost", "url": "http://${host}/"})
|
| 15 |
+
assert cfg["url"] == "http://localhost/"
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
def test_two_references_in_one_value():
|
| 19 |
+
cfg = Config({"host": "db", "port": 5432, "dsn": "${host}:${port}"})
|
| 20 |
+
assert cfg["dsn"] == "db:5432"
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
def test_chained_reference():
|
| 24 |
+
cfg = Config({"a": "x", "b": "${a}-y", "c": "${b}-z"})
|
| 25 |
+
assert cfg["c"] == "x-y-z"
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
def test_missing_key_raises():
|
| 29 |
+
cfg = Config({"a": "${nope}"})
|
| 30 |
+
with pytest.raises(MissingKeyError):
|
| 31 |
+
cfg["a"]
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
def test_missing_key_message():
|
| 35 |
+
cfg = Config({"a": "${nope}"})
|
| 36 |
+
with pytest.raises(MissingKeyError) as excinfo:
|
| 37 |
+
cfg["a"]
|
| 38 |
+
assert str(excinfo.value) == "missing key 'nope' referenced from 'a'"
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
def test_non_string_passthrough():
|
| 42 |
+
cfg = Config({"n": 3, "flag": True, "nested": {"x": 1}})
|
| 43 |
+
assert cfg["n"] == 3
|
| 44 |
+
assert cfg["flag"] is True
|
| 45 |
+
assert cfg["nested"] == {"x": 1}
|
| 46 |
+
|
| 47 |
+
|
| 48 |
+
def test_get_returns_default_for_unknown_key():
|
| 49 |
+
cfg = Config({"a": "1"})
|
| 50 |
+
assert cfg.get("zzz") is None
|
| 51 |
+
assert cfg.get("zzz", "dflt") == "dflt"
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
def test_getitem_raises_keyerror_for_unknown_key():
|
| 55 |
+
cfg = Config({"a": "1"})
|
| 56 |
+
with pytest.raises(KeyError):
|
| 57 |
+
cfg["zzz"]
|
| 58 |
+
|
| 59 |
+
|
| 60 |
+
def test_raw_does_not_resolve():
|
| 61 |
+
cfg = Config({"a": "1", "b": "${a}"})
|
| 62 |
+
assert cfg.raw("b") == "${a}"
|
| 63 |
+
|
| 64 |
+
|
| 65 |
+
def test_from_layers_override():
|
| 66 |
+
cfg = Config.from_layers({"a": "1", "b": "2"}, {"b": "3"})
|
| 67 |
+
assert cfg.as_dict() == {"a": "1", "b": "3"}
|
| 68 |
+
|
| 69 |
+
|
| 70 |
+
def test_load_layers_rejects_non_dict():
|
| 71 |
+
with pytest.raises(ConfigError):
|
| 72 |
+
load_layers({"a": 1}, ["not", "a", "dict"])
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
def test_as_dict_resolves_all():
|
| 76 |
+
cfg = Config({"a": "1", "b": "${a}2"})
|
| 77 |
+
assert cfg.as_dict() == {"a": "1", "b": "12"}
|
| 78 |
+
|
| 79 |
+
|
| 80 |
+
def test_config_rejects_non_dict():
|
| 81 |
+
with pytest.raises(TypeError):
|
| 82 |
+
Config([("a", 1)])
|
repos/semverlite/README.md
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# semverlite
|
| 2 |
+
|
| 3 |
+
Semantic version parsing, precedence and range matching, following
|
| 4 |
+
[semver 2.0.0](https://semver.org/).
|
| 5 |
+
|
| 6 |
+
## Precedence
|
| 7 |
+
|
| 8 |
+
`Version` objects sort by semver precedence:
|
| 9 |
+
|
| 10 |
+
1. Compare `major`, then `minor`, then `patch` numerically.
|
| 11 |
+
2. A version **with** a prerelease is lower than the same version without one
|
| 12 |
+
(`1.0.0-alpha < 1.0.0`).
|
| 13 |
+
3. Prerelease identifiers are compared left to right. A **numeric** identifier
|
| 14 |
+
always compares numerically, so `1.0.0-alpha.2 < 1.0.0-alpha.10`. An
|
| 15 |
+
alphanumeric identifier compares by ASCII order and is always **greater**
|
| 16 |
+
than a numeric identifier of the same position (`1.0.0-1 < 1.0.0-alpha`).
|
| 17 |
+
A longer prerelease wins when all shared identifiers are equal
|
| 18 |
+
(`1.0.0-alpha < 1.0.0-alpha.1`).
|
| 19 |
+
4. Build metadata is ignored for precedence, so `1.0.0+a == 1.0.0+b`.
|
| 20 |
+
|
| 21 |
+
## Ranges
|
| 22 |
+
|
| 23 |
+
`VersionRange(text)` accepts a comma-separated conjunction of clauses using
|
| 24 |
+
`=`, `==`, `>`, `>=`, `<`, `<=`, `^` and `~`. `^1.2.3` means
|
| 25 |
+
`>=1.2.3, <2.0.0`; `~1.2.3` means `>=1.2.3, <1.3.0`.
|
| 26 |
+
|
| 27 |
+
**Prereleases are excluded from a range unless the range itself opts in.** A
|
| 28 |
+
version carrying a prerelease satisfies the range only when some clause names a
|
| 29 |
+
prerelease with the *same* `major.minor.patch` triple. So `>=1.0.0` does not
|
| 30 |
+
allow `1.0.1-alpha`, while `>=1.0.0-alpha` does allow `1.0.0-beta`.
|
| 31 |
+
|
| 32 |
+
## Usage
|
| 33 |
+
|
| 34 |
+
```python
|
| 35 |
+
from semverlite import Version, VersionRange
|
| 36 |
+
|
| 37 |
+
Version.parse("1.2.3-alpha.1") < Version.parse("1.2.3")
|
| 38 |
+
Version.parse("1.2.3") in VersionRange(">=1.0.0, <2.0.0")
|
| 39 |
+
```
|
| 40 |
+
|
| 41 |
+
## Tests
|
| 42 |
+
|
| 43 |
+
```
|
| 44 |
+
python -m pytest -q
|
| 45 |
+
```
|
repos/semverlite/semverlite/__init__.py
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""semverlite - semantic version parsing, precedence and ranges."""
|
| 2 |
+
|
| 3 |
+
from .range import VersionRange
|
| 4 |
+
from .version import Version
|
| 5 |
+
|
| 6 |
+
__version__ = "0.5.2"
|
| 7 |
+
|
| 8 |
+
__all__ = ["Version", "VersionRange"]
|
repos/semverlite/semverlite/range.py
ADDED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Version range expressions such as ``">=1.2.3, <2.0.0"``."""
|
| 2 |
+
|
| 3 |
+
import re
|
| 4 |
+
|
| 5 |
+
from .version import Version
|
| 6 |
+
|
| 7 |
+
_CLAUSE_RE = re.compile(r"^(>=|<=|==|=|>|<|\^|~)?\s*(\S+)$")
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def _parse_clause(text):
|
| 11 |
+
match = _CLAUSE_RE.match(text)
|
| 12 |
+
if match is None:
|
| 13 |
+
raise ValueError("invalid range clause: %r" % (text,))
|
| 14 |
+
operator = match.group(1) or "=="
|
| 15 |
+
return operator, Version.parse(match.group(2))
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
def _caret_upper(target):
|
| 19 |
+
"""Upper bound of ``^target``: the next release that breaks the left-most
|
| 20 |
+
non-zero component."""
|
| 21 |
+
if target.major > 0:
|
| 22 |
+
return Version(target.major + 1, 0, 0)
|
| 23 |
+
if target.minor > 0:
|
| 24 |
+
return Version(0, target.minor + 1, 0)
|
| 25 |
+
return Version(0, 0, target.patch + 1)
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
def _tilde_upper(target):
|
| 29 |
+
"""Upper bound of ``~target``: the next minor release."""
|
| 30 |
+
return Version(target.major, target.minor + 1, 0)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
class VersionRange:
|
| 34 |
+
"""A comma-separated list of clauses that must all hold."""
|
| 35 |
+
|
| 36 |
+
def __init__(self, text):
|
| 37 |
+
self.text = text.strip()
|
| 38 |
+
parts = [part.strip() for part in self.text.split(",") if part.strip()]
|
| 39 |
+
if not parts:
|
| 40 |
+
raise ValueError("empty range")
|
| 41 |
+
self.clauses = [_parse_clause(part) for part in parts]
|
| 42 |
+
|
| 43 |
+
def allows(self, version):
|
| 44 |
+
return all(self._allows_clause(op, target, version) for op, target in self.clauses)
|
| 45 |
+
|
| 46 |
+
@staticmethod
|
| 47 |
+
def _allows_clause(operator, target, version):
|
| 48 |
+
if operator in ("=", "=="):
|
| 49 |
+
return version == target
|
| 50 |
+
if operator == ">":
|
| 51 |
+
return version > target
|
| 52 |
+
if operator == ">=":
|
| 53 |
+
return version >= target
|
| 54 |
+
if operator == "<":
|
| 55 |
+
return version < target
|
| 56 |
+
if operator == "<=":
|
| 57 |
+
return version <= target
|
| 58 |
+
if operator == "^":
|
| 59 |
+
return target <= version < _caret_upper(target)
|
| 60 |
+
if operator == "~":
|
| 61 |
+
return target <= version < _tilde_upper(target)
|
| 62 |
+
raise ValueError("unknown operator: %r" % (operator,))
|
| 63 |
+
|
| 64 |
+
def __contains__(self, version):
|
| 65 |
+
return self.allows(version)
|
| 66 |
+
|
| 67 |
+
def __str__(self):
|
| 68 |
+
return self.text
|
| 69 |
+
|
| 70 |
+
def __repr__(self):
|
| 71 |
+
return "VersionRange(%r)" % (self.text,)
|
repos/semverlite/semverlite/version.py
ADDED
|
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Semantic version parsing and precedence."""
|
| 2 |
+
|
| 3 |
+
import re
|
| 4 |
+
|
| 5 |
+
_VERSION_RE = re.compile(
|
| 6 |
+
r"^(?P<major>0|[1-9]\d*)\.(?P<minor>0|[1-9]\d*)\.(?P<patch>0|[1-9]\d*)"
|
| 7 |
+
r"(?:-(?P<pre>[0-9A-Za-z.-]+))?"
|
| 8 |
+
r"(?:\+(?P<build>[0-9A-Za-z.-]+))?$"
|
| 9 |
+
)
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
class Version:
|
| 13 |
+
"""A semantic version: major.minor.patch[-prerelease][+build]."""
|
| 14 |
+
|
| 15 |
+
def __init__(self, major, minor, patch, prerelease=(), build=()):
|
| 16 |
+
self.major = int(major)
|
| 17 |
+
self.minor = int(minor)
|
| 18 |
+
self.patch = int(patch)
|
| 19 |
+
self.prerelease = tuple(prerelease)
|
| 20 |
+
self.build = tuple(build)
|
| 21 |
+
|
| 22 |
+
@classmethod
|
| 23 |
+
def parse(cls, text):
|
| 24 |
+
match = _VERSION_RE.match(text.strip())
|
| 25 |
+
if match is None:
|
| 26 |
+
raise ValueError("invalid version: %r" % (text,))
|
| 27 |
+
prerelease = tuple(match.group("pre").split(".")) if match.group("pre") else ()
|
| 28 |
+
build = tuple(match.group("build").split(".")) if match.group("build") else ()
|
| 29 |
+
return cls(
|
| 30 |
+
match.group("major"),
|
| 31 |
+
match.group("minor"),
|
| 32 |
+
match.group("patch"),
|
| 33 |
+
prerelease,
|
| 34 |
+
build,
|
| 35 |
+
)
|
| 36 |
+
|
| 37 |
+
@property
|
| 38 |
+
def core(self):
|
| 39 |
+
"""The (major, minor, patch) triple."""
|
| 40 |
+
return (self.major, self.minor, self.patch)
|
| 41 |
+
|
| 42 |
+
def precedence_key(self):
|
| 43 |
+
"""Sort key implementing semver precedence (build metadata ignored)."""
|
| 44 |
+
core = self.core
|
| 45 |
+
if not self.prerelease:
|
| 46 |
+
return core + (1, ())
|
| 47 |
+
return core + (0, tuple(str(part) for part in self.prerelease))
|
| 48 |
+
|
| 49 |
+
def __str__(self):
|
| 50 |
+
text = "%d.%d.%d" % self.core
|
| 51 |
+
if self.prerelease:
|
| 52 |
+
text += "-" + ".".join(str(part) for part in self.prerelease)
|
| 53 |
+
if self.build:
|
| 54 |
+
text += "+" + ".".join(str(part) for part in self.build)
|
| 55 |
+
return text
|
| 56 |
+
|
| 57 |
+
def __repr__(self):
|
| 58 |
+
return "Version(%r)" % (str(self),)
|
| 59 |
+
|
| 60 |
+
def __eq__(self, other):
|
| 61 |
+
if not isinstance(other, Version):
|
| 62 |
+
return NotImplemented
|
| 63 |
+
return self.precedence_key() == other.precedence_key()
|
| 64 |
+
|
| 65 |
+
def __ne__(self, other):
|
| 66 |
+
result = self.__eq__(other)
|
| 67 |
+
if result is NotImplemented:
|
| 68 |
+
return result
|
| 69 |
+
return not result
|
| 70 |
+
|
| 71 |
+
def __lt__(self, other):
|
| 72 |
+
return self.precedence_key() < other.precedence_key()
|
| 73 |
+
|
| 74 |
+
def __le__(self, other):
|
| 75 |
+
return self.precedence_key() <= other.precedence_key()
|
| 76 |
+
|
| 77 |
+
def __gt__(self, other):
|
| 78 |
+
return self.precedence_key() > other.precedence_key()
|
| 79 |
+
|
| 80 |
+
def __ge__(self, other):
|
| 81 |
+
return self.precedence_key() >= other.precedence_key()
|
| 82 |
+
|
| 83 |
+
def __hash__(self):
|
| 84 |
+
return hash(self.precedence_key())
|
repos/semverlite/tests/test_version.py
ADDED
|
@@ -0,0 +1,106 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import pytest
|
| 2 |
+
|
| 3 |
+
from semverlite import Version, VersionRange
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
def v(text):
|
| 7 |
+
return Version.parse(text)
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def test_parse_basic():
|
| 11 |
+
version = v("1.2.3")
|
| 12 |
+
assert version.major == 1
|
| 13 |
+
assert version.minor == 2
|
| 14 |
+
assert version.patch == 3
|
| 15 |
+
assert version.prerelease == ()
|
| 16 |
+
assert version.build == ()
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
def test_parse_prerelease_and_build():
|
| 20 |
+
version = v("1.2.3-rc.1+build.5")
|
| 21 |
+
assert version.prerelease == ("rc", "1")
|
| 22 |
+
assert version.build == ("build", "5")
|
| 23 |
+
assert version.core == (1, 2, 3)
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
def test_str_roundtrip():
|
| 27 |
+
for text in ("0.0.1", "1.2.3-alpha", "1.2.3-alpha.1+build.5"):
|
| 28 |
+
assert str(v(text)) == text
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
def test_invalid_versions_rejected():
|
| 32 |
+
for text in ("1.2", "1.2.3.4", "v1.2.3", "1.2.x"):
|
| 33 |
+
with pytest.raises(ValueError):
|
| 34 |
+
Version.parse(text)
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
def test_ordering_by_patch():
|
| 38 |
+
assert v("1.2.3") < v("1.2.4")
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
def test_ordering_by_minor():
|
| 42 |
+
assert v("1.2.9") < v("1.3.0")
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
def test_ordering_by_major():
|
| 46 |
+
assert v("1.9.9") < v("2.0.0")
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
def test_release_beats_prerelease():
|
| 50 |
+
assert v("1.0.0-alpha") < v("1.0.0")
|
| 51 |
+
assert v("1.0.0") > v("1.0.0-rc.9")
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
def test_shorter_prerelease_is_lower():
|
| 55 |
+
assert v("1.0.0-alpha") < v("1.0.0-alpha.1")
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
def test_equality_ignores_build_metadata():
|
| 59 |
+
assert v("1.0.0+a") == v("1.0.0+b")
|
| 60 |
+
assert v("1.0.0-alpha+x") == v("1.0.0-alpha+y")
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
def test_equality_and_hash():
|
| 64 |
+
assert v("1.2.3") == v("1.2.3")
|
| 65 |
+
assert len({v("1.2.3"), v("1.2.3")}) == 1
|
| 66 |
+
|
| 67 |
+
|
| 68 |
+
def test_inequality():
|
| 69 |
+
assert v("1.2.3") != v("1.2.4")
|
| 70 |
+
|
| 71 |
+
|
| 72 |
+
def test_range_lower_and_upper_bound():
|
| 73 |
+
version_range = VersionRange(">=1.0.0, <2.0.0")
|
| 74 |
+
assert version_range.allows(v("1.0.0"))
|
| 75 |
+
assert version_range.allows(v("1.5.0"))
|
| 76 |
+
assert not version_range.allows(v("2.0.0"))
|
| 77 |
+
assert not version_range.allows(v("0.9.9"))
|
| 78 |
+
|
| 79 |
+
|
| 80 |
+
def test_range_membership_operator():
|
| 81 |
+
assert v("1.5.0") in VersionRange(">=1.0.0, <2.0.0")
|
| 82 |
+
assert v("2.5.0") not in VersionRange(">=1.0.0, <2.0.0")
|
| 83 |
+
|
| 84 |
+
|
| 85 |
+
def test_caret_range():
|
| 86 |
+
version_range = VersionRange("^1.2.3")
|
| 87 |
+
assert version_range.allows(v("1.2.3"))
|
| 88 |
+
assert version_range.allows(v("1.9.9"))
|
| 89 |
+
assert not version_range.allows(v("2.0.0"))
|
| 90 |
+
assert not version_range.allows(v("1.2.2"))
|
| 91 |
+
|
| 92 |
+
|
| 93 |
+
def test_tilde_range():
|
| 94 |
+
version_range = VersionRange("~1.2.3")
|
| 95 |
+
assert version_range.allows(v("1.2.9"))
|
| 96 |
+
assert not version_range.allows(v("1.3.0"))
|
| 97 |
+
|
| 98 |
+
|
| 99 |
+
def test_exact_clause():
|
| 100 |
+
assert VersionRange("==1.2.3").allows(v("1.2.3"))
|
| 101 |
+
assert not VersionRange("==1.2.3").allows(v("1.2.4"))
|
| 102 |
+
|
| 103 |
+
|
| 104 |
+
def test_invalid_clause_rejected():
|
| 105 |
+
with pytest.raises(ValueError):
|
| 106 |
+
VersionRange(">=nope")
|
repos/ttlcache/README.md
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# ttlcache
|
| 2 |
+
|
| 3 |
+
A bounded cache with per-entry time-to-live and LRU eviction. The cache never
|
| 4 |
+
reads the system clock directly; pass a `clock` object exposing `now()` (see
|
| 5 |
+
`ttlcache.clock.FakeClock`) so expiry is deterministic under test.
|
| 6 |
+
|
| 7 |
+
## Contract
|
| 8 |
+
|
| 9 |
+
1. **Expiry.** An entry written at time `t` is live while `now < t + ttl` and
|
| 10 |
+
expired once `now >= t + ttl`. Reading an expired key deletes it and raises
|
| 11 |
+
`KeyError` (`get` returns its default instead). `key in cache` is `False` for
|
| 12 |
+
expired entries.
|
| 13 |
+
2. **Recency.** `__getitem__`, `get` and `setdefault` mark the entry as most
|
| 14 |
+
recently used. Writing an existing key also moves it to the most-recently-used
|
| 15 |
+
end *and* restarts its TTL.
|
| 16 |
+
3. **Eviction.** When a write would exceed `maxsize`, expired entries are dropped
|
| 17 |
+
first; if the cache is still full, the least recently used live entry is
|
| 18 |
+
dropped.
|
| 19 |
+
4. **`len(cache)`** counts live entries only. Expired-but-not-yet-collected
|
| 20 |
+
entries must not inflate the count.
|
| 21 |
+
5. **`items()`** yields live pairs in LRU order (least recently used first).
|
| 22 |
+
|
| 23 |
+
## Usage
|
| 24 |
+
|
| 25 |
+
```python
|
| 26 |
+
from ttlcache import FakeClock, TTLCache
|
| 27 |
+
|
| 28 |
+
clock = FakeClock()
|
| 29 |
+
cache = TTLCache(maxsize=2, ttl=10.0, clock=clock)
|
| 30 |
+
cache["a"] = 1
|
| 31 |
+
cache["b"] = 2
|
| 32 |
+
cache["a"] # 'a' is now most recently used
|
| 33 |
+
cache["c"] = 3 # evicts 'b', not 'a'
|
| 34 |
+
```
|
| 35 |
+
|
| 36 |
+
## Tests
|
| 37 |
+
|
| 38 |
+
```
|
| 39 |
+
python -m pytest -q
|
| 40 |
+
```
|
repos/ttlcache/tests/test_cache.py
ADDED
|
@@ -0,0 +1,111 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import pytest
|
| 2 |
+
|
| 3 |
+
from ttlcache import FakeClock, TTLCache
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
def make(clock, maxsize=3, ttl=10.0):
|
| 7 |
+
return TTLCache(maxsize=maxsize, ttl=ttl, clock=clock)
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
def test_set_and_get():
|
| 11 |
+
clock = FakeClock()
|
| 12 |
+
cache = make(clock)
|
| 13 |
+
cache["a"] = 1
|
| 14 |
+
assert cache["a"] == 1
|
| 15 |
+
assert cache.get("a") == 1
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
def test_missing_key_returns_default():
|
| 19 |
+
cache = make(FakeClock())
|
| 20 |
+
assert cache.get("nope") is None
|
| 21 |
+
assert cache.get("nope", 7) == 7
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
def test_getitem_raises_for_missing_key():
|
| 25 |
+
cache = make(FakeClock())
|
| 26 |
+
with pytest.raises(KeyError):
|
| 27 |
+
cache["nope"]
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
def test_entry_expires():
|
| 31 |
+
clock = FakeClock()
|
| 32 |
+
cache = make(clock)
|
| 33 |
+
cache["a"] = 1
|
| 34 |
+
clock.advance(10)
|
| 35 |
+
assert "a" not in cache
|
| 36 |
+
assert cache.get("a") is None
|
| 37 |
+
with pytest.raises(KeyError):
|
| 38 |
+
cache["a"]
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
def test_entry_survives_before_deadline():
|
| 42 |
+
clock = FakeClock()
|
| 43 |
+
cache = make(clock)
|
| 44 |
+
cache["a"] = 1
|
| 45 |
+
clock.advance(9.5)
|
| 46 |
+
assert cache["a"] == 1
|
| 47 |
+
|
| 48 |
+
|
| 49 |
+
def test_eviction_order_on_insert():
|
| 50 |
+
clock = FakeClock()
|
| 51 |
+
cache = make(clock, maxsize=2)
|
| 52 |
+
cache["a"] = 1
|
| 53 |
+
cache["b"] = 2
|
| 54 |
+
cache["c"] = 3
|
| 55 |
+
assert "a" not in cache
|
| 56 |
+
assert "b" in cache and "c" in cache
|
| 57 |
+
|
| 58 |
+
|
| 59 |
+
def test_overwrite_updates_value():
|
| 60 |
+
clock = FakeClock()
|
| 61 |
+
cache = make(clock)
|
| 62 |
+
cache["a"] = 1
|
| 63 |
+
cache["a"] = 2
|
| 64 |
+
assert cache["a"] == 2
|
| 65 |
+
assert len(cache) == 1
|
| 66 |
+
|
| 67 |
+
|
| 68 |
+
def test_setdefault_uses_factory_once():
|
| 69 |
+
clock = FakeClock()
|
| 70 |
+
cache = make(clock)
|
| 71 |
+
calls = []
|
| 72 |
+
|
| 73 |
+
def factory():
|
| 74 |
+
calls.append(1)
|
| 75 |
+
return "v"
|
| 76 |
+
|
| 77 |
+
assert cache.setdefault("k", factory) == "v"
|
| 78 |
+
assert cache.setdefault("k", factory) == "v"
|
| 79 |
+
assert len(calls) == 1
|
| 80 |
+
|
| 81 |
+
|
| 82 |
+
def test_pop_removes_entry():
|
| 83 |
+
clock = FakeClock()
|
| 84 |
+
cache = make(clock)
|
| 85 |
+
cache["a"] = 1
|
| 86 |
+
assert cache.pop("a") == 1
|
| 87 |
+
assert cache.pop("a", "gone") == "gone"
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
def test_items_skips_expired():
|
| 91 |
+
clock = FakeClock()
|
| 92 |
+
cache = make(clock)
|
| 93 |
+
cache["a"] = 1
|
| 94 |
+
clock.advance(4)
|
| 95 |
+
cache["b"] = 2
|
| 96 |
+
clock.advance(7)
|
| 97 |
+
assert list(cache.items()) == [("b", 2)]
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
def test_invalid_arguments():
|
| 101 |
+
with pytest.raises(ValueError):
|
| 102 |
+
TTLCache(maxsize=0)
|
| 103 |
+
with pytest.raises(ValueError):
|
| 104 |
+
TTLCache(ttl=0)
|
| 105 |
+
|
| 106 |
+
|
| 107 |
+
def test_repr_reports_live_count():
|
| 108 |
+
clock = FakeClock()
|
| 109 |
+
cache = make(clock)
|
| 110 |
+
cache["a"] = 1
|
| 111 |
+
assert "live=1" in repr(cache)
|
repos/ttlcache/ttlcache/__init__.py
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""ttlcache - a bounded TTL + LRU cache."""
|
| 2 |
+
|
| 3 |
+
from .cache import TTLCache
|
| 4 |
+
from .clock import FakeClock, SystemClock
|
| 5 |
+
|
| 6 |
+
__version__ = "0.3.0"
|
| 7 |
+
|
| 8 |
+
__all__ = ["FakeClock", "SystemClock", "TTLCache"]
|
repos/ttlcache/ttlcache/cache.py
ADDED
|
@@ -0,0 +1,83 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""A bounded cache with per-entry time-to-live and LRU eviction."""
|
| 2 |
+
|
| 3 |
+
from collections import OrderedDict
|
| 4 |
+
|
| 5 |
+
|
| 6 |
+
class TTLCache:
|
| 7 |
+
"""Fixed-capacity cache where every entry expires ``ttl`` seconds after write.
|
| 8 |
+
|
| 9 |
+
Entries are kept in least-recently-used order. When the cache is full the
|
| 10 |
+
least recently used entry is dropped first. See README.md for the full
|
| 11 |
+
contract.
|
| 12 |
+
"""
|
| 13 |
+
|
| 14 |
+
def __init__(self, maxsize=128, ttl=60.0, clock=None):
|
| 15 |
+
if maxsize <= 0:
|
| 16 |
+
raise ValueError("maxsize must be positive")
|
| 17 |
+
if ttl <= 0:
|
| 18 |
+
raise ValueError("ttl must be positive")
|
| 19 |
+
self.maxsize = maxsize
|
| 20 |
+
self.ttl = float(ttl)
|
| 21 |
+
self._clock = clock
|
| 22 |
+
self._data = OrderedDict()
|
| 23 |
+
|
| 24 |
+
def _now(self):
|
| 25 |
+
return self._clock.now() if self._clock is not None else 0.0
|
| 26 |
+
|
| 27 |
+
@staticmethod
|
| 28 |
+
def _expired(deadline, now):
|
| 29 |
+
return now >= deadline
|
| 30 |
+
|
| 31 |
+
def __len__(self):
|
| 32 |
+
return len(self._data)
|
| 33 |
+
|
| 34 |
+
def __contains__(self, key):
|
| 35 |
+
if key not in self._data:
|
| 36 |
+
return False
|
| 37 |
+
_, deadline = self._data[key]
|
| 38 |
+
return not self._expired(deadline, self._now())
|
| 39 |
+
|
| 40 |
+
def __setitem__(self, key, value):
|
| 41 |
+
now = self._now()
|
| 42 |
+
self._data[key] = (value, now + self.ttl)
|
| 43 |
+
while len(self._data) > self.maxsize:
|
| 44 |
+
self._data.popitem(last=False)
|
| 45 |
+
|
| 46 |
+
def __getitem__(self, key):
|
| 47 |
+
value, deadline = self._data[key]
|
| 48 |
+
if self._expired(deadline, self._now()):
|
| 49 |
+
del self._data[key]
|
| 50 |
+
raise KeyError(key)
|
| 51 |
+
return value
|
| 52 |
+
|
| 53 |
+
def get(self, key, default=None):
|
| 54 |
+
try:
|
| 55 |
+
return self[key]
|
| 56 |
+
except KeyError:
|
| 57 |
+
return default
|
| 58 |
+
|
| 59 |
+
def setdefault(self, key, factory):
|
| 60 |
+
"""Return the live value for ``key``, computing and storing it if absent."""
|
| 61 |
+
if key in self:
|
| 62 |
+
return self[key]
|
| 63 |
+
value = factory()
|
| 64 |
+
self[key] = value
|
| 65 |
+
return value
|
| 66 |
+
|
| 67 |
+
def pop(self, key, default=None):
|
| 68 |
+
if key not in self._data:
|
| 69 |
+
return default
|
| 70 |
+
value, _ = self._data.pop(key)
|
| 71 |
+
return value
|
| 72 |
+
|
| 73 |
+
def items(self):
|
| 74 |
+
"""Iterate over live ``(key, value)`` pairs in LRU order."""
|
| 75 |
+
now = self._now()
|
| 76 |
+
return [(k, v) for k, (v, d) in self._data.items() if not self._expired(d, now)]
|
| 77 |
+
|
| 78 |
+
def __repr__(self):
|
| 79 |
+
return "TTLCache(maxsize=%d, ttl=%r, live=%d)" % (
|
| 80 |
+
self.maxsize,
|
| 81 |
+
self.ttl,
|
| 82 |
+
len(self),
|
| 83 |
+
)
|
repos/ttlcache/ttlcache/clock.py
ADDED
|
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Clocks used by ttlcache.
|
| 2 |
+
|
| 3 |
+
The cache never calls ``time`` directly: it asks a clock object for the current
|
| 4 |
+
time, which makes expiry deterministic in tests.
|
| 5 |
+
"""
|
| 6 |
+
|
| 7 |
+
import time
|
| 8 |
+
|
| 9 |
+
|
| 10 |
+
class SystemClock:
|
| 11 |
+
"""Monotonic wall clock for production use."""
|
| 12 |
+
|
| 13 |
+
def now(self):
|
| 14 |
+
return time.monotonic()
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
class FakeClock:
|
| 18 |
+
"""A clock that only moves when you tell it to."""
|
| 19 |
+
|
| 20 |
+
def __init__(self, start=0.0):
|
| 21 |
+
self._now = float(start)
|
| 22 |
+
|
| 23 |
+
def now(self):
|
| 24 |
+
return self._now
|
| 25 |
+
|
| 26 |
+
def advance(self, seconds):
|
| 27 |
+
self._now += float(seconds)
|
| 28 |
+
return self._now
|
run_harness.py
ADDED
|
@@ -0,0 +1,238 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""SWE-DockerLess evaluation harness.
|
| 2 |
+
|
| 3 |
+
No Docker, no network, no prebuilt images: every instance carries its own repo
|
| 4 |
+
snapshot inside the JSONL row, so the harness only needs Python, git and pytest.
|
| 5 |
+
|
| 6 |
+
Usage
|
| 7 |
+
-----
|
| 8 |
+
# sanity check: score the gold patches (expect 4/4 resolved)
|
| 9 |
+
python run_harness.py --dataset SWE-DockerLess.jsonl --gold
|
| 10 |
+
|
| 11 |
+
# score a predictions file: one JSON object per line with
|
| 12 |
+
# {"instance_id": ..., "model_name_or_path": ..., "model_patch": ...}
|
| 13 |
+
python run_harness.py --dataset SWE-DockerLess.jsonl \
|
| 14 |
+
--predictions predictions.jsonl --out report.json
|
| 15 |
+
|
| 16 |
+
A prediction is "resolved" when every FAIL_TO_PASS test passes and every
|
| 17 |
+
PASS_TO_PASS test still passes, which is the SWE-bench rule.
|
| 18 |
+
"""
|
| 19 |
+
|
| 20 |
+
import argparse
|
| 21 |
+
import json
|
| 22 |
+
import os
|
| 23 |
+
import re
|
| 24 |
+
import shutil
|
| 25 |
+
import subprocess
|
| 26 |
+
import sys
|
| 27 |
+
import tempfile
|
| 28 |
+
import time
|
| 29 |
+
from pathlib import Path
|
| 30 |
+
|
| 31 |
+
GIT_FLAGS = [
|
| 32 |
+
"-c", "core.autocrlf=false",
|
| 33 |
+
"-c", "core.safecrlf=false",
|
| 34 |
+
"-c", "user.name=harness",
|
| 35 |
+
"-c", "user.email=harness@example.invalid",
|
| 36 |
+
"-c", "commit.gpgsign=false",
|
| 37 |
+
]
|
| 38 |
+
|
| 39 |
+
RESULT_RE = re.compile(
|
| 40 |
+
r"^(tests/[^\s:]+::[^\s]+)\s+(PASSED|FAILED|ERROR|SKIPPED|XFAIL|XPASS)"
|
| 41 |
+
)
|
| 42 |
+
|
| 43 |
+
|
| 44 |
+
def run(cmd, cwd, check=True):
|
| 45 |
+
env = os.environ.copy()
|
| 46 |
+
env.setdefault("GIT_AUTHOR_NAME", "harness")
|
| 47 |
+
env.setdefault("GIT_AUTHOR_EMAIL", "harness@example.invalid")
|
| 48 |
+
env.setdefault("GIT_COMMITTER_NAME", "harness")
|
| 49 |
+
env.setdefault("GIT_COMMITTER_EMAIL", "harness@example.invalid")
|
| 50 |
+
proc = subprocess.run(
|
| 51 |
+
cmd, cwd=str(cwd), capture_output=True, text=True, encoding="utf-8", env=env
|
| 52 |
+
)
|
| 53 |
+
if check and proc.returncode != 0:
|
| 54 |
+
raise RuntimeError(
|
| 55 |
+
"command failed (%d): %s\n%s\n%s"
|
| 56 |
+
% (proc.returncode, " ".join(cmd), proc.stdout[-4000:], proc.stderr[-4000:])
|
| 57 |
+
)
|
| 58 |
+
return proc
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
def git(args, cwd, check=True):
|
| 62 |
+
return run(["git"] + GIT_FLAGS + args, cwd, check=check)
|
| 63 |
+
|
| 64 |
+
|
| 65 |
+
def write_snapshot(files, dest):
|
| 66 |
+
for rel, content in files.items():
|
| 67 |
+
path = dest / rel
|
| 68 |
+
path.parent.mkdir(parents=True, exist_ok=True)
|
| 69 |
+
with open(path, "w", encoding="utf-8", newline="\n") as handle:
|
| 70 |
+
handle.write(content)
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
def apply_patch(patch_text, work, tmp, label):
|
| 74 |
+
"""Apply a unified diff; return (ok, message)."""
|
| 75 |
+
if not patch_text or not patch_text.strip():
|
| 76 |
+
return False, "empty patch"
|
| 77 |
+
patch_file = Path(tmp) / (label + ".patch")
|
| 78 |
+
patch_file.write_text(patch_text, encoding="utf-8", newline="\n")
|
| 79 |
+
first = git(["apply", "--check", str(patch_file)], work, check=False)
|
| 80 |
+
if first.returncode == 0:
|
| 81 |
+
git(["apply", str(patch_file)], work)
|
| 82 |
+
return True, "applied"
|
| 83 |
+
second = git(["apply", "--check", "--recount", "-C1", str(patch_file)], work, check=False)
|
| 84 |
+
if second.returncode == 0:
|
| 85 |
+
git(["apply", "--recount", "-C1", str(patch_file)], work)
|
| 86 |
+
return True, "applied (recount)"
|
| 87 |
+
return False, (first.stderr or first.stdout).strip().splitlines()[0:3]
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
def run_tests(work, test_files, cmd_template):
|
| 91 |
+
cmd = [sys.executable, "-m", "pytest", "-v", "--tb=no", "-p", "no:cacheprovider"]
|
| 92 |
+
cmd += list(test_files)
|
| 93 |
+
proc = run(cmd, work, check=False)
|
| 94 |
+
outcomes = {}
|
| 95 |
+
for raw in proc.stdout.splitlines():
|
| 96 |
+
match = RESULT_RE.match(raw.strip())
|
| 97 |
+
if match:
|
| 98 |
+
outcomes[match.group(1)] = match.group(2)
|
| 99 |
+
return outcomes, proc
|
| 100 |
+
|
| 101 |
+
|
| 102 |
+
def evaluate(row, model_patch):
|
| 103 |
+
test_files = row["dockerless"]["test_files"]
|
| 104 |
+
f2p = json.loads(row["FAIL_TO_PASS"])
|
| 105 |
+
p2p = json.loads(row["PASS_TO_PASS"])
|
| 106 |
+
|
| 107 |
+
with tempfile.TemporaryDirectory(prefix="swedl-eval-") as tmp:
|
| 108 |
+
work = Path(tmp) / "work"
|
| 109 |
+
work.mkdir()
|
| 110 |
+
write_snapshot(row["dockerless"]["files"], work)
|
| 111 |
+
git(["init", "-q", "-b", "main"], work)
|
| 112 |
+
git(["add", "-A"], work)
|
| 113 |
+
git(["commit", "-q", "-m", row["base_commit"]], work)
|
| 114 |
+
|
| 115 |
+
ok, message = apply_patch(row["test_patch"], work, tmp, "test")
|
| 116 |
+
if not ok:
|
| 117 |
+
return {
|
| 118 |
+
"instance_id": row["instance_id"],
|
| 119 |
+
"resolved": False,
|
| 120 |
+
"reason": "test_patch failed to apply: %s" % (message,),
|
| 121 |
+
"f2p_passed": 0,
|
| 122 |
+
"f2p_total": len(f2p),
|
| 123 |
+
"p2p_passed": 0,
|
| 124 |
+
"p2p_total": len(p2p),
|
| 125 |
+
}
|
| 126 |
+
|
| 127 |
+
ok, message = apply_patch(model_patch, work, tmp, "model")
|
| 128 |
+
if not ok:
|
| 129 |
+
return {
|
| 130 |
+
"instance_id": row["instance_id"],
|
| 131 |
+
"resolved": False,
|
| 132 |
+
"reason": "model patch failed to apply: %s" % (message,),
|
| 133 |
+
"f2p_passed": 0,
|
| 134 |
+
"f2p_total": len(f2p),
|
| 135 |
+
"p2p_passed": 0,
|
| 136 |
+
"p2p_total": len(p2p),
|
| 137 |
+
}
|
| 138 |
+
|
| 139 |
+
outcomes, proc = run_tests(work, test_files, row["dockerless"]["test_cmd"])
|
| 140 |
+
|
| 141 |
+
if not outcomes:
|
| 142 |
+
return {
|
| 143 |
+
"instance_id": row["instance_id"],
|
| 144 |
+
"resolved": False,
|
| 145 |
+
"reason": "no tests collected (collection error?)",
|
| 146 |
+
"f2p_passed": 0,
|
| 147 |
+
"f2p_total": len(f2p),
|
| 148 |
+
"p2p_passed": 0,
|
| 149 |
+
"p2p_total": len(p2p),
|
| 150 |
+
}
|
| 151 |
+
|
| 152 |
+
f2p_pass = [t for t in f2p if outcomes.get(t) == "PASSED"]
|
| 153 |
+
p2p_pass = [t for t in p2p if outcomes.get(t) == "PASSED"]
|
| 154 |
+
resolved = len(f2p_pass) == len(f2p) and len(p2p_pass) == len(p2p)
|
| 155 |
+
failing = sorted(t for t in f2p + p2p if outcomes.get(t) != "PASSED")
|
| 156 |
+
return {
|
| 157 |
+
"instance_id": row["instance_id"],
|
| 158 |
+
"resolved": resolved,
|
| 159 |
+
"reason": "ok" if resolved else "unmet tests: %d" % (len(failing),),
|
| 160 |
+
"f2p_passed": len(f2p_pass),
|
| 161 |
+
"f2p_total": len(f2p),
|
| 162 |
+
"p2p_passed": len(p2p_pass),
|
| 163 |
+
"p2p_total": len(p2p),
|
| 164 |
+
"failing": failing[:12],
|
| 165 |
+
}
|
| 166 |
+
|
| 167 |
+
|
| 168 |
+
def load_predictions(path):
|
| 169 |
+
predictions = {}
|
| 170 |
+
with open(path, encoding="utf-8") as handle:
|
| 171 |
+
for line in handle:
|
| 172 |
+
line = line.strip()
|
| 173 |
+
if not line:
|
| 174 |
+
continue
|
| 175 |
+
record = json.loads(line)
|
| 176 |
+
predictions[record["instance_id"]] = (
|
| 177 |
+
record.get("model_patch") or record.get("patch") or ""
|
| 178 |
+
)
|
| 179 |
+
return predictions
|
| 180 |
+
|
| 181 |
+
|
| 182 |
+
def main():
|
| 183 |
+
parser = argparse.ArgumentParser(description=__doc__)
|
| 184 |
+
parser.add_argument("--dataset", default="SWE-DockerLess.jsonl")
|
| 185 |
+
parser.add_argument("--predictions", help="JSONL of {instance_id, model_patch}")
|
| 186 |
+
parser.add_argument("--gold", action="store_true", help="score the gold patches")
|
| 187 |
+
parser.add_argument("--out", help="write a JSON report here")
|
| 188 |
+
args = parser.parse_args()
|
| 189 |
+
|
| 190 |
+
rows = [json.loads(line) for line in open(args.dataset, encoding="utf-8") if line.strip()]
|
| 191 |
+
|
| 192 |
+
if args.gold:
|
| 193 |
+
predictions = {row["instance_id"]: row["patch"] for row in rows}
|
| 194 |
+
elif args.predictions:
|
| 195 |
+
predictions = load_predictions(args.predictions)
|
| 196 |
+
else:
|
| 197 |
+
parser.error("pass --gold or --predictions")
|
| 198 |
+
|
| 199 |
+
started = time.time()
|
| 200 |
+
results = []
|
| 201 |
+
print("%-22s %-9s %-7s %-7s %s" % ("instance", "resolved", "F2P", "P2P", "note"))
|
| 202 |
+
for row in rows:
|
| 203 |
+
result = evaluate(row, predictions.get(row["instance_id"], ""))
|
| 204 |
+
results.append(result)
|
| 205 |
+
print(
|
| 206 |
+
"%-22s %-9s %d/%-5d %d/%-5d %s"
|
| 207 |
+
% (
|
| 208 |
+
result["instance_id"],
|
| 209 |
+
"yes" if result["resolved"] else "no",
|
| 210 |
+
result["f2p_passed"],
|
| 211 |
+
result["f2p_total"],
|
| 212 |
+
result["p2p_passed"],
|
| 213 |
+
result["p2p_total"],
|
| 214 |
+
result["reason"],
|
| 215 |
+
)
|
| 216 |
+
)
|
| 217 |
+
|
| 218 |
+
resolved = sum(1 for result in results if result["resolved"])
|
| 219 |
+
elapsed = time.time() - started
|
| 220 |
+
summary = {
|
| 221 |
+
"dataset": str(Path(args.dataset).name),
|
| 222 |
+
"instances": len(rows),
|
| 223 |
+
"resolved": resolved,
|
| 224 |
+
"resolve_rate": round(resolved / len(rows), 4) if rows else 0.0,
|
| 225 |
+
"seconds": round(elapsed, 2),
|
| 226 |
+
"results": results,
|
| 227 |
+
}
|
| 228 |
+
print(
|
| 229 |
+
"\nresolved %d/%d (%.1f%%) in %.1fs"
|
| 230 |
+
% (resolved, len(rows), 100.0 * summary["resolve_rate"], elapsed)
|
| 231 |
+
)
|
| 232 |
+
if args.out:
|
| 233 |
+
Path(args.out).write_text(json.dumps(summary, indent=2), encoding="utf-8")
|
| 234 |
+
print("report -> %s" % args.out)
|
| 235 |
+
|
| 236 |
+
|
| 237 |
+
if __name__ == "__main__":
|
| 238 |
+
main()
|
tools/make_controls.py
ADDED
|
@@ -0,0 +1,168 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Generate negative-control prediction files for SWE-DockerLess.
|
| 2 |
+
|
| 3 |
+
Three controls, each of which must score 0/4:
|
| 4 |
+
|
| 5 |
+
empty.jsonl every prediction is an empty patch
|
| 6 |
+
cross.jsonl each instance receives another instance's gold patch
|
| 7 |
+
partial.jsonl a plausible but incomplete fix: the gold patch with one of the
|
| 8 |
+
bugs reintroduced (or, for semverlite, only one of its two
|
| 9 |
+
files fixed)
|
| 10 |
+
|
| 11 |
+
Run with: python tools/make_controls.py
|
| 12 |
+
"""
|
| 13 |
+
|
| 14 |
+
import json
|
| 15 |
+
import shutil
|
| 16 |
+
import subprocess
|
| 17 |
+
import sys
|
| 18 |
+
import tempfile
|
| 19 |
+
from pathlib import Path
|
| 20 |
+
|
| 21 |
+
ROOT = Path(__file__).resolve().parent.parent
|
| 22 |
+
REPOS = ROOT / "repos"
|
| 23 |
+
FIXED = ROOT / "fixed"
|
| 24 |
+
OUTDIR = ROOT / "predictions"
|
| 25 |
+
DATASET = ROOT / "SWE-DockerLess.jsonl"
|
| 26 |
+
|
| 27 |
+
GIT_FLAGS = [
|
| 28 |
+
"-c", "core.autocrlf=false",
|
| 29 |
+
"-c", "core.safecrlf=false",
|
| 30 |
+
"-c", "user.name=control",
|
| 31 |
+
"-c", "user.email=control@example.invalid",
|
| 32 |
+
"-c", "commit.gpgsign=false",
|
| 33 |
+
]
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
def git(args, cwd):
|
| 37 |
+
proc = subprocess.run(
|
| 38 |
+
["git"] + GIT_FLAGS + args, cwd=str(cwd), capture_output=True, text=True, encoding="utf-8"
|
| 39 |
+
)
|
| 40 |
+
if proc.returncode != 0:
|
| 41 |
+
raise RuntimeError("git %s failed: %s %s" % (args, proc.stdout, proc.stderr))
|
| 42 |
+
return proc
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
def read_tree(root):
|
| 46 |
+
files = {}
|
| 47 |
+
for path in sorted(Path(root).rglob("*")):
|
| 48 |
+
if path.is_dir() or ".git" in path.parts:
|
| 49 |
+
continue
|
| 50 |
+
files[path.relative_to(root).as_posix()] = path.read_text(encoding="utf-8")
|
| 51 |
+
return files
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
def write_tree(files, dest):
|
| 55 |
+
for rel, content in files.items():
|
| 56 |
+
path = Path(dest) / rel
|
| 57 |
+
path.parent.mkdir(parents=True, exist_ok=True)
|
| 58 |
+
with open(path, "w", encoding="utf-8", newline="\n") as handle:
|
| 59 |
+
handle.write(content)
|
| 60 |
+
|
| 61 |
+
|
| 62 |
+
def diff_with_overrides(name, overrides):
|
| 63 |
+
"""Return a unified diff that turns the base repo into base+overrides."""
|
| 64 |
+
with tempfile.TemporaryDirectory(prefix="swedl-ctl-") as tmp:
|
| 65 |
+
work = Path(tmp) / "w"
|
| 66 |
+
work.mkdir()
|
| 67 |
+
write_tree(read_tree(REPOS / name), work)
|
| 68 |
+
git(["init", "-q", "-b", "main"], work)
|
| 69 |
+
git(["add", "-A"], work)
|
| 70 |
+
git(["commit", "-q", "-m", "base"], work)
|
| 71 |
+
write_tree(overrides, work)
|
| 72 |
+
return git(["diff", "--no-color"], work).stdout
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
def main():
|
| 76 |
+
OUTDIR.mkdir(exist_ok=True)
|
| 77 |
+
rows = [json.loads(line) for line in open(DATASET, encoding="utf-8") if line.strip()]
|
| 78 |
+
ids = [row["instance_id"] for row in rows]
|
| 79 |
+
|
| 80 |
+
# --- control 1: empty patches ----------------------------------------
|
| 81 |
+
with open(OUTDIR / "empty.jsonl", "w", encoding="utf-8", newline="\n") as handle:
|
| 82 |
+
for row in rows:
|
| 83 |
+
handle.write(
|
| 84 |
+
json.dumps(
|
| 85 |
+
{
|
| 86 |
+
"instance_id": row["instance_id"],
|
| 87 |
+
"model_name_or_path": "control-empty",
|
| 88 |
+
"model_patch": "",
|
| 89 |
+
}
|
| 90 |
+
)
|
| 91 |
+
+ "\n"
|
| 92 |
+
)
|
| 93 |
+
|
| 94 |
+
# --- control 2: cross-instance gold patches ---------------------------
|
| 95 |
+
with open(OUTDIR / "cross.jsonl", "w", encoding="utf-8", newline="\n") as handle:
|
| 96 |
+
for index, row in enumerate(rows):
|
| 97 |
+
other = rows[(index + 1) % len(rows)]
|
| 98 |
+
handle.write(
|
| 99 |
+
json.dumps(
|
| 100 |
+
{
|
| 101 |
+
"instance_id": row["instance_id"],
|
| 102 |
+
"model_name_or_path": "control-cross",
|
| 103 |
+
"model_patch": other["patch"],
|
| 104 |
+
}
|
| 105 |
+
)
|
| 106 |
+
+ "\n"
|
| 107 |
+
)
|
| 108 |
+
|
| 109 |
+
# --- control 3: plausible but incomplete fixes ------------------------
|
| 110 |
+
partial = {}
|
| 111 |
+
|
| 112 |
+
# miniconf: cycle detection kept, but the $$ escape dropped again
|
| 113 |
+
config = read_tree(FIXED / "miniconf")["miniconf/config.py"]
|
| 114 |
+
broken_escape = config.replace(
|
| 115 |
+
' if text.startswith("$$", index):\n'
|
| 116 |
+
' out.append("$")\n'
|
| 117 |
+
" index += 2\n"
|
| 118 |
+
" continue\n",
|
| 119 |
+
"",
|
| 120 |
+
)
|
| 121 |
+
assert broken_escape != config, "miniconf control replacement did not apply"
|
| 122 |
+
partial["swedl__miniconf-1"] = diff_with_overrides(
|
| 123 |
+
"miniconf", {"miniconf/config.py": broken_escape}
|
| 124 |
+
)
|
| 125 |
+
|
| 126 |
+
# ttlcache: len() fixed, but reads still do not refresh recency
|
| 127 |
+
cache = read_tree(FIXED / "ttlcache")["ttlcache/cache.py"]
|
| 128 |
+
broken_recency = cache.replace(" self._data.move_to_end(key)\n", "")
|
| 129 |
+
assert broken_recency != cache, "ttlcache control replacement did not apply"
|
| 130 |
+
partial["swedl__ttlcache-1"] = diff_with_overrides(
|
| 131 |
+
"ttlcache", {"ttlcache/cache.py": broken_recency}
|
| 132 |
+
)
|
| 133 |
+
|
| 134 |
+
# cronlite: Sunday alias fixed, but `<value>/<step>` still collapses
|
| 135 |
+
fields = read_tree(FIXED / "cronlite")["cronlite/fields.py"]
|
| 136 |
+
broken_step = fields.replace(
|
| 137 |
+
" start = self._check(int(base))\n stop = self.maximum\n",
|
| 138 |
+
" start = self._check(int(base))\n stop = start\n",
|
| 139 |
+
)
|
| 140 |
+
assert broken_step != fields, "cronlite control replacement did not apply"
|
| 141 |
+
partial["swedl__cronlite-1"] = diff_with_overrides(
|
| 142 |
+
"cronlite", {"cronlite/fields.py": broken_step}
|
| 143 |
+
)
|
| 144 |
+
|
| 145 |
+
# semverlite: precedence fixed, range gating left out
|
| 146 |
+
version = read_tree(FIXED / "semverlite")["semverlite/version.py"]
|
| 147 |
+
partial["swedl__semverlite-1"] = diff_with_overrides(
|
| 148 |
+
"semverlite", {"semverlite/version.py": version}
|
| 149 |
+
)
|
| 150 |
+
|
| 151 |
+
with open(OUTDIR / "partial.jsonl", "w", encoding="utf-8", newline="\n") as handle:
|
| 152 |
+
for instance_id in ids:
|
| 153 |
+
handle.write(
|
| 154 |
+
json.dumps(
|
| 155 |
+
{
|
| 156 |
+
"instance_id": instance_id,
|
| 157 |
+
"model_name_or_path": "control-partial",
|
| 158 |
+
"model_patch": partial[instance_id],
|
| 159 |
+
}
|
| 160 |
+
)
|
| 161 |
+
+ "\n"
|
| 162 |
+
)
|
| 163 |
+
|
| 164 |
+
print("wrote controls for %d instances -> %s" % (len(ids), OUTDIR))
|
| 165 |
+
|
| 166 |
+
|
| 167 |
+
if __name__ == "__main__":
|
| 168 |
+
main()
|
tools/validate.py
ADDED
|
@@ -0,0 +1,174 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
"""Structural validation for SWE-DockerLess.jsonl.
|
| 2 |
+
|
| 3 |
+
Checks that every row is SWE-bench shaped, that the two patches apply to the
|
| 4 |
+
embedded snapshot, that they touch disjoint files, and that the gold patch stays
|
| 5 |
+
inside a budget a 2048-token completion can produce.
|
| 6 |
+
|
| 7 |
+
Run with: python tools/validate.py
|
| 8 |
+
"""
|
| 9 |
+
|
| 10 |
+
import json
|
| 11 |
+
import subprocess
|
| 12 |
+
import sys
|
| 13 |
+
import tempfile
|
| 14 |
+
from pathlib import Path
|
| 15 |
+
|
| 16 |
+
ROOT = Path(__file__).resolve().parent.parent
|
| 17 |
+
DATASET = ROOT / "SWE-DockerLess.jsonl"
|
| 18 |
+
|
| 19 |
+
SWE_BENCH_FIELDS = [
|
| 20 |
+
"instance_id",
|
| 21 |
+
"repo",
|
| 22 |
+
"base_commit",
|
| 23 |
+
"patch",
|
| 24 |
+
"test_patch",
|
| 25 |
+
"problem_statement",
|
| 26 |
+
"hints_text",
|
| 27 |
+
"created_at",
|
| 28 |
+
"version",
|
| 29 |
+
"environment_setup_commit",
|
| 30 |
+
"FAIL_TO_PASS",
|
| 31 |
+
"PASS_TO_PASS",
|
| 32 |
+
]
|
| 33 |
+
|
| 34 |
+
MAX_PROBLEM_CHARS = 4000
|
| 35 |
+
MAX_GOLD_ADDED_LINES = 120
|
| 36 |
+
|
| 37 |
+
failures = []
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
def check(condition, message):
|
| 41 |
+
if not condition:
|
| 42 |
+
failures.append(message)
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
def write_tree(files, dest):
|
| 46 |
+
for rel, content in files.items():
|
| 47 |
+
path = dest / rel
|
| 48 |
+
path.parent.mkdir(parents=True, exist_ok=True)
|
| 49 |
+
with open(path, "w", encoding="utf-8", newline="\n") as handle:
|
| 50 |
+
handle.write(content)
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
def git(args, cwd):
|
| 54 |
+
return subprocess.run(
|
| 55 |
+
[
|
| 56 |
+
"git",
|
| 57 |
+
"-c", "core.autocrlf=false",
|
| 58 |
+
"-c", "user.name=v",
|
| 59 |
+
"-c", "user.email=v@example.invalid",
|
| 60 |
+
]
|
| 61 |
+
+ args,
|
| 62 |
+
cwd=str(cwd),
|
| 63 |
+
capture_output=True,
|
| 64 |
+
text=True,
|
| 65 |
+
encoding="utf-8",
|
| 66 |
+
)
|
| 67 |
+
|
| 68 |
+
|
| 69 |
+
def touched_paths(patch):
|
| 70 |
+
paths = set()
|
| 71 |
+
for line in patch.splitlines():
|
| 72 |
+
if line.startswith("+++ b/") or line.startswith("--- a/"):
|
| 73 |
+
paths.add(line[6:])
|
| 74 |
+
elif line.startswith("+++ ") or line.startswith("--- "):
|
| 75 |
+
value = line[4:].strip()
|
| 76 |
+
if value != "/dev/null":
|
| 77 |
+
paths.add(value.split("/", 1)[-1])
|
| 78 |
+
return paths
|
| 79 |
+
|
| 80 |
+
|
| 81 |
+
def added_lines(patch):
|
| 82 |
+
return sum(1 for line in patch.splitlines() if line.startswith("+") and not line.startswith("+++"))
|
| 83 |
+
|
| 84 |
+
|
| 85 |
+
def main():
|
| 86 |
+
rows = [json.loads(line) for line in open(DATASET, encoding="utf-8") if line.strip()]
|
| 87 |
+
check(len(rows) == 4, "expected 4 rows, found %d" % len(rows))
|
| 88 |
+
|
| 89 |
+
seen = set()
|
| 90 |
+
for row in rows:
|
| 91 |
+
iid = row.get("instance_id", "<missing>")
|
| 92 |
+
for field in SWE_BENCH_FIELDS:
|
| 93 |
+
check(field in row, "%s: missing field %s" % (iid, field))
|
| 94 |
+
check(iid not in seen, "%s: duplicate instance_id" % iid)
|
| 95 |
+
seen.add(iid)
|
| 96 |
+
|
| 97 |
+
check(isinstance(row["FAIL_TO_PASS"], str), "%s: FAIL_TO_PASS must be a JSON string" % iid)
|
| 98 |
+
check(isinstance(row["PASS_TO_PASS"], str), "%s: PASS_TO_PASS must be a JSON string" % iid)
|
| 99 |
+
f2p = json.loads(row["FAIL_TO_PASS"])
|
| 100 |
+
p2p = json.loads(row["PASS_TO_PASS"])
|
| 101 |
+
check(len(f2p) >= 1, "%s: empty FAIL_TO_PASS" % iid)
|
| 102 |
+
check(len(p2p) >= 1, "%s: empty PASS_TO_PASS" % iid)
|
| 103 |
+
check(not (set(f2p) & set(p2p)), "%s: F2P/P2P overlap" % iid)
|
| 104 |
+
check(len(row["base_commit"]) == 40, "%s: base_commit is not a 40-char sha" % iid)
|
| 105 |
+
|
| 106 |
+
problem = row["problem_statement"]
|
| 107 |
+
check(len(problem) <= MAX_PROBLEM_CHARS, "%s: problem statement too long" % iid)
|
| 108 |
+
for test_id in f2p + p2p:
|
| 109 |
+
check("::" not in problem, "%s: problem statement leaks a test id" % iid)
|
| 110 |
+
|
| 111 |
+
gold_paths = touched_paths(row["patch"])
|
| 112 |
+
test_paths = touched_paths(row["test_patch"])
|
| 113 |
+
check(not (gold_paths & test_paths), "%s: patch and test_patch touch the same files" % iid)
|
| 114 |
+
check(
|
| 115 |
+
all(not path.startswith("tests/") for path in gold_paths),
|
| 116 |
+
"%s: gold patch edits test files" % iid,
|
| 117 |
+
)
|
| 118 |
+
check(
|
| 119 |
+
all(path.startswith("tests/") for path in test_paths),
|
| 120 |
+
"%s: test_patch edits non-test files" % iid,
|
| 121 |
+
)
|
| 122 |
+
check(
|
| 123 |
+
added_lines(row["patch"]) <= MAX_GOLD_ADDED_LINES,
|
| 124 |
+
"%s: gold patch adds %d lines (budget %d)"
|
| 125 |
+
% (iid, added_lines(row["patch"]), MAX_GOLD_ADDED_LINES),
|
| 126 |
+
)
|
| 127 |
+
|
| 128 |
+
files = row["dockerless"]["files"]
|
| 129 |
+
check(bool(files), "%s: empty snapshot" % iid)
|
| 130 |
+
|
| 131 |
+
with tempfile.TemporaryDirectory(prefix="swedl-val-") as tmp:
|
| 132 |
+
work = Path(tmp) / "w"
|
| 133 |
+
work.mkdir()
|
| 134 |
+
write_tree(files, work)
|
| 135 |
+
git(["init", "-q", "-b", "main"], work)
|
| 136 |
+
git(["add", "-A"], work)
|
| 137 |
+
git(["commit", "-q", "-m", "base"], work)
|
| 138 |
+
|
| 139 |
+
for label, patch_text in (("test_patch", row["test_patch"]), ("patch", row["patch"])):
|
| 140 |
+
patch_file = Path(tmp) / (label + ".patch")
|
| 141 |
+
patch_file.write_text(patch_text, encoding="utf-8", newline="\n")
|
| 142 |
+
result = git(["apply", "--check", str(patch_file)], work)
|
| 143 |
+
check(
|
| 144 |
+
result.returncode == 0,
|
| 145 |
+
"%s: %s does not apply (%s)" % (iid, label, result.stderr.strip()[:200]),
|
| 146 |
+
)
|
| 147 |
+
if label == "test_patch":
|
| 148 |
+
git(["apply", str(patch_file)], work)
|
| 149 |
+
|
| 150 |
+
print(
|
| 151 |
+
"%-22s fields=ok F2P=%-2d P2P=%-2d +%-3d lines problem=%d chars files=%d"
|
| 152 |
+
% (
|
| 153 |
+
iid,
|
| 154 |
+
len(f2p),
|
| 155 |
+
len(p2p),
|
| 156 |
+
added_lines(row["patch"]),
|
| 157 |
+
len(problem),
|
| 158 |
+
len(files),
|
| 159 |
+
)
|
| 160 |
+
)
|
| 161 |
+
|
| 162 |
+
total_tests = sum(len(json.loads(row["FAIL_TO_PASS"])) for row in rows)
|
| 163 |
+
print("\n%d instances, %d FAIL_TO_PASS tests total" % (len(rows), total_tests))
|
| 164 |
+
if failures:
|
| 165 |
+
print("\nFAILED %d checks:" % len(failures))
|
| 166 |
+
for message in failures:
|
| 167 |
+
print(" - %s" % message)
|
| 168 |
+
return 1
|
| 169 |
+
print("all checks passed")
|
| 170 |
+
return 0
|
| 171 |
+
|
| 172 |
+
|
| 173 |
+
if __name__ == "__main__":
|
| 174 |
+
sys.exit(main())
|