DJ Lab v1.0.9 manifest row (archive 15cc4844)

BACKUP_MANIFEST.md

@@ -190,9 +190,10 @@ The archives are live snapshots of 2026-11-02 taken while DJ Lab Campaign Wave E
 | File | Size | SHA-256 | HF commit | What it is |
 |---|---|---|---|---|
 | `djlab-app-v1.0.8-final.exe` | 23,396,864 | `a0326fc686b32d6b744d23a417df11f321925a61998950cbf566cf509cde6025` | `20a2e3ab894c5b9745589408ea01aed93b71542e` | DJ Lab **v1.0.8** — the AI wave (report: proof/AI_WAVE_108_REPORT.md): Item 0 Shazam peak-encoding reconcile (byte-identical to the user's own SongRec port on synthetic fixtures; live interoperability NOT claimed — the user's machine is the first live proof) · the six AI programs (Auto-Cue Prepper · Duplicate & Lost-Track Rescue · Clean-Up Crew · Mixable Next + Set Builder · Forgotten Gems Digger · Client List Matcher) on the AI hub + the track-menu AI group · real-run tranches D56–D78: D57 native HTTP transport for every lane (typed errors), D56 same-chart-or-named-failure, D58 cell render atomicity, D59/D60 year provenance + field-write governance, D61 direction corroboration, D62 transliteration gate fold (Asala outcome change stated in the report for the user's verdict), D63 audio-lane trigger + always-printed lane line, D64–D66 deck triggers/Open Folder/multi-resolution waveform + grid editing, D67 Utility grouping, D68 full-index hunter, D69 structured relocate, D70 input normalisation, D71 scrollbars, D72 header measurements, D73 genre aggregation, D74 Camelot default key format, D75 write-visibility honesty, **D76 the DIRECT-WRITE LAW** (the Sync screen/queue erased; every mutation lands in master.db in the gesture; legacy queue drained once with named outcomes), D77 importer mapping audit + duplicate STATUS chips + rail counts, D78 the 900-problem sweep (numbered found-and-fixed list in the report, incl. two disposition-round product fixes: duration-alone-is-not-evidence in the hunter + Auto Replace, and the deck BPM tag following a BPM-stepped grid). Release profile, version 1.0.8 in all five places + Release Notes head, **static CRT preserved** — PE census proof/pe_census_108.txt = **18** Windows system DLLs, **set-diff vs v1.0.7 EMPTY** (no additions, no removals; census taken by direct PE import-directory parse — llvm-objdump segfaults and llvm-readobj is 0 bytes in the post-VM-replacement sandbox, disclosed in the census file). Core battery **166/166, 0 failed** on the final source (log proof/battery_108b.log). GUI gate: every phase green on the final source EXCEPT Z4 (1/3) and Z5 (2/2), dispositioned environment-bound with probe evidence in the report §12 (the sandbox egress proxy's TLS interception breaks rustls for itunes.apple.com and api.deezer.com — InvalidCertificate(UnknownIssuer) through the app's exact native transport; MusicBrainz answers 200; the cascade behaved per D60 with the single answering lane; the stubbed phases AQ/AC/AJ carry the cascade proofs). Gate-environment disclosures (write-guard window truncating three fixtures, all regenerated + affected phases re-proven; two further VM replacements) are in report §12 notes c–d. Served bytes verified after upload (X-Linked-Size 23,396,864 and X-Linked-ETag = the local sha256). The repo's `djlab-app.exe` is NOT touched by this upload — the parent verifies and replaces it. |
+| `djlab-app-v1.0.9-final.exe` | 23,639,040 | `98cf5c7b21b179fc9fee15e4a8df729b96766ea5b5239e63dd80bfbfdf27fcdf` | `15cc4844aa483f1f02aa0c613173b898dab4bb0b` | DJ Lab **v1.0.9** — the real-run defect wave D79–D98 (report: proof/DEFECT_WAVE_109_REPORT.md): D79 Apple Music genre chips removed where no keyless same-chart lane exists (iTunes keeps its own) · D80 Charts preview probe-gated + WAV rewrap, no engine panics · D81 canonical per-track Open URLs · D82 Beatport browser headers + webview-assisted fetch on 403 (his machine is the live proof) · D83 YouTube lane re-derived (Kworb trending, WW + per-country) + full lane URL census (EU aggregate paths fixed) · D84 nested-roots dedupe · D85 full-index grading truth · **D86 one-press hunt** (finds, analyses, applies; only true ties ask) · D87 per-prefix dead-root remedy (existence-verified remap) · D88 key-format first-run migration to Camelot · D89 independent jobs · D90 Prepper one-press + generated cues land as HOT CUES in free slots 1–8 · **D91 operation-level undo law** · **D92 scope popup picker law** · D93 zombie-job remedy · D94 pad delete pins to the corner · D95 denormalized-field masking fixed (unlinked values never read as already-in-library) · D96 title markers preserved on identify write-back · D97 History completion (selection, bulk undo, reasoned states, Redo) · D98 second sweep: 7 found-and-fixed (incl. the Hunter's dead Cancel button and the Relocate '/' contradiction). Release profile, version 1.0.9 in all five places + Release Notes head, **static CRT preserved** — PE census proof/pe_census_109.txt = **18** Windows system DLLs, **set-diff vs v1.0.8 EMPTY**. Core battery **172 passed / 0 failed** (proof/battery_109_final.log). GUI gate: every phase green on the final source EXCEPT Z4 (0/4) and Z5 (2/2), carried environment dispositions from the v1.0.8 report (sandbox egress TLS interception; not re-cut), and AA 19/21 with SAA14/SAA15 dispositioned in the report (the at-scale cancel press window is not exercisable on the fixture in an unloaded sandbox; the mechanism fix is sweep #7). Served bytes verified after upload (X-Linked-Size 23,639,040 and the LFS object hash = the local sha256). The repo's `djlab-app.exe` is NOT touched by this upload — the parent verifies and replaces it. |
 
 - Content files after this addition: **45**, plus this manifest = 46. Sum of file sizes: **9,987,274,115 bytes**.
 
 **Canonical replacement (parent, 2026-11-05):** `djlab-app.exe` replaced with the v1.0.8 final build — 23,396,864 B, sha256 `a0326fc686b32d6b744d23a417df11f321925a61998950cbf566cf509cde6025`, HF commit `001055da91cd90fef82e9c9d8cf431ca23b3070c`. Parent independently re-verified before replacing: local stat + sha256 recomputed (exact match with the archive), own PE census = exactly the same 18 Windows system DLLs as v1.0.7 (set-diff vs proof/pe_census_107.txt shows only the census file's own header comments — no DLL added or removed; census taken by direct PE import-directory parse, method disclosed in the file), core battery independently re-summed from proof/battery_108b.log = **166 passed / 0 failed**, version 1.0.8 grep-verified in all five places, gate record read in full in proof/AI_WAVE_108_REPORT.md (GUI gate: every phase green on the final source EXCEPT Z4 (1/3) and Z5 (2/2), dispositioned environment-bound with probe evidence — the sandbox egress proxy's TLS interception breaks rustls for iTunes/Deezer; the D59–D63 cascade proofs are carried by the stubbed phases AQ/AC/AJ; the live two-lane proof is the user's machine). Served-bytes proof follows in this file's verification note (canonical downloaded back and re-hashed by the parent).
 
-- Content files after the delivery: **45**, plus this manifest = 46. Sum of file sizes: **9,987,388,803 bytes** (the canonical replacement: −23,282,176 +23,396,864).
+- Content files after the v1.0.9 archive upload: **46**, plus this manifest = 47. Sum of file sizes: **10,011,027,843 bytes** (9,987,388,803 + 23,639,040 for the v1.0.9 archive; the canonical still holds v1.0.8 until the parent replaces it).