The cross-reference loop body became its own function, but one exit
stayed a continue and left the module unparseable, which took the whole
test suite down with it. The other ten continues are inside genuine
inner loops and stand.
Output verified identical to the version before the split.
A declared hash that the local dump contradicts is not one situation. An
existence platform never reads the bytes, so withholding the file lets an
upstream list error remove something the frontend would have loaded; a
hash platform would reject it, so shipping it is pointless.
The mode now decides, at every point that had an opinion: pack building,
core complement, emulator packs, manifests, conformance and
_intentional_hash_exclusion. verify.find_undeclared_files follows, since
verify and generate_pack must agree file for file.
Also here: resolution reports which evidence matched rather than a flat
"exact", a path or filename can no longer override a declared hash, and
safe_extract_zip treats a Windows backslash as the separator it is
instead of refusing the archive.
Single source of truth for gap page: verification status from
verify.py (verified/untested/missing/mismatch), file provenance
from cross_reference (bios/data/large_file/missing).
cross_reference.py: _find_in_repo -> _resolve_source returning
source category, stop skipping storage: release/large_file,
add by_path_suffix lookup, all_declared param for global check.
generate_site.py: gap page now shows verification by platform,
18 hash mismatches, and core complement with provenance breakdown.
Run ruff check --fix: remove unused imports (F401), fix f-strings
without placeholders (F541), remove unused variables (F841), fix
duplicate dict key (F601).
Run isort --profile black: normalize import ordering across all files.
Run ruff format: apply consistent formatting (black-compatible) to
all 58 Python files.
3 intentional E402 remain (imports after require_yaml() must execute
after yaml is available).
Files with storage: release are in GitHub release assets,
not in bios/. Eliminates donpachi/sfz3mix/twotiger false
positives. 149/149 tests pass. Cross-ref: 10 -> 7.
Files with explicit path: null are UI-imported (Dolphin NAND,
Hatari cartridge) and not resolvable by pack placement. Skip
them in find_undeclared_files and cross_reference. Also add
desc.dat (SDLPAL fan-made descriptions) to data/. 149/149 OK.
Many emulator profiles use descriptive names (e.g., "SeaBIOS
(128 KB)") while files exist under their path: field basename
(e.g., "bios.bin"). Try path: when name: fails. Eliminates
206 false positives. True missing: 448 -> 242.
_build_supplemental_index scans both data/ directories and
contents of bios/ ZIP files. Eliminates 197 false positives
where files existed inside archive ZIPs (neogeo.zip, pgm.zip,
stvbios.zip, etc.) but were counted as missing. True missing
drops from 645 to 448.
_find_in_repo and _name_in_index now scan data/ in addition to
bios/ via database.json. Eliminates 129 false positives from
game data migrated to data/ (OpenTyrian, ScummVM, SDLPAL, Cave
Story, Syobon Action). True missing: 782 -> 653.
- fix KeyError in compute_coverage (generate_readme, generate_site)
- fix comma-separated MD5 handling in generate_pack check_inside_zip
- fix _verify_file_hash to handle multi-MD5 for large files
- fix external downloads not tracked in seen_destinations/file_status
- fix tar path traversal in _is_safe_tar_member (refresh_data_dirs)
- fix predictable tmp path in download.py
- fix _sanitize_path to filter "." components
- remove blanket data_dir suppression in find_undeclared_files
- remove blanket data_dir suppression in cross_reference
- add status_counts to verify_platform return value
- add md5_composite cache for repeated ZIP hashing
New files: OpenTyrian data (11), Cave Story (2), SeaBIOS,
VGA BIOS, OpenSBI, Cromwell, xbox_hdd, Sega CD Model 2 (3),
NGP Color BIOS, Pentagon 128p-1.rom, X1 font, BK TERAK.
cross_reference.py: basename + case-insensitive lookup.
Abdessamad Derraz