Sample configs shipped with newer releases can reference media types or
options unknown to older versions; previously the loader called
sys.exit(3) and refused to boot. Now drop unknown scan.media entries,
fall back to defaults for unknown gamelist media thumbnail/image values,
and recover from YAML parse errors with a clear critical log instead of
a traceback. Type-mismatch validation still hard-fails since those are
real misconfigurations.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
UPDATE scans with a known launchbox_id were calling get_rom_by_id(),
which only returns remote data and bypassed get_rom()'s local-first
merge — so local-only fields like Notes were getting clobbered and
local media matching fidelity dropped. get_rom_by_id() now optionally
takes fs_name/platform_slug and merges the local entry when its
DatabaseID matches the requested id.
Also fixes populate_rom_specific_paths writing every video as
video.mp4 regardless of source extension; since store_media_file is a
byte copy, .mkv/.webm contents would be served as .mp4 and break MIME
sniffing.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
magic.Magic(mime=True) loads the magic database from disk on construction;
instantiating it per request was adding pointless overhead to every avatar
and artwork upload. Share a module-level instance guarded by a lock (the
underlying magic_t handle is not thread-safe), and surface MagicException
as a 400 so a sniffing failure fails closed instead of bubbling a 500.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds rejection + acceptance tests for update_rom, add_collection, and
update_collection artwork uploads, mirroring the existing avatar tests:
non-image content returns 400, and a real PNG uploaded under a misleading
filename like payload.html is stored with the trusted .png extension.
Also fixes two `return HTTPException(...)` → `raise` in raw.py so the 404
path actually surfaces instead of silently returning the exception object.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Avatar, ROM artwork, and collection artwork uploads now sniff the file
header with libmagic and reject anything that isn't PNG/JPEG/WebP/GIF,
saving the file with an extension derived from the detected MIME rather
than the user-supplied filename. Pairs with the raw asset endpoint,
which decides inline vs attachment from the on-disk extension.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Switch the get_platform_fs_structure, get_firmware_fs_structure,
get_platforms_directory and get_roms_fs_structure tests from mocking
os.path.exists to mocking glob.glob, matching the new STRUCTURE_PATH_B
detection logic. Rename the existing high_priority/normal_structure
variants to structure_a/structure_b for clarity, and fix
test_platform_specific_behavior so its monkeypatch wraps the calls
instead of being applied after them.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Replace the single HIGH_PRIO_STRUCTURE_PATH config attribute with two
glob patterns (STRUCTURE_PATH_A = roms/*, STRUCTURE_PATH_B = */roms) and
update all call sites to detect Structure B via glob.glob, defaulting to
Structure A when no match is found.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Renaming a rom via PUT /roms/{id} only updated fs_name and its
derivatives, leaving regions/languages/tags/revision/version stale
against the new filename. Re-parse them whenever fs_name changes so
edits like "patapon (Fr En)" -> "Patapon (Fr, En)" are reflected in
the database.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The sample_metadata.zip test fixture used invented tag names
(<Description>, FileName ending in .zip). Real LaunchBox Mame.xml
(see backend/tasks/fixtures/launchbox/mame.xml) uses <Name> as the
full title and keys MameFile entries by the stem (e.g. wrlok_l3),
no extension. Read Name instead of Description, and fall back to
the stem when the raw fs_name lookup misses.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
remote_media_req built a MediaRequest with platform_name=None, which
made _build_local_media_context bail and never search on-disk
Images/Manuals/Videos. Any rom that matched only via Metadata.xml (no
local XML entry) ended up with images: [] even though LaunchBox art
was sitting on disk. Thread platform_name and fs_name into
remote_media_req, falling back to the remote entry's Platform field
when no slug is available (e.g. lookups by database id).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
LaunchBox stores videos under /romm/launchbox/Videos/<Platform>/<Game>.<ext>
but the handler only extracted YouTube IDs; LAUNCHBOX_VIDEOS_DIR was
defined and unused. Add _get_video() to surface local videos as
launchbox-file:// URLs, thread a populate_rom_specific_paths() through
the scan pipeline to set video_path once the rom is known, and mirror
the SS/gamelist branch in the scan socket so store_media_file actually
copies the video into the rom's resource directory. Rom.path_video now
also reads from launchbox_metadata.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The scheduled task populated LAUNCHBOX_MAME_KEY from Mame.xml but no
handler code ever read it, so arcade ROMs like pacman.zip never matched
against Metadata.xml (which keys on full titles like "Pac-Man"). Add
RemoteSource.get_mame_entry() and an ARCADE branch in get_rom() that
resolves the MAME filename to its Description before name lookup.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
LaunchBox produced file:// URIs relative to /romm/launchbox, but the
resources handler resolved them under /romm/library via fs_rom_handler,
so local images/manuals/screenshots were never found. Switch LaunchBox
to a distinct launchbox-file:// scheme and add FSLaunchboxHandler +
_resolve_local_file_uri to route each scheme to the correct root.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
