NanoBoyAdvance opens general.bios_path, default bios.bin (src/platform/core/src/config.cpp:36 at 3bb6f478). gba_bios.bin is the path ROCKNIX's config.toml sets, recorded under launchers: in rocknix.yml.
Without the alias the entry met only PCem's and QEMU's bios.bin, 64 and 128 KB, which BIOSLoader refuses (bios.cpp:17-48). verify counted it held by name, so the launcher layer left it alone and the ROCKNIX pack gained PCem's image. A hashless entry whose every same-named copy fails its size is now not held, and the gba_bios.bin ROCKNIX declares settles it.
archive_members scanned the whole file list on every call, and callers ask once per entry of the same profile: 38 million entries walked in one verify --all over mame.yml. Lists of 64 entries or more are now indexed by archive once (held so the id is not reused, indexed again when the length changes), members still read from the entries at each call. archive_members drops from 2.30 s to 0.14 s tottime under the profiler. verify --all (text and --json), slots --json and the retroarch manifest are byte-identical.
Step 2a advanced mame.yml's core_version and upstream_commit to a new
release and left the device ROM sets on the previous one. A plain run
that moves the pin now fetches that release's listxml into .cache/dats
and runs the device merge, which checks version and commit itself.
The acquisition report never read a profile, so XM7's required DICROM.ROM
sat among 2 287 targets at the rank of dumps no core reads. A missing
entry that a profile reads and the collection lacks is now listed first
with the profile, matched by hash or, for a hashless entry with a size,
by name and size. 29 entries on the current snapshots.
A crc32 and a size can be met by crafted bytes, so a remote download
matched on them could land in bios/. verify_content now requires every
sha256, sha1 and md5 the entry declares to match, and an entry declaring
none is reported as not verifiable from the network and nothing is
written. crc32 and size still find a file the database already holds.
check_crypto_validation keeps its answer keyed by the file and keys file as compute_hashes keys them (path, inode, size, mtime, ctime). verify --all ran the verifiers 35 times for 5 files (0.88 s), now 5 times (0.13 s). verify --all --json output is byte-identical.
Basilisk II opens rom_path or ROM_FILE_NAME "ROM" (BasiliskII/src/Unix/main_unix.cpp:109, 685) and never mac.rom. RetroPie's basilisk.sh:56 passes $romdir/macintosh/mac.rom with --rom, recorded under launchers: in retropie.yml.
load_dd_rom builds only <system>/Mupen64plus/IPL.n64 (mupen64plus-core/src/main/main.c:958-961 at 12edd2c7). 64DD_IPL.bin is Batocera's standalone mupen64plus name (mupenConfig.py:188-192), held in batocera.yml files and launchers. standalone_path described a build this libretro profile does not have.
Fusion.exe fixes only us_scd1_9210.bin, jp_mcd1_9111.bin and eu_mcd1_9210.bin (0x44b662-0x44b74d). bios_CD_U/J/E.bin are the names RetroBat's generator writes into Fusion.ini (KegaFusion.Generator.cs:105-116), and platforms/retrobat.yml declares all three.
The loader at speccy 0x3d970 gzopens only the name it is given and clears the model bit when that fails. ZXS16.ROM is passed with mask 1 (0x3dcb2), ZXS48.ROM separately with mask 2 (0x3dcc8), and the binary holds no 48.rom string (Speccy59-Ubuntu-x86-bin). The aliases let a root 48.rom settle the 16K ROM, which the gap report now lists.
mame2000 opens hiscore.dat, history.dat and mameinfo.dat from IMAMEBASEPATH, the content directory (src/libretro/fileio.c:1002-1014, libretro.c:1044-1052). mame2014 and mame2015 open hiscore.dat through emu_file with no search path, the working directory (src/emu/hiscore.c:330-337, fileio.c:141-143). With no load_from the entries claimed system/ and the packs shipped files neither core opens there. ANoteOutsideTheSystemDirCarriesLoadFrom holds libretro notes and load_from together.
ares finds firmware by sha256 (desktop-ui/settings/firmware.cpp:116-136) and names no file. Its FDS entry (RP2C33-01, crc32 1c7ae5d5) was labelled disksys.rom, the name nestopia, fceumm and four more cores read for RP2C33A-01A: ROCKNIX and EmuDeck got ares's revision at disksys.rom. It now carries MAME's member name rp2c33-01.bin.
The LaserActive v1.05 entry, the PAC-S10, PAC-N10, PAC-N1, PCE-LP1 and Neo Geo Pocket aliases are names RetroBat's launcher passes (Ares.Generator.cs:435-555): mega-ld.cpp:16-17 holds only the v1.04 US and v1.02 Japan hashes. RetroBat declares each name in its own list.
AHashScanLabelNamesNoOtherDump compares every found_by: content label with other profiles' entries of that name through the hash indexes.
Ten profile notes graded a .info (correct, accurate, still, missing, omits). Each now says what the file declares and what the code reads. test_no_grade_on_another_projects_file reads every profile line that names a .info, core-info or System.dat.
verify_content read sha1 then md5 only, so an entry pinned by sha256 or
crc32 could be searched and never accepted, and one without any hash was
announced as searched. The strongest declared hash now decides, crc32
counts with its size, the database lookup reads sha256 and crc32, and an
entry with nothing to prove a download by is reported, not searched.
Equal files have equal sizes, and the scan only uses the sha1. Files alone at their size are no longer read and the rest get one digest instead of five. Same 278 groups (645 paths), same dry-run output, 36.7 s to 1.9 s.
check_freshness exited 0 when no upstream answered. An unknown row now
holds the exit code, and a profile skipped by nature is reported skipped
so it does not hold it forever. --only names the valid areas, a stale
native row names its platform, and the summary names refresh_stale.py.
The summary in exports/<platform>.json read '82 hashes corrected' and '854 additions left out'. It now counts what the written file carries, in the exports page's terms (hashes updated to the emulator source, required flags aligned with the code). Files the format cannot carry, withheld updates and content-found labels are printed on a line of their own and stay in the report's counts.
refresh_stale ran a parent's rescrape and its children's truth and
export jobs side by side, so a child read the parent mid-rewrite. A job
now starts once the rescrapes of its ancestors have ended, and a child
no row names gets its export rewritten after them. A rescrape chain ends
on verify.py --platform, whose summary closes the report.
When two cores declare different values for one file, the truth keeps the first and marks the entry hash_conflict. Its hashes pair an entry and never replace the platform's: armsx2's GameIndex.yaml no longer overwrites System.dat's line for LRPS2.
Step 1b read 18 924 held files and members for 7 407 distinct contents and ran every transform on each read. A content already transformed is now skipped (the first path still keeps the match), the power-of-two crc shift operators are squared once for the pass instead of once per length, and a group pairs members by the sizes it holds. derivable.json and the step output are byte-identical, 33.4 s to 25.1 s.
export_native and cross_reference read data/ without the shared hold, so
a refresh could swap a tree mid-walk. hold_data_dirs skipped a cache not
yet on disk, so a refresh could install it between verify and the packs
of one pipeline run. Both now hold every registered tree.
The FBNeo import recorded the listing's blob sha for bytes fetched from a
branch URL a CDN caches, and --fetch <revision> still read the tip. The
revision is now resolved to a commit, listed and downloaded there, each
DAT is held against its blob sha, and a DAT read at a named revision is
labelled with it without replacing the tip's recorded blobs.
The pipeline locked only --output-dir while it also writes database.json,
install/, README.md and docs/. A run on a scratch output directory could
start beside the run on dist/ and rewrite them. Every run now holds the
repository's .lock first.
_patched_entry read '#{ ... }' lines as live items, found no match in the model and gave up, so atarist was written again whole (52 lines changed for 3 additions). A commented item is now left as written.
release_record rebuilt nothing from the name builder: a substring match
took RetroArch_v1.22.2_AndroidArm64V8A_BIOS_Pack.zip for the RetroArch
pack. The record now refuses any pack that is not a registered platform's
full pack and stores the platforms each pack serves, which the readers
use instead of the pack name.
RomM lists the platform firmware folder without recursing and looks each file up as slug:file_name (models/firmware.py, base_handler.list_files at 5.4.0). An addition named with a directory (fuse/48.rom) is not written.
The render writes one biosFiles item per path, as checkBios reads one md5 per item. A second addition at a held path was dropped from the file and still counted added (78 announced, 75 written).
DecideNAND (MelonDS.cs:684-706) keeps, for a DSi title, only the NAND
slots whose bit is set in header byte 0x1B0. The six DSi_Nand entries
now carry that territory, so --region keeps one 240 MB image, not six.
AUS answers oceania, as every other Australian slot does.
Systems a core covers without a file were added while walking a set,
so the hash seed placed them and two runs wrote dist/truth/<platform>.yml
in two orders. The normalized system map was built from a set too.
The buildbot serves nicaiemu_libretro (frozen at 4c9b73f7, last built
2026-10-16 05:15) beside coolbaremu_libretro. They are two builds: each
reads its own orientation list from the system directory, so both
profiles stay, each with its own core name. Both read the same .cbe
archives from the same workspace, which builds a standalone binary at
both revisions: one system id and one build type for the two.
The slot and claim decisions above change which copy some paths take
(EmuDeck neocdz.zip, RetroBat DS firmware, hiscore.dat on six platforms)
and settle case variants on folding filesystems.
Ranking contested core claims sorted the whole gap report, which
reordered every manifest. Only the claims on one contested path move now,
grouped where the first of them stood, and two spellings of a path are
one group where the filesystem folds case (RetroBat's 64DD_IPL_DEV.bin).
An entry without a destination (EmuDeck's system:md5 whitelist) no
longer counts as a pack path in verify, and step 5 prints one status for
counts and paths together. A full offline run compared all 12 archives
and 930 emulator packs path by path, every one equal.
Thirty one-line labels in verify.py, generate_site.py and
generate_truth.py named the statement under them ("# Load registry",
"# Copy branding assets"). Comments that give a reason stay, as do the
module section headers. A test reads the three files for new ones.
The which-pack page announced itself ("This page picks the right pack
for a setup") before naming a pack. The sentence goes, the pointer to
Getting started stays.
Ymir reads sega_101.bin or saturn_bios.bin, so the sega_101.bin entry
carries saturn_bios.bin as an alias. kronos, yabause and yabasanshiro
open saturn_bios.bin and fall back to sega_101.bin: they were in that
entry's _cores but not its _required_by, while their saturn_bios.bin
entry had them requiring it. A reader that requires one of an entry's
aliases now requires the entry. No reader is added, and same-name
entries kept apart by path are left alone. No export changes.
Pin moves to 2de3bf8 (v8.4.2). roms/aros-ext.bin, roms/aros-rom.bin and whdboot/game-data/whdload_db.json changed upstream and are collected at that revision. The WHDLoad database is matched on the archive SHA1 first and refreshed from the bundled copy when newer. kick02019.AROS and its .ext are names the Recalbox launcher writes, so the profile now documents the kickstart_rom_file settings instead.
Pins move to libretro/hatari 5092994 and hatari/hatari 90949cf. The port now reads floppy.raw from the system directory (floppy_sound.c:146-186), documented as an optional user sample. Upstream decodes IPF without capsimage, the icon block moved by four lines, and the launcher file names leave the tos.img note for the platform layer.
Abdessamad Derraz