libretro/hatari rewrote main on 2026-11-04: it is now hatari/hatari
c453b85f with 23 port commits on top, and the old pin 50929946 survives
only as the tip of main-old. Every citation is read again at 25f488b1 and
c453b85f. The libretro build now reads files through
src/retro/file_retro.c, with no .gz or archive unpacking, and a missing
tos.img clears bUseTos in retro_init instead of a have_bios gate in
retro_run, so the core runs the built-in test stub.
A bulk advance moved fheroes2, supertux and supermariowar past the
revision EmuELEC compiles, and nothing stopped it: bump_commit refused
built_by only. It now leaves a profile a single registered platform ships
until it carries pin_frozen or the new pin_follows, which says that
platform builds the branch tip.
hode, bermuda, bennugd-monolithic and ccleste are held on the
PKG_VERSION of their EmuELEC v4.8 package.mk, equal to their pins.
rigelengine was one commit past it (README.md only) and goes back to
f05996f9.
bermuda.yml (the EmuELEC build) and bsyndrome.yml (RetroBat's bs.exe)
document the same cyxx/bermuda engine under two slugs, bermuda-syndrome
and bsyndrome, so the site gave the game two system pages. bsyndrome now
files its entries under bermuda-syndrome.
In a fork's clone origin/main is the fork: once pushed it is HEAD, and
profile_sync --changed --strict checked nothing and passed while the PR's
CI then failed. The default base is now the line validate_pr.py reads:
upstream/main or origin/main when that remote is the project, the
project's main fetched otherwise.
openMSX reaches a ROM in share/systemroms through its file pool by SHA1
alone. A filename is only looked up beside the machine's XML
(src/memory/Rom.cc:121-170, HardwareConfig.cc:279-280,
FileContext.cc:152-156 at a557de7b). The labels read as names the core
opens, and the export added 15 of them to KNULLI and 2 to Batocera.
DEV9 opens eeprom.dat and flash.dat by bare name (pcsx2/DEV9/DEV9.cpp:144
at f0523f8a, AetherSX2 DEV9/flash.cpp:89, PCSX2 1.6 dev9ghzdrk/flash.cpp:73),
so the process working directory answers, never the BIOS folder. Only
yaps2 said so: the other PCSX2 profiles had the file packed into the BIOS
folder and the RetroDECK export add it there.
PCSX2 and its forks keep the first 4 to 8 MB image with a romdir in
the BIOS folder (pcsx2/ps2/BiosTools.cpp:258-282, 317-332 at f0523f8a),
and PCSX2 1.6 lists the same test in its BIOS picker
(BiosSelectorPanel.cpp:151-177). The names are labels. Without found_by
the truth model read ps2-0230a-20080220.bin as a name the core opens, and
the RetroDECK export added it as a required file.
Notes compared a core with another profile of this directory (unlike
beetle_pce, unlike the other engines in this directory), which means
nothing once a profile is read alone, and shouted NOT, ALL and MUST. The
facts stay, the comparisons and capitals go. Comparisons with a core's
own upstream remain.
The profile declared libretro/redream, which answers 404, with no pin and
no source_ref: profile_sync skipped it and the boot.bin md5 came from the
libretro docs. It now reads inolen/redream at ffb73022, the last public
tree, whose libretro core opens dc/boot.bin and dc/flash.bin in the system
directory. boot.bin is checked against four md5 and the flash against its
size, and both fall back to the high level BIOS, so neither is required.
needs_review counts an orphaned pin and a moved repository, but the
triage line only added stale prose and unchecked refs, so pcem read
"1 to review" for an upstream_commit outside its branch history.
idtech4a++ cites xash3d-fwgs filesystem/filesystem.c, starship and
ghostship Torch src/Companion.cpp: files of submodules of the declared
repository. No blob of the declared tree answers, so the parts stayed
external and a bump moved the submodule commit with nothing checked.
A project word now also resolves to the one submodule named after it,
and a file under a submodule is read in the repository .gitmodules
names, at the commit the declared revision pins it to.
core_version went null although doc/CHANGES.html at the pin opens on
Version 0.16.0. ti99sim cites v0.16.0 and v0.0.11 parts side by side,
and with no documented version every part read as another program:
profile_sync reported 6 external, 0 to review, nothing checked. The v0.16.0
parts are anchored at the pin again, the v0.0.11 parts stay external.
XRoar reads the Dragon 200-E and Alice ROMs for those machine configs
(src/xroar.c:437-444, 552-558 at 1.12.1), and VICE the Japanese kernal
and chargen for the C64 JAP and VIC-1001 models: a (Japan) file name
selects the NTSC model (libretro-core.c:2018-2034). No code branches on a
territory for these files, so region: is removed, as profiling.md rules
for a model enum.
CLK reads every export cartridge as Europe or USA and boots the western
BIOS for any region but Japan (StaticAnalyser.cpp:49-66,
MasterSystem.cpp:137-140). The tag named brazil without its parent, so
--region mx dropped the BIOS from the Batocera pack.
higan and nSide look program.rom up in the option's folder, SGB.sfc/,
then Super GameBoy.sfc/. Only the SGB1.sfc/ and SGB2.sfc/ folders were
entries, and both names are files bsnes, Mesen-S and ares read from the
same system directory, so no pack could give the higan cores a Super
Game Boy. SGB.sfc/ collides with SuperSnes9x's sgb.sfc on a
case-insensitive file system. Super GameBoy.sfc/ collides with nothing.
x1-box and hakuX both import the retail partition C tree. It sat in
hakuX's tree, so for x1-box the C/xbox.xtf tail was set aside as foreign
and the name step served Xenia's Xbox 360 dashboard font. xbox.xtf and
xboxdash.xbe now live in bios/Microsoft/Xbox/C/ and resolve by path for
both. msdash.xbe stays: its bytes are cxbx-reloaded's xboxdash.xbe, whose
path the RetroBat manifest names.
A one-member archive whose member carries the zipped_file md5 under
another name has that md5 as its composite. The name step returned it as
md5_composite_exact, which slots and the site take as content proof, while
Batocera looks the member up by name and finds nothing.
reminiscence's Flashback module reunion resolved to skyline's
zoneinfo/Indian/Reunion through the case-insensitive name step, and the
emulator pack shipped the time-zone file. Another casing found only in
another emulator's tree says nothing about the entry.
GitLab refused the hundred raw reads per IP. A GitLab directory is now
read from one archive request, else a shallow sparse git fetch, and the
cache keeps its file list so offline runs know what is absent.
Each board builds system.ini against its defconfig: a core whose package
the board does not enable is dropped (pisnes was the SNES default
everywhere), and a system with no core is not built. The row names the
default most boards share, board_defaults the others. A files: list
names whole files and is carried as files through the index and the
installer.
A folder no name resolves, run by a core of several machines, now holds
those machines: thomson, multivision and pcv2 had no fallback, so their
games got no file. Arcade-indexed cores keep naming files per set.
board_cores walked only the plain [core.N] sections and replaced a board
section whole. romfs2 systemholder.py reads a slot present as
[core.N.<board>] alone (gb and gbc bsnes on x86_64) and each key from the
board section, then <key>.<board>, then the plain key. The committed
target lists are unchanged on all seven boards.
A Sony fixture split by console recorded its pages in the shared map, and the index tests that ran after it linked PlayStation to the split page. They now start from an empty map, and the fixture clears it when done.
Tests that built a maker's page with sections used Sony, which now gives each console a page: they use Sega, which keeps its sections. A Sony fixture split by console leaked its recorded pages into the index tests that followed. Built-page checks read the cards on the console pages and the opening lines on the Sony summary.
A section shares its page's search fields, and "playstation 3 bios" found the PlayStation page before the Sony page that held the PS3. CONSOLE_PAGES splits Sony by console whatever the size. The search gate now expects the PlayStation 3 page in the top two for that query, the order of words being something Pagefind does not weigh, and the page first for "ps3 bios".
The games importers wrote through fixed .part names, the new options had
no help text, and romfolders stopped on the first setup the offline cache
lacked. RetroArch's rom folder map is regenerated from its source spec
(path null, as the schema says). The installer's Python 3.8 job imports
the games tests without pyyaml, so the region rules load lazily.
In a console folder an FBNeo set is looked up under its driver name
(msx_1942), the bare name being the arcade set's. A run given its list as
a file reads the games index from beside it or the cache only when
--for-games was not asked. The installer keeps to Python 3.8 syntax, and
the testing guide names test_install_games.
Same files per platform. The manifests catch up with the per-system MAME
filing, the platform system join and declared_in, and name sfiii2a.zip as
missing. The games index fallback sets and the installer page samples
follow.
--for-games reads the card's rom folders through install/games/ and keeps
what the games need: per system what its default core reads, one region
where the content proves it (crc32 and size, a disc serial, a header the
core reads), an arcade set's BIOS parents. A game not recognised gets
every BIOS of its system and is named. --required-only keeps only what
starts. The plan and the line menu open on it when the card holds games,
everything one press away. A cache keeps what reading a card found.
mame2000, mame2003 and mame2003-plus carry a DAT their code moved past. Their parents and BIOS sets now come from src/driver.c and src/drivers/ at the pin, named by the src tree sha. The DAT is still read to report gaps. FBNeo DATs are read one namespace each, console sets keyed by driver. A BIOS with no dumped ROM is no longer listed.
The games index now names, per arcade core, the game sets that load no
BIOS set and every BIOS set, so a set on a card the DAT does not know is
told apart from one that needs nothing. Each board default core of a rom
folder gets its own fallback set, and index.json maps every folder name
a setup uses to its systems, for a games folder on a computer.
nintendo-sgb and nintendo-super-game-boy are one id now, so the flat
declared-set test uses two ids no alias joins. The search weighs "3" as
nothing: "playstation 3 bios" is held to the top three results.
MAME's first card part showed 1,931 words against the 1,693 ceiling once
888 entries named their system: each row gained a "for <system>" badge
while parts were cut by card count. Parts are now cut by the words each
row shows with its card closed, balanced, at most 1,350 a part and still
at most 180 entries. MAME goes from 9 parts to 10, no part disappears.
Abdessamad Derraz