The target scraper treated every select in Config.in as unconditional
and never set the symbols Buildroot sets, so nearly every board listed
the same 171 cores and xroar, which needs OpenGL, reached none. Selects
now fire from their own config, to a fixed point, from the flags each
board file sets, with Buildroot's arch and OpenGL provider rules read at
the commit Batocera pins. A board that builds no image is no target.
RetroDECK filed coco as tandy-coco and Dragon as dragon-32-64, MiSTer
filed COCO3 as tandy-coco and Recalbox kept its raw ids, so none of
them met xroar's coco, coco3 and dragon64 and its ROMs reached no
system. The scraper maps now agree and a test holds every shared
native id to one system.
The truth model merged a pathless entry into another core's entry of
the same name even when their contents disagreed, so the export wrote
one core's size next to the other core's md5.
BizHawk embeds ares as Ares64, its N64 core, whose 64DD IPL its own
FirmwareDatabase declares. Mapped to the whole ares profile, it pulled
thirty systems into the BizHawk pack and made slots accuse
FirmwareDatabase.cs of a wrong MSX.rom it never declared for ares.
Two --from-md5 runs with different hash lists wrote the same file, the
second replacing the first, and a pack limited to thirty systems built
a name longer than a file name may be. The custom pack now carries a
digest of its hash list, and a long system list keeps its first system,
a count and a digest, as the emulator pack does.
Dolphin files the Australian country code under Region::PAL, whose
IPL lives in GC/EUR. Tagged europe alone, --region australia dropped
it and kept the American IPL, which Dolphin never reads for those
discs.
verify counted a declaration whose hash the frontend refuses as
occupying its path, while the builder leaves it out and a core file
takes the place; the report then called that core file missing from
the pack. It now reads the builder's verdict. The hidden-extras tests
move to RetroArch and RetroBat, which still fold case, and test the
same-bytes twin again instead of skipping.
Both run on Linux but inherited RetroArch's case folding, so a core
file differing from a platform name only by case left their packs:
MesenCE's bios.col and np2kai's lowercase 2608_*.wav, which those
cores open with that exact case. They are now case-sensitive, the
shared groups fold only where the platform does, and Lakka gets its
own pack instead of sharing RetroArch's.
hypseus-singe asks for gamecontrollerdb.txt by name alone and, first
in profile order, took the BIOS root of Batocera, RetroBat and
RetroPie with C-Dogs' small table while Solarus pins its own by sha1.
A destination two cores claim now goes to the one naming its content.
lrps2 names no hash for its GameIndex.yaml, and the only full-tail
copy was ARMSX2's: it counted as path proof and replaced the file
RetroArch pins on desktop targets, which have no ARMSX2. Upstream
PCSX2 took ARMSX2's patches.zip by name the same way. A copy in
another emulator's tree now proves nothing at any tail length, and
LRPS2's own GameIndex comes from the tree the buildbot publishes.
The registry list added clk and gsplus, which RetroDECK does not ship
as components, current or archived, and overrode the list the scraper
reads from the component manifests.
Recalbox compiles libretro/mame at 0.258 and 0.278 itself; typed
libretro, both profiles were resolved for every all_libretro platform
although the buildbot RetroArch installs from has neither. A profile
now names the platforms that build it (built_by), and is reached only
when a platform lists it by name.
RetroBat's platform file carried no core list, and a registry list
copied from Batocera stood in with ports RetroBat never runs and none
of its standalone emulators marked as such. The scraper now reads
es_systems.cfg at the release tag: libretro cores, and every other
emulator as standalone.
es_bios.xml names only the cores that need a BIOS, so a hand-kept
registry list stood in, with mame and mamemess that Recalbox does not
ship and without vice, ymir, vircon32 and others it does. The scraper
now reads each system.ini of recalbox-romfs2 at the scraped tag, and
the registry list is gone.
Recalbox ships libretro/mame built at 0.258 and at 0.278, and no
profile described either release. mame_hash_scraper gains --release,
which reads the BIOS root sets at one mamedev tag and writes a profile
pinned to it; mame0258 and mame0278 are generated from it, every
source_ref anchored at its tag.
Batocera's cores were read from configgen-defaults.yml, one per system,
and a hand-kept registry list overrode even that; DOSBox Staging,
melonDS DS and nine other cores a user can pick in ES were outside the
pack, while mamemess, which Batocera does not ship, was in. The
scraper now reads es_systems.yml, the registry list is gone, and the
target scraper names standalone emulators (amiberry, fsuae, vice)
instead of only the machines listed under them.
The emulator pack recorded only loose files as a path's occupant, so a
path an archive or data tree had filled was reported missing even with
the same bytes in the ZIP, and an archive several profiles ask for was
counted once per profile. verify registered neither, and credited one
profile's file at a path holding another's. Both now record archives
and tree members in pack order; --system snk-neogeo-mvs went from 113
missing in the pack and 51 in verify to 59 on both sides.
realign_prose read a failing git log as an empty history and stopped
silently at an unreadable revision, so the stale-prose check answered
"nothing to move" and let --bump-commit through. It now reports such
citations to read again instead of passing them.
Step 2 refreshed the data directories, then the pack build and the
manifest build refreshed them again, so a tree tracking a moving
branch could differ between the pack and the manifest. generate_pack
gains --no-refresh-data, which packs the cached trees and still
fetches release assets; the online pipeline passes it to both builds.
The RetroArch and Lakka pack was built under RetroArch's own name and
renamed afterwards, which overwrote then carried away a --platform
retroarch pack in the same directory, and failed on Windows where
os.rename refuses an existing target.
--system snes joined 23 emulator names into the ZIP name and died
with ENAMETOOLONG before writing anything. The system pack now passes
its own name in, and an --emulator selection too long to join keeps
its first name, a count and a digest of the selection.
Abdessamad Derraz