Online pipeline at 169776790, generated twice with identical output: database, recipes, install manifests and games index, native exports, README, installer and report excerpts.
KNULLI's RetroArch is Batocera's build with its core downloader on.
Its armv7 hard-float boards (atm7039, r16, rk3128) fetch from buildbot
linux/armhf: atm7039's shipped retroarch.cfg hides the menu entry but
leaves the URL, and Settings offers the entry back. Its aarch64 boards
get no buildbot URL. ecwolf, squirreljme, stella2014 and vemulator come
back as installable cores, and openmsx, which no buildbot list carries,
stays out.
Batocera builds RetroArch with its core downloader (configure defaults,
qb/config.params.sh:149-150), shown in the menu since it is no Lakka
build (config.def.h:743-745), and configgen sets no URL, so the compiled
one applies: buildbot linux/x86_64 or linux/armhf, none on aarch64
(config.def.h:1855-1861). The targets scraper names that list per board
(downloads, an armv7-A CPU read from arch/Config.in.arm), the builder
and the published target lists add it, and the platform names the union
as installable_cores. applewin, gong, squirreljme and vemulator come
back that way. A build option that compiles the downloader out, or a
shipped retroarch.cfg that blanks its URL, turns it off.
ROCKNIX listed retroarch, which stood for every libretro profile (351
resolved), while its devices build 124 libretro packages. The platform
list now names the cores those packages install on at least one device.
The targets scraper read a package's name as its core: vice-lr installs
vice_x128, vice_x64, vice_xpet, vice_xplus4 and vice_xvic, which reached
no device. It now reads each package's makeinstall_target, the file
copied to ${INSTALL}/usr/lib/libretro under its target name, through
TARGET_ARCH and ENABLE_32BIT (pcsx_rearmed-lr, desmume-lr). 155
profiles resolve, each built by a device, and no device loses one.
No Batocera board turns on BATOCERA_CORES_STILL_COMMERCIALIZED or
BATOCERA_CORES_EXPERIMENTAL (Config.in:109, 1670), so citron, ryujinx,
applewin and eka2l1 are not built, and freej2me, gong, squirreljme, uqm
and vemulator are selected by no Kconfig file at batocera-43.1. Read
with every sourced Kconfig file, the Pi 4 boards gain dxx-rebirth,
sonic2013, soniccd and xroar (mesa3d-rpi4 gives them OpenGL).
parallel_dc (libretro/flycast renamed), hyperscanemu, hypseushw, n64emu
and odin reached the RetroArch buildbot after the last targets scrape and
resolved to no profile. Targets rescraped, each core profiled from its
source.
EmuELEC's scripts/image writes COREELEC_DEVICE to os-release (v4.8, line
189), as a GameForce image shows. The installer reads it the way it does
for AmberELEC and applies install/targets/emuelec.json. MiSTer FPGA takes
EmuELEC's place in the test of a platform without targets.
AmberELEC's metapackage adds LIBRETRO_CORES_EXTRA (bsnes, Mesen, melonDS, geolith) only on RG552, RG353 and RG503, and deletes those cores from es_systems.cfg on every other device (packages/amberelec/package.mk:23-31, 45-58 at a21975bf). The new target scraper reads the Makefile's device images, that test and es_systems.cfg the way the platform scraper does: RG351P, RG351V and RG351MP list 125 cores, the other three 136. The installer reads the device from COREELEC_DEVICE in os-release (scripts/image:168).
The target scraper kept its own copy of the EmuScripts table and took
.ps1 files the platform scraper never read, so emudeck-we's template
skeleton was published as a core and removed afterwards by an override.
Both scrapers now read emulator_of_script, and the platform scraper
lists the installers at the pinned revision, as launchers: and the
SteamOS target already did. Targets rescraped, override dropped.
The target scraper stripped -sa from every standalone package while the platform scraper keeps both spellings. vice-standalone, scummvm-standalone and supermodel3 answer to vice-sa, scummvm-sa and supermodel-sa only, so every targeted ROCKNIX build left them out. Both now read standalone_emulators(), and the targets are rescraped at tag 20261001.
Batocera and KNULLI write the image name, not the board, to the card's board file (post-image-script.sh). The target scraper now records each board's images from Config.in, the target lists carry them, and the installer maps a card's image to its board. KNULLI's boot/knulli.board is read too.
KNULLI builds per board on Batocera's Kconfig model from its own tree (knulli-cfw/knulli-linux at the tag platforms/knulli.yml pins): configs/knulli-*.board, package/system/knulli-system/Config.in, knulli-es-system/es_systems.yml and the knulli-cfw/buildroot submodule. Ten boards, the armhf library builds excluded.
Continued if-conditions were cut at their first line, so GLES3 held on every board, and selects guarded by a negation stayed on once fired. Derived flags are read line by line, each pass rebuilds the symbol set from the board, and the architecture comes from the board's BR2_<arch> choice.
The pack builder accepts the aliases declared in the target overrides,
so --target switch works there, but the installer's target manifests
carried only canonical names. The documented word was the one that
failed, and the installer then carried on with every file: 1911 files
and 4.1 GB where the user had asked for the 863 that target needs.
Aliases are emitted beside their canonical target, and an unknown
target now stops the run and lists what is available. A filter is
applied or refused, never ignored.
YAML 1.1 reads an unquoted 01 as the integer 1 and 81 as 81, so the
manager set stopped naming the file its driver opens and the API
published "cores": [2048] as a number where consumers match strings.
Ground truth from MAME 0.289 at the revision the profile cites:
src/mame/vtech/crvision.cpp:957 loads "01" and "23".
The root cause was the scraper: _hash_merge wrote these names through an
f-string, so quoting the profiles alone would be undone on the next
refresh. _yaml_scalar now quotes every name, archive and description it
writes.
Abdessamad Derraz