Each pin advance left website/src/data/source-paths.json behind and the
site's freshness test red until sitesources.py --fetch ran by hand.
profile_sync now rebuilds it after any run that wrote a profile, online
and not in a dry run.
After the second bring-to-head pass: 40 bumps, ten profiles by hand,
five off-forge apps, the G-NET BIOS sets and the new hiscore.dat.
Manifests generated twice, identical.
The Apple II page named its start files with no visible way to get one: the download sat after the collection path, fourth of a card's facts. Each card now carries a Download button in its title, each start-line name links to its card, and a card held on another part page is reached by its permanent link.
Branch tips, tags and releases are cached for six hours, and the
EKA2L1, trzy/Supermodel and VICE heads read as behind their real HEAD.
--fresh asks the forges again, and is refused with --offline.
A recale on a profile whose pins are keyed by build mode said the prose
had to move first. It now says the pins are set by hand.
Seven HID and UI commits (keyboard packets, Esc opens the in-game menu). The XeFu warning moved to emulator_window.cc:1995-2005; no file loading changed. xenia-libretro is still at f23c14c.
Three trzy commits: billboard output device, per-game GUI settings, wnaspi32 link flag. No file loading changed. The notes now say the GUI also rewrites Supermodel.ini (Gui.cpp:1492). Libretro-Supermodel is still at 84bc106.
16 upstream commits (MIDI, Avkon message queries, guest screen mode, iOS card mount flags). Android, iOS and Qt citations moved by one to three lines; firmware and ROM loading unchanged. The libretro branch is still at e39ccac.
The fork rebased on trzy e048c09 (Src/Version.h:32 now 0.3a-e048c09-mm), one commit touching the drive board outputs only; upstream_commit held there with upstream_frozen. The 0455189 release archives carry the same Games.xml, Music.xml and crosshair bitmaps. Notes name the arm branch commit ROCKNIX builds.
Bare tiletex.cc, database.cc and maps.cc resolved in crawl/crawl once upstream_commit moved. They describe the libretro fork and now carry crawl-ref/source/, read at cc2530f. upstream_commit to a5ed934.
26 VICE-relative citations resolved in the VICE mirror once upstream_commit moved. They describe the tree vendored under C64.emu/src/vice and now name it, read at 1c12fac. upstream_commit to 2e2a43e.
Bare a7800.cpp and xm.cpp resolved in mainline MAME once upstream_commit moved. Each citation now names its path in the fork, read at 7a2afdc. upstream_commit to mamedev 9d8a672, where the PAL BIOS is still c300558-001b.u7 and the 7800 ROMs carry no ROM_OPTIONAL.
Upstream 453b679c adds the Taito G-NET driver to zn.c with two BIOS roots, coh3002t (eight members, four BIOS EPROM revisions) and coh1002t. metadata/hiscore.dat gained three Asura Buster sets: new hash. Pin 03fc843.
Read from NO$GBA.EXE of no$gba-w.zip v3.06. The boot ROM table, loader and every file read match v3.04, which RetroBat still ships; only the addresses moved.
Pro 7.4.17 has no genuine copy anywhere; no_genuine_copy moves to it. The native references hold at the same lines in the DBTecno-signed Basic 4.1.15 bundle; its Java slot gate moved to MainActivity.java:3207-3213.
Pro 3.9.0 has no genuine copy anywhere; no_genuine_copy moves to it. The native and Basic Java references hold at the same lines in the DBTecno-signed Basic 2.3.22 bundle, which also carries the multiplayer BIOS CRC32 exchange.
Linkboy 4.0.0 keeps the three firmware slots, sizes, md5 table and loader behaviour of 3.11.4. Class names and library addresses are re-read from the 4.0.0 Play build, and the firmware import path is documented.
The README, the installer steps, the choice of what to install and platform detection still described the line prompts as what a terminal shows. They now name the full-screen interface first, with --plain for the questions, and say that no default folder is taken where detection was refused.
Same chain as the first pass: no ref to review, no file name or hash
literal added to a cited file since the pin, no profile a single
platform builds at a fixed commit, refs recaled with their pins, and the
written prose read again.
The Flatpak RetroArch's folder at 000 hid the retroarch.cfg naming ~/Games/bios, and the run offered 6.6 GB to ~/.config/retroarch/system. A closed folder under HOME or above the default now leaves no default: --dest is asked for. Python 3.14's quiet Path.exists() refusals are told from absence by os.stat, and an assumed folder reads as the one a frontend usually reads.
A platform list that declares one path for two files fails one of the
two checks on every install. The export report lists each declaration
with its hash and the emulators whose profile reads that file.
Core extras placed under the bios prefix are now one claim per path, so
DuckStation's region-tagged PS2 images carry their regions in the
manifests. Generated twice, identical.
_MISSING_DATA_DIRS is module state: a test that called main() after one
whose data directories were not cached exited 1 with the earlier run's
notice. main() now clears it first.
The collector deduplicated a destination already carrying the extras
prefix at prefix/prefix/name while the builder wrote it at prefix/name.
DuckStation's standalone bios/<name> on Recalbox and PCSX2's scan then
held one path under two keys, and --region withdrew only one. The
second pass also no longer claims a path another core's extra holds
for a name only that extra covers.
One closed folder under HOME (~/.config/EmuDeck at 000) made the whole
detection return nothing: a RetroArch whose retroarch.cfg was readable
was lost and the run fell back to a folder it does not read. Every
detection probe now goes through _exists, _is_dir or _is_file, which
count a refused place as absent, name the closed folder once, and let
the other probes go on. The decorator stays as the last net.
The refusal is an unsearched event before detect. The interface keeps
the line above the list of setups. At HOME or a drive root A goes into
a visible subfolder, never a card or drive row nor .cache, and with
none it stays on Use this folder and says where it is. A default folder
taken because the platform was not found is no longer called detected.
After section A of the round 5 audit and the ROCKNIX standalone profiles
(data_root on 58 profiles, 67 display names, three alias folds, 33 core
names mapped, scummvm and vice standalone, yabasanshiro both modes).
Manifests generated twice, identical.
A branch tip is cached for six hours. scummvm-standalone was pinned on
6f7b060d while the cache still held 77d80d7, its ancestor, so profile_sync
showed a HEAD behind the pin and a bump would have moved it backwards. A
tip that is a strict ancestor of the pin is now asked for again.
Covers ROCKNIX scummvm-sa and AmberELEC SCUMMVMSA, SDL2 builds with SDL2_net. Pinned to master 6f7b060d with hashes of the files there. Adds eem.dat, macventure.dat, fonts-imgui.dat, wwwroot.zip and monkey4-patch.m4b to the set the SDL1 build reads.
ROCKNIX vice-sa builds the 3.8 tarball (svn r44900) with the SDL2 UI. Lists the ROMs, palettes and keymaps make install ships, which ROMs mem_load needs at startup, and the three CMD drive ROMs the release does not carry.
The retro_arena front end of sydarn/yabause pi4-update (a40dace1), which ROCKNIX and AmberELEC build, opens a BIOS only at the -b path and runs HLE otherwise. The three named files stay libretro-only.
With no usual folder the browser opened at HOME on Use this folder, and A pressed through the screens installed the BIOS files into ~. At HOME or a drive's root A now opens the first folder. Picking the old card to copy from is unchanged.
Abdessamad Derraz