An install writes retrobios-backup/<folder>.install.json beside the BIOS
folder: each file it placed with its sha1, size and source, and the choice
it made. The next run for the same setup and folder takes that choice when
the command line chose nothing, in line mode and on the full screen, and
says so. A record that cannot be read is said once and ignored. A record
event reports it.
With one setup found, the full screen opens on its plan, Change where one
row away. The plan's Install row says the size its headline says.
--verbose set it and no run cleared it, so a later run in the same process ended its clock message with the raw SSL text. Also: --target named again in its refusal, three names mypy read as another type, a semicolon and a tooling file named in test comments.
test_install loads its own install module. Its free-space test ran
main() in process without naming the OS, so on a WSL host main()
entered WSL mode and left _WSL active in that module, for the rest of
the process. Any test of the module run after it, as when the module
is named twice on one command line, read D:\bios as /mnt/d/bios. The
test now runs as Linux and restores _WSL on exit.
Run on the device detection found (KNULLI, Batocera), the last line said to start that system, which is already running. A computer still names its emulator, and a card mounted on one is still put back.
Six messages on stderr ran 26 to 35 words before the option to give. Each now says the problem and the remedy in at most 25 words, and a test measures every stderr message against that budget.
Five lines the default text hid are back where the user reads them:
- the result names each failed file again, after Exit too;
- a run, --check, the plan and the result count the files an emulator
needs that are not available here, --verbose or Details naming them;
- --check names each copy that fails its check, the files that sit too
deep and what a run copies, and says such a copy is backed up; only
the plain missing files wait on --verbose;
- the plan and the result keep the count of save files left out and the
remedy, the reason going to Details;
- the plan counts the --from files that could not be read where no row
names them, and says another drive to add.
Exit codes are unchanged. The result's word budget rises by the failed
file line.
The plan, Confirm and the result say what happens, its size, where and
the next step. Their counts, free space, backup and save notes, kinds,
reasons a file is left out, per-file failures and commands move to a
Details screen: d on the plan, Confirm and the result, the Details row
on the plan, Y, o or Tab on the result. Opened on the result, they are
also said in the terminal after Exit. Installing lists the files that
went well only with --verbose, and --verbose draws every line in place.
The log line reads Log: so that Details names one thing.
--check prints its count, the folder and the remedy in both modes, so
that --verbose prints every default line. Each golden screen gets a word
budget per size, and a test proves the screen and its Details together
equal the --verbose screen.
By default a run prints what it installs, the BIOS folder, the progress
at a terminal, the result and the next step. The steps, the per-file log
lines, the counts and the reasons a file is left out come back with
--verbose. --check gives the count and the folder, and lists each file
with --verbose. The installer page shows the default run and walks
through the --verbose one.
Batocera and KNULLI start shadPS4 from CONFIGS/shadps4, so its user
folder sits below /userdata/system/configs. The scrapers read that root
from batoceraPaths and write configs_root with its source_ref. The
manifest names it configs_directory, and install.py honours it through
the same checks as roms_directory: character rules, at most two ups,
the ancestor match and the lexical bound.
A drive is refused only at the start of a value, so a later segment
such as ../C:x joins as a drive-relative path on Windows. Every
manifest path now refuses a segment Windows reads as a drive and one
ending with a colon. roms_directory, emulators_directory and the dests
below them refuse any colon. Elsewhere a colon inside a name is kept
where the host is not Windows: RetroDECK's ares reads
bios/snes:st018.data.rom and bios/zxs:48.rom on Linux, three dests of
install/retrodeck.json. A Windows host refuses them as alternate data
streams. The roms root is also checked to stay below the BIOS folder's
ancestor it climbs to.
roms_directory.from_bios and the emulators_directory folders come from a
downloaded list. from_bios now meets the character rules of every
manifest path (shared with _safe_relative_path: no backslash, drive,
empty segment, trailing slash or control), climbs at most two folders
(../../Roms, the deepest the rom folders files record) before plain
names, and its joined form must equal the BIOS folder's ancestor at that
depth: "." is the BIOS folder, null leaves the files out. Every
emulator folder the installer builds goes through _safe_relative_path
and stays below the folder holding the BIOS folder.
Run from the retrobios package, a hint and --help named python3 install.py, a file that user does not have. The package's launcher sets RETROBIOS_BOOTSTRAP=pypi, and the command is read from the environment it runs in: uvx retrobios from uv's cache, pipx run retrobios from pipx's, retrobios once installed.
- The installer speaks without "you": the first screen asks "Where
to install?", rows read "Copy the BIOS files from an old card or a
folder", a close match reads "Close matches: ...". The handheld
launchers and their port text follow, and the pronoun guard no
longer lists any installer message.
- A setup that runs as an app opens on its one-line command, the
folder table after it; a setup that runs on its own device keeps
the network folder first, which its ZIP route needs.
- The home steps say what to do: "Its page gives one line to copy",
"Paste it, press Enter".
- The installer says "Every file on the list is in place." where it
said "Everything is in place.": a check covers the files the
collection holds, and the required ones it lacks follow.
- The FAQ, first install, legal and home pages state only what the
code shows (PlayStation 1 and Saturn run without a BIOS on some
emulators, save files are never touched, records are added).
- MinUI and MyMinUI show their own boot logos, and the Android,
Linux handheld and iPhone tiles show the Android robot, Tux and a
plain device icon, each credited in sources.json.
- Each Markdown copy, llms.txt and llms-full.txt name the commit
they were built from, version.json the current one, and the
answering rules tell assistants to fetch a stale copy again.
- The deploy lists the pages whose words changed (digests.json) and
posts them to IndexNow once the site is live.
The SHA-256 pin already refuses any changed byte: the cap in install.sh
and install.ps1 only stops a sign-in portal's page before it is hashed,
so it can be generous. The size test now fails once install.py passes
3 MiB, so its growth is looked at, and checks both bootstraps accept the
same 8 MiB. The separate size test of the earlier commit is folded into
the one test_install.py already had.
--for-games --required-only on a dArkOS card from a computer kept every
file of the list: --dest skipped the card's detection, so its rom
folders were never found. The games part beside the BIOS folder now
names them. The manifest marks a data tree every core runs without
(optional), and --required-only leaves it out: one GBA and one SNES game
took FBNeo's cheats and samples, 3,472 files no game opened.
tests/__init__.py keeps a run offline, its scratch under tmp/ and its
installer cache its own, but only discover -s tests -t . imported it.
discover tests ran with RETROBIOS_OFFLINE unset, scratch on the /tmp
tmpfs and the user's real installer cache, the setting in which e2e
fixtures had written into the shared .cache/large. Every test module now
imports the package first, and a test fails on a module that does not.
test_verify_listing.py took its line with the previous commit.
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.
RetroPie was the only platform with status: archived, which kept it out
of verify --all, slots, the pipeline and list_platforms unless
--include-archived was typed. release_pack already decides which packs a
release publishes and the installer serves every platform, so the status
set nothing apart. The field, its schema entry and the flag are gone, and
RetroPie is covered by every full run like any other platform.
A data archive cut mid-download was deleted on every failure: the retry and the next run started from zero, while the run said nothing was lost. The archive is now kept beside its tree with the validator its server sent, and the rest is asked for with Range and If-Range. Checked against the release: fbneo-samples stopped at 16 MB, the next run downloaded 34.6 MB of 50.5.
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).
Every change to install.py meant a sha256sum pasted by hand into
install.sh and install.ps1, and the tests that caught a stale pin
failed on two bare hashes. scripts/pin_installer.py writes both pins
(--check only reports), the pin tests name it, check_freshness reads
the pins through it and refresh_stale runs it for a stale pin row.
A tarball tree was announced at its extracted size: yaps2 at 6.5 MB for a 21.2 MB download, Cemu's game profiles at 9.9 KB for 21.4 MB. refresh_data_dirs records the size of each archive it downloads (a tree current from before is measured, not refetched), the install lists carry it as archive_size, and the plan, the menu and the data event count it.
RetroPie, ROCKNIX, AmberELEC, EmuELEC and RomM keep their save directory in no fixed place, and their lists leave those files out with no hash or address. The run said to name the folder with --save-dir, and said it again when --save-dir was given. It now says the list holds none of them.
Two listing options printed one list, and --list-platforms, --list-targets and --list-emulators printed the same output with --system, --region or --core as without, code 0. A run now prints one list, and an option a listing ignores is refused.
A BIOS folder that is a file, or lies under one, failed at mkdir with EEXIST or ENOTDIR, and the run blamed permissions and offered sudo. It is now refused before anything is checked, with an error event whose reason is not_a_folder.
Native RetroArch on Linux keeps retroarch.cfg under $XDG_CONFIG_HOME/retroarch when the variable is set, and falls back to ~/.retroarch.cfg (file_path_special.c fill_pathname_application_data, configuration.c open_default_config_file at e98e1da0). Detection read ~/.config/retroarch only.
Abdessamad Derraz