Between the check and the creation of the run folder, retrobios-backup could be swapped for a link. The holder's device and inode are confirmed unchanged once the run folder exists. Directory file descriptors would close the window entirely but do not exist on Windows; the remaining window is the instant between that confirmation and the move.
When every dated name in retrobios-backup already existed, the loop ended on the last one and wrote into it. It now raises, so the next location is tried and the file stays in place if none works. A holder another user owns is refused on systems with user ids.
A BIOS folder can sit on a share other people write to. A retrobios-backup that is a link is no longer written through (the backup goes inside the BIOS folder instead), run folders are created for the user only and carry a marker, and --restore only takes a folder holding that marker, walking it without following links. Beginner pages label their links to the installer page as details.
A file in place whose contents differ (another revision, a personal
dump) was overwritten and lost. It is now moved to
retrobios-backup/<date-time>/ beside the folder written to, under the
same path, before the verified copy replaces it. Beside, not inside:
DuckStation, SwanStation and Rustation read every file of the BIOS
folder and BizHawk walks its firmware folder recursively.
Identical files are never kept. --dry-run counts what would be kept,
the summary and the summary event name the folder, files and bytes,
and --restore FOLDER puts the files back. Data trees and standalone
copies keep their replaced files the same way.
install.py --no-game-data leaves out the files and omissions the manifest
marks game data, in an install, --check and --dry-run, and names them with
their category. The --events summary gains left_out, one group per
category with every destination. A list without categories is refused.
minarch hands every core SDCARD_PATH/Bios/<TAG> as its system directory
and opens no file in it. The scrapers read each pak's launch.sh (core,
tag, devices), each device's core makefile (the project a library is
built from: MyMinUI's dosbox_libretro.so is dosbox-pure) and the minarch
statement, at a pinned commit. The installer recognises the card from
.system/version.txt and refuses a folder inside it.
AmberELEC, PAN4ELEC and EmuELEC detection now installs their own lists
at /storage/roms, the root those lists are written against; KNULLI its
own list at /userdata/bios. JELOS, KOLOS, UnofficialOS and Arch R keep
RetroArch's list with the folder their retroarch.cfg names.
One JSON object per line on stdout (detection, per-file and per-system
check, download progress, summary, exit code) and the usual text on
stderr, so a program driving install.py never parses prose. EVENT_KEYS
is the one schema. The installer page documents it, and a test holds the
two equal. The handheld port reads it.
The Google Play and retroarch.com builds share com.retroarch and com.retroarch.aarch64 but build their defaults under different roots (RetroArch v1.22.2 MainMenuActivity.java:182-183). The build is told by the version code it records in retroarch.cfg, else by the folder it wrote. A Google Play folder Termux cannot write is named, as for DraStic and Flycast.
DraStic keeps its BIOS in its private data and Flycast in Android/data,
which Termux cannot write on Android 11 and later. An installed one is now
named with the files not copied, and a writable Flycast folder is a
standalone copy target.
RetroDECK moves each of its folders on its own and records the new place
in retrodeck.json paths (api_data_processing.sh:913-935). Only
rd_home_path was read, so files went to <home>/bios after a move.
AmberELEC, EmuELEC, JELOS, KOLOS, UnofficialOS, ArchR, PAN4ELEC, dArkOS,
RetroOZ, spruceOS, CrossMix-OS, Onion, Allium, Koriki, MiyooCFW and plumOS
are found by what their builds write, named as they name themselves, and
installed to the system_directory of their RetroArch config, which follows
ArkOS and dArkOS to a second card. TrimUI stock, MinUI and a plumOS build
without a known folder are named and not installed to. Two unrelated
systems matching stop a run without a terminal.
The builder keeps a held regional fallback when the requested file is
missing (since 2026-10-17), and the comment beside the manifest's
omissions still said the opposite. A file the builder holds but the
manifest cannot download sat in omitted_files counted unheld, so the
installer kept a fallback the pack drops. Such an omission now carries
held: true and narrow() counts it held, as the builder does.
jp, us, kr and gb were aliases but au was unknown, and the first
spelling offered for it was auto, which means world. Every territory
of the region tree now answers to its ISO 3166-1 code, in the build
scripts and in the installer, whose SHA-256 is pinned again.
Without Python, install.ps1 at a console now offers to run winget install Python.Python.3.12, rereads PATH from the registry and carries on. The Store's python.exe placeholder fails the version check like a missing Python.
CommentsSayWhy read three scripts while 22 comments in 13 other files
named the next statement. They are removed or replaced by the reason,
and the test reads every tracked script and install.py.
With no platform detected, a machine running standalone emulators (ES-DE with DuckStation and PCSX2) was offered RetroArch's folder. The installer now fills the emulator folders of the standalone copies list that exist here, from the EmuDeck then RetroArch file lists, asking first at a terminal.
Every install ended on 'Standalone copies skipped', on machines with no standalone emulator as on one whose DuckStation lacked its BIOS. The line now lists the emulator folders that exist here, and nothing is said when there are none.
A mistyped --dest was created with every missing parent and reported done. The run now names the missing folder and the nearest one that exists, asks at a terminal before creating more than one level, and prints the folder it created. The free-space check now runs before the folder is made and ends with 2, nothing installed.
--list-systems, --list-cores, --list-emulators and --list-targets with nothing detected were answered with an install command that dropped the listing and added --dest. The hint now names the platform and keeps the listing.
The manifest now carries system_names, each system's native_system names from the platform's own file, and --system, its suggestions and --list-systems read them: --platform batocera --system psx installs sony-playstation.
EROFS was advised like a permission refusal, and sudo fails the same way on a read-only mount. The advice now reads the mount flag and says to check and remount the card or drive.
EmuDeck's macOS build reads ~/.config/EmuDeck/settings.sh like the Linux one (functions/all.sh:3-4,24 and vars.sh:2 at dddb3022). The settings and ~/Emulation/bios rules now run on darwin too.
RomM 5.4 reads each platform's firmware where filesystem.structure.firmware
in config.yml says, bios/{platform} by default, and suggests {platform}/bios
for a {platform}/roms library (rommapp/romm 5.4.0,
backend/config/config_manager.py:58,101-109,205-225,564-568,1105-1124).
The installer only wrote bios/<platform>/. It now reads the template from
config.yml beside the library, takes --romm-firmware when that file is
elsewhere, refuses a platform-first library with no template, and detects
RomM by /romm/config/config.yml too.
No Batocera build writes /etc/batocera-version: batocera-system.mk:21-22
(batocera.linux 38845725, unchanged since batocera-30) writes
/usr/share/batocera/batocera.version. Knulli (knulli-linux 57afa662) adds
OS_NAME=knulli to os-release, writes /usr/share/knulli/knulli.version and
removes /usr/share/batocera, and no Knulli source writes /etc/knulli-release.
Neither was detected. A Knulli board is named and not narrowed by
Batocera's per-board lists, and an unlisted board is named.
muOS runs RetroArch on a config copied from its default, which reads
/run/muos/storage/bios, bound to MUOS/bios on the second card when that
card holds one, and loads cores from /opt/muos/share/core (MustardOS/internal
f41f4a89, script/control/retroarch.sh:8-20, script/device/bind.sh:116-150,
retroarch.default.cfg:2806,3151). The installer wrote to /mnt/mmc/MUOS/bios
and --installed-cores found no core. The config now decides both.
functions/all.ps1:4-14 (emudeck-we 037460ef) reads the APPDATA
settings.ps1, else the one under the profile, and takes C:\Emulation when
neither names a folder. Only the first was read and a forced platform fell
back to %USERPROFILE%\Emulation\bios. A network install names its share
through $VolName (JSONtoPS1.ps1:166-172), and a UTF-16 settings file is
decoded.
RetroBat.exe writes its folder to HKCU\SOFTWARE\RetroBat
LatestKnownInstallPath on every start (retrobat-executable 702da66a,
Program.cs:57 and 99, RegistryTools.cs:15-50). Only C:\RetroBat and the
profile folder were looked at, so an install on D: was missed and a forced
--platform retrobat wrote into C:\RetroBat\bios. The value is read
through winreg, or reg.exe under WSL, before the fixed folders.
A run whose downloads were all cut ended on "Location that was being
written to" and "once that is resolved", pointing at the disk. A
failure caused by the connection (lookup, reset, timeout, TLS, a cut
transfer, HTTP 5xx or 429) is now marked as such, and the summary says to
check the connection and HTTPS_PROXY.
An unknown --emulator name pointed at --list-cores and counted 441 names
while --list-cores prints 189: the buildbot names --emulator accepts
(mednafen_psx for beetle_psx) were listed nowhere. --list-emulators prints
each accepted name beside the emulator it stands for, and the hint names
it.
kronos tags the ST-V BIOS epr-19854.ic8 taiwan, mednafen_stv tags it
asia-ntsc, the SMPC area Mednafen names for Taiwan (ss.cpp:1119,
2363-2364). As siblings under asia they never matched, so a request in
one spelling dropped the file tagged in the other, and --region
taiwan,us dropped the Saturn sega_101.bin. AREA_TERRITORIES relates
each area to the territory it serves, china for asia-pal included, in
region.py and its installer copy.
A [Parameter()] attribute made install.ps1 an advanced script: its common
parameters took -v (-Verbose), -vb, -d and -ea before install.py saw them,
and the scriptblock form printed PowerShell's own VERBOSE lines. The
wrapper now forwards $args as typed.
A link already at the destination was skipped without a word and counted
as "dir not found" while the folder existed, so the emulator kept the file
the link pointed at. The link now gives way to the file, written beside it
and renamed over the name, and a name leading out of the folder is a named
failure.
On Windows a verified partial copy is closed before its rename. A second
run that claimed it in that moment truncated it and the first rename failed
with a sharing violation, reported as an unwritable folder. A complete
partial is now left whole there, the rename waits for a sharing violation
to clear, and a sharing violation no longer reads as a refused folder.
install.region_drops now skips one-file groups (:file:) and lets no
omitted match displace an installable file, as region.resolve_region_drops
does on the builder side. The property test draws both cases and
compares against the builder.
Names are matched on letters and digits, so the site's display names
work, and a typo gets the closest names instead of a full list. Hints
and advice give back the command the user ran (curl or PowerShell form,
every flag kept), --dest alone follows detection and says whose list it
installs, Ctrl+C at a question exits 130, and --cores-dir names a core
folder. install.ps1 runs in its own scope so irm | iex leaves the
session alone, and an override that causes a fingerprint mismatch is
named. --check lists hash mismatches first, missing required files are
grouped by system, and --help ends with examples.
On Windows the partial copy of a file of 16 MB or more is locked with msvcrt.locking, a LockFile region: the locking process cannot access that region through a second handle (Microsoft LockFile remarks). Hashing reopened the path, the read failed with a sharing violation, and every large file was reported as 'folder not writable' (issue 82). Both digests are now read in one pass through the handle that holds the lock. A read-only file already at the destination had the same misleading report: on Windows its attribute is now cleared before the verified copy replaces it, and a refusal that remains is named for what it is.
install.py --emulator NAME keeps what the named emulators read on the detected platform: their own files and the platform's files of the systems they run, as a pack built for them carries, without the systems no core is mapped to. Each emulator page now gives that command for Linux, macOS and Windows, beside the per-file links.
A copy refused by a read-only emulator folder raised the error count with no failures entry: the summary named no file, gave the BIOS folder as the location and no advice. The failure now carries its path, and the summary names the emulator folder and what to do.
--check counted the required files the list leaves out and named none, then said all files were up to date. A Recalbox user asking why Caesar III would not start learned nothing, though the list names c3.555. The count line no longer calls a file not found unsafe.
The live line divided every byte by the run's time, a partial copy left by an earlier run included, and showed a 3 GB resume as 2.5 GB/s with nothing left.
Abdessamad Derraz