| Commit message (Collapse) | Author | Age | Files | Lines |
| ... | |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Setting UMU_ID="fluorine" triggered protonfixes' setup_mount_drives
(GE-Proton/utilities.py), which mounts X:=$HOME, U:=/media, V:=/run/media,
W:=/mnt every launch — undoing pruneExtraDrives. Wine's find_drive_nt_root
walks the path innermost-out and picks the first matching drive, so
X:=$HOME wins over Z:=/ for any path under home, producing X:\games\...
launches.
Empty UMU_ID:
- protonfixes utilities.py:38 `os.environ.get('UMU_ID','')` -> '' falsy
-> setup_mount_drives early-exits, no X/U/V/W created
- proton:1590 same falsy check -> 'gamedrive' not added to compat_config
-> setup_game_dir_drive's setup_dir_drive takes the elif branch and
actively removes any stale s: symlink
- proton:2340 `"UMU_ID" in os.environ` is still True for "" -> still
skips the c:\Program Files (x86)\Steam\steam.exe bridge for non-Steam
executables (the original reason we set UMU_ID)
Verified against installed Proton-GE proton script + utilities.py and
GloriousEggroll's confirmation in umu-launcher#634.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
| |
Blanking DISPLAY/WAYLAND_DISPLAY made xrandr exit 1 inside the
container, which Proton-GE 10-34's protonfixes treats as fatal during
wineboot -u. wineboot -u is unattended and won't pop UI, so leaking the
host display vars is safe.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
New "Enable Steam Overlay" checkbox in Instance Manager (off by
default, key fluorine/steam_overlay). When enabled and the executable
has a Steam App ID, ProtonLauncher now:
- LD_PRELOADs gameoverlayrenderer.so (32+64 bit) from the running
Steam install for legacy GL/X11 hooks,
- sets SteamOverlayGameId so Steam's IPC handshake matches the app,
- exports ENABLE_VK_LAYER_VALVE_steam_overlay_1=1 and clears the
DISABLE_... counterpart so the Steam Vulkan implicit layer loads
even under pressure-vessel (this is what makes overlay work for
DXVK-rendered Bethesda games — most "manual" overlay setups skip
this and only get the GL hook),
- ensures Steam is running, and
- exposes <steam>/ubuntu12_{32,64} as SLR --filesystem= mounts so
the .so paths resolve inside the container.
No reaper wrapper, so Steam's "in-game" indicator may stay blank, but
the overlay itself attaches because the .so/Vulkan-layer hook the
process directly. Requires the user to actually own the game on the
running Steam account.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
| |
Pruning only the dosdevices symlinks let Wine recreate them on the
next prefix start from [Software\\Wine\\Drives] in system.reg, so
external mappings like X:=/home/user/Games kept winning over Z: when
Wine canonicalised launch paths (game spawned with X:\Games\... cwd
instead of Z:\home\user\Games\...). Strip those registry lines too.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
| |
std::locale("") at static-init throws std::runtime_error on systems
where the requested locale (e.g. en_US.UTF-8) isn't generated, killing
Fluorine before main() runs. Wrap in try/catch and fall back to the
classic locale.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Wine canonicalises paths against the dosdevices map and prefers the most
specific (longest-prefix) drive letter. An external symlink such as
dosdevices/x: -> /home/<user> turns every Z:\home\<user>\... binary path
into X:\..., which confuses tools and obscures launch logs.
PrefixSetupRunner::stepDriveCleanup already enforces the {C:, Z:} policy
but only runs as part of the explicit prefix-setup pipeline; drives
added later (Faugus, manual edits, modlist installers) survive forever.
Mirror that cleanup at launch in spawn.cpp's pruneExtraDrives() so every
Proton launch sees a known-good dosdevices layout.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add IPluginGame::isNativeLinux() (virtual, default false) so a plugin
can advertise that the discovered installation is a native Linux build.
Bound through the Python plugin trampoline so basic_games subclasses can
override it.
Stardew Valley plugin: when the install dir contains the StardewValley
launcher script and no Stardew Valley.exe, report isNativeLinux=true and
return Linux-side binary names (StardewValley, StardewModdingAPI). Also
fix GameDataPath to "Mods" — Linux ships the SMAPI mods dir with a
capital M.
ExecutablesList::getPluginExecutables: skip the UseProton flag when the
plugin reports native Linux, so the spawn path falls into launchDirect()
and the game runs without a Wine prefix.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Wrap every FUSE callback in a noexcept thunk that catches std::bad_alloc
(→ ENOMEM) and other exceptions (→ EIO) so they never unwind into
libfuse3 (C, no unwind support). Stops std::terminate from orphaning
the mount.
Crash handler: replace raise(sig) with _exit(128+sig) — raise can hang
when the signal is masked on libfuse worker threads. Try direct
umount2(MNT_DETACH) before fork+exec fusermount3 -uz, and double-fork
the fallback so we don't waitpid on a stuck child. Switch signal() to
sigaction(SA_RESETHAND), add SIGBUS, install std::set_terminate.
Cap readdir/readdirplus kernel-supplied size at 1 MiB to defend against
oversized vector(size) allocations.
|
| |
|
|
|
|
|
|
| |
Last commit renamed the QProcess variable but missed the final
extractor->start(), breaking the build. No logic change beyond
completing the rename.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
| |
GitHub Releases serves whatever's uploaded, but some CI flows produce
.zip artifacts (or users re-upload as zip). Teach the release-info
parser to pick up .zip when no .tar.gz is attached, and teach the
install flow to call `unzip` vs `tar` based on the URL extension.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Adds an "Install update & restart" button that:
1. Downloads the release asset (with progress bar).
2. Extracts tarball to ~/.local/share/fluorine/update-staging/extract/.
3. Writes a small install.sh helper that waits for the current Fluorine
PID to exit, then execs the new launcher.
4. Launches the helper detached and quits the current process. The new
launcher runs its existing sync-into-bin/ logic and starts from there.
Button stays disabled until a check has reported an update with a valid
.tar.gz asset attached.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
| |
Upstream MO2 gates the one-time MO2 2.4 -> 2.5 category-migration dialog
on `lastVersion < 2.5`. Fluorine's 0.x.y schema change means lastVersion
is 0.x.y, which is always < 2.5, so the dialog nagged on every startup.
Skip it when lastVersion has major=0 (Fluorine).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- settingsdialogupdates.cpp missed an include of settings.h (it was only
getting the forward-declared version from settingsdialog.h), which
broke every settings().setX() call. Now included directly.
- UpdatesSettingsTab is not a QObject so it has no tr() member; add
Q_DECLARE_TR_FUNCTIONS to get a static tr() via QCoreApplication.
- CI: the docker run chained commands with `;` so a ninja failure got
masked by the trailing `ccache -s`. Capture the build-inner.sh exit
status and `exit $status` at the end so real build failures fail the
step instead of bleeding into the tar packaging step and producing a
confusing "Cannot stat: No such file" error.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- New Settings > Updates tab (src/src/settingsdialogupdates.{h,cpp}):
current build info (version, channel, commit, timestamp), channel
picker, startup-check toggle, and a "Check for updates now" button
with live status. Legacy widgets on the General tab are now hidden.
- Downgrade warning: suppress the "your version is lower than before"
log when lastVersion is 2.x.y (upstream MO2 / pre-Fluorine schema)
and currentVersion is 0.x.y — that's a schema switch, not a
downgrade.
- Titlebar: beta builds show "beta @ <short-sha>" instead of a static
semver so each rolling build is identifiable at a glance. Stable
still shows semver. Renamed to "Fluorine Manager" in the title.
- Controllers: remove the four xinput* "native" overrides from the
prefix registry seed. Games that ship their own xinput1_3.dll
(Fallout 4, Skyrim, etc.) make Wine's builtin (SDL-backed) lose to
a Windows stub that can't talk to /dev/input, so controllers went
silent. Leaving xinput unset lets Wine builtin win.
- ntsync fallback: detect a missing /dev/ntsync at launch and set
PROTON_NO_NTSYNC=1 / WINENTSYNC=0 / WINE_DISABLE_FAST_SYNC=1 so
pre-6.14 kernels stop dying with "Cannot open synchronization
device: No such file or directory".
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- VFS: add mo2_rmdir + OverwriteManager::removeDirectory so Wrye Bash
shutil.rmtree stops failing with WinError 267 (issue #47).
- FileRenamer: case-insensitive fallback via new resolvePathCaseInsensitive
walks each path component; bulk hide in Conflict tab no longer dies with
a misleading "Input/output error" when DirectoryEntry normalised parent
dir case (issue #54). formatSystemMessage stops aliasing Windows error
code 5 to errno 5 (EIO).
- Versioning: FLUORINE_VERSION_* + fluorine_build_info.h generated from
CMake; Linux createVersionInfo returns Fluorine's version. About dialog
shows Fluorine version + MO2 engine + commit. Beta builds stamp
"<semver>B<yyyymmddhhmm>" (issue #51).
- CI: ci.yml determines channel from ref (v* tags = stable, main = beta),
publishes a rolling `beta` GitHub release whose body carries a
machine-parseable fluorine-meta block (timestamp, commit) for the
in-app updater.
- Updater: new FluorineUpdater polls GitHub API on startup, compares
against embedded metadata, notifies on new release without auto-install.
Channel toggle piggybacks on the existing "beta versions" checkbox.
- Bethesda plugins: determineMyGamesPath no longer creates
Documents/My Games/<GameName> for every possible title at plugin load;
only games with a detected install path get their dirs created
(issue #55).
- Starfield: dataDirectory() on Linux now points at <game>/Data instead
of My Games/Starfield/Data so the single FUSE mount lands where SFSE
loads plugins; My Games Data becomes a secondary symlinked mapping
(issue #56).
- Prefix resolution: prefer Fluorine config, then explicit
fluorine/prefix_path, then legacy Settings/* keys (with a warning) so
auto-detected Heroic/Bottles prefixes can't silently override a user-
configured Fluorine prefix (issue #52). compatDataPathFromPrefix no
longer hands Proton the wrong parent dir for plain wine prefixes.
- xrandr: ensureXrandrInstalled back-fills the helper for existing SLR
installs, prefix init runs it before wineboot, SLR wrap exposes
xrandr-bin via --filesystem= and prepends PATH inside the container
(issue #49).
- Process tracking: drop the wineserver-as-last-resort fallback that made
MO2 hang on Proton's session manager after game exit. Rescan the prefix
for matching game executables instead (covers reparented launcher
children like f4se_loader → Fallout4.exe). Unlock button now kills
wineserver for the prefix (SIGTERM then SIGKILL on timeout).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Replace the per-launch __MO_Saves symlink with a kernel bind mount
inside an unprivileged user+mount namespace. Writes from the game
land directly in the profile's saves dir via the mount, which Wine
can't replace with a real directory the way it could with a symlink.
The namespace (and therefore the bind) tears down automatically when
the game process tree exits, so afterRun no longer needs to sync
saves or undeploy anything. Fall back to the symlink path on
kernels where unprivileged_userns_clone is disabled (Debian buster,
linux-hardened); detection is via /proc/sys/kernel/unprivileged_userns_clone.
DownloadListView::issueDeleteSelected captured indices up-front and
looped emit removeDownload(row). DownloadManager::removeDownload
calls refreshList() at the end, which clears finished entries from
m_ActiveDownloads and rebuilds from disk enumeration — stale indices
after the first delete pointed at the wrong file or triggered the
invalid-index error. Capture filenames instead and re-resolve the
current row each iteration.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Wine picks its Unix codepage from LC_ALL/LC_CTYPE/LANG via
nl_langinfo(CODESET); a C/POSIX locale falls back to CP1252, so
non-ASCII Linux filenames (CJK, Cyrillic) fail UTF-8→UTF-16
conversion and MSVC std::filesystem throws "Invalid name" when
Windows tools like PGPatcher iterate mod dirs through the Z:\ drive.
Steam's pressure-vessel can strip the user locale, triggering this.
Detect UTF-8 in LC_ALL/LC_CTYPE/LANG and only inject C.UTF-8 when
none is UTF-8 already — preserves de_DE.UTF-8, ja_JP.UTF-8, etc.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Bethesda Plugin Manager disabled in settings still swapped the espTab
at UI init — the callback ran unconditionally. Gate it on
isPluginEnabled and warn + flag restart when toggling panel plugins,
since the swap is one-shot.
Pre-deploy sync-back: if prefix Plugins.txt mtime is newer than
profile (external LOOT run), syncPluginsBack first so the deploy
doesn't clobber the external edits.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Replace copy-in/copy-out save deployment with a direct symlink from the
prefix's __MO_Saves to the profile's saves/. Writes land in the profile
immediately; afterRun reverts the symlink + prefix INI so a vanilla
launch outside MO2 uses the default Saves dir.
- Parse Steam's binary appinfo.vdf (v41) via a new Rust FFI crate backed
by steam-vdf-parser. Rank detected games by common.type so real Games
beat Tools/Editors when multiple prefixes share a My Games folder name
(fixes Creation Kit shadowing Skyrim SE). Top-ranked candidate
overwrites stale symlinks from earlier runs.
- Switch bundled Steam Linux Runtime downloader from sniper (steamrt3)
to steamrt4 so Proton 11+ works out of the box. steamrt4 omits xrandr
so we inject it from the Debian x11-xserver-utils .deb into a
dedicated dir and prepend it to PATH via `env PATH=...` inside the
pressure-vessel container (PATH env doesn't propagate).
- Auto-kill stale wineboot/wineserver/pv-adverb processes still bound
to the prefix before starting a new wineboot -u; match by
/proc/<pid>/environ WINEPREFIX / STEAM_COMPAT_DATA_PATH since Wine
cmdlines are Windows-style. Same sweep runs in destroyPrefix so a
delete actually frees file handles.
- Blacklist Proton 11 (Steam-bundled) — its current Wine tree deadlocks
during wineboot -u on modern kernels via ntsync / futex_wait_multiple
races. Use "waitforexitandrun" (not "run") for prefix init so
use_sessions=1 Protons don't fork into a persistent session manager.
- Drop WINEDEBUG=-all during prefix init so wineboot progress is
visible. Suppress focus-stealing by setting WA_ShowWithoutActivating
on the PrefixSetupDialog and SLR download progress dialogs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Proton 11's toolmanifest sets use_sessions=1, so the "run" verb forks
into a persistent session manager that never exits — prefix init hung
forever. Use "waitforexitandrun" instead. Also drop WINEDEBUG=-all so
init output is visible, and migrate the bundled Steam Linux Runtime
downloader from steamrt3/sniper to steamrt4 (Proton 11's required
runtime). SLR lookup paths fall back to sniper for users with an
existing install.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
deployPlugins now writes only the plugin-list file the game actually
reads (Plugins.txt for PluginsTxt games, lowercase plugins.txt for
FileTime games). loadorder.txt is MO2-internal and never belonged in
AppData/Local/<Game>/; previously writing a stale copy there defeated
LOOT's edits because syncPluginsBack pulled the stale loadorder.txt
back into the profile, and getLoadOrder() preferred it over the
freshly-synced plugins.txt.
syncSavesBack now mirrors deletions: files present in the profile but
absent from the prefix are removed before the copy pass. Without this,
saves the user deleted in-game stayed in the profile forever because
copyTreeContents only copies files that exist in the source.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When stepProtonInit failed, the runner kept marching through every
downstream step against a half-initialized prefix, producing a cascade
of misleading "wine client error: version mismatch" failures that
masked the real problem. Seen in the wild with a GE-Proton10-34 install
that was missing concrt140.dll from its bundled default_pfx template:
setup_prefix crashes with FileNotFoundError, then seven unrelated steps
are reported as broken.
Break out of the step loop as soon as proton_init fails, and inspect the
captured Proton output for the "FileNotFoundError" + "default_pfx" hint
so the error message points at the broken Proton distribution instead
of a generic "exit code 1". runProcess gains an optional QByteArray*
out param so the caller can pattern-match on the merged output without
changing any existing call sites.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
OrganizerCore::modStatusChanged updates the DirectoryStructure in
place (enabling/disabling origins) but never fires
directoryStructureReady, which is the only signal MainWindow used to
drive DataTab::updateTree. Result: after enabling or disabling a mod
the Data tab kept showing the old virtual tree until the user hit
Refresh manually.
Connect ModList::modStatesChanged directly to DataTab::updateTree so
the file tree re-queries the structure as soon as the toggle lands.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
syncPluginsBack previously updated the profile from the newest case
variant but left the stale sibling (e.g. lowercase plugins.txt after
LOOT edited Plugins.txt) untouched in the prefix, so anything reading
the prefix before the next deployPlugins saw divergent content. Mirror
the newest variant into every sibling after the profile sync.
installer_omod.py called mobase.log / mobase.LogLevel which the Python
bindings do not expose, crashing .omod installs with AttributeError.
Route messages through stderr so the Python runner forwards them to
MOBase::log::error.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
LOOT (and any other tool that edits the plugin list) writes to the
deployed copies in the Wine prefix AppData. afterRun was syncing
saves and INIs back but not plugins, so LOOT's reordering was
silently reverted by savePluginList() writing the old in-memory
order on top of the untouched profile file.
Add WinePrefix::syncPluginsBack that picks the newest case variant
of plugins.txt and loadorder.txt from the prefix and copies them
back to the profile. Run it in afterRun alongside the existing
save/INI sync. For FileTime-based games (Skyrim LE, FO3, FNV) the
loadorder.txt removal now happens after the sync, so the file-time
fallback still works.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
| |
The auto-prompt in organizercore already offers to fix the registry
on launch, so the manual button is redundant.
BethINI Pie plugin compared against managedGame()->gameName(), but
GameFalloutNV returns "New Vegas" while BethINI's app folder is
"Fallout New Vegas" — add an alias map so NV is detected correctly.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
| |
Adds per-tick rusage snapshot (user/sys seconds, delta since last
tick, wall delta, busy%, max RSS) to the [VFS] stats lines so we
can tell CPU-bound from IO-bound slowness without external tools.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
VFS:
- Open backing fd at mo2_open time for read-only handles so mo2_read can
splice without re-opening the file on every read call.
- Bump RLIMIT_NOFILE at mount time to fit the resulting fd pressure when
the game keeps hundreds of BSAs open.
- Dedicated node_cache_mutex so resolveByInode / mo2_lookup stop racing
unordered_map writes while multiple tree_mutex readers are active.
- max_read=1MB + matching conn->max_read and raised max_readahead/max_write
so the kernel merges Wine's small sequential reads into bigger FUSE
requests.
- Per-op wall-clock counters in mo2_init() logs so we can distinguish VFS
latency from game-side work.
Proton launch:
- Write dxvk.conf to the prefix and set DXVK_CONFIG_FILE to force
dxvk.enableGraphicsPipelineLibrary=False, avoiding long GPL compile
stalls on first run.
UI / plugin compat:
- Clamp IconDelegate's per-icon width to [8, 16] so narrow content
columns don't trigger QIcon::pixmap(0,0) returning null and logging
"failed to load icon" every repaint.
- Pre-create lowercase symlinks in the stylesheet directory so QSS files
authored on Windows resolve url(foo.svg) when on-disk files are Foo.svg.
- Flip content icons empty-chessboard.png and facegen.png to 8-bit RGBA
so Qt's built-in PNG reader loads them uniformly.
- Add mobase.Version.canonicalString shim for legacy Python plugins that
still call the old VersionInfo method on appVersion()'s return.
- BSPluginList: compare normalized plugin name lists instead of byte
hashes so the post-run case-fix refresh stops spuriously firing the
"load order changed" dialog.
- BSPluginList disabled by default.
Download manager:
- Parse CDN URL with QUrl + QUrlQuery and read the filename from the
response-content-disposition query param (RFC 6266 quoted / unquoted /
ext-value forms), not QFileInfo on the raw URL.
- When the API or URL gives us a CDN object key (no archive extension),
prefer the actual Content-Disposition header from the live response in
metaDataChanged and downloadFinished.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bsatk:
- Archive::extractToMemory() extracts a file directly into a byte buffer
instead of a filesystem path, backed by a std::ostringstream sink.
- Archive::findFile() locates a file by relative path, case-insensitive
and separator-agnostic, walking the folder tree recursively.
- extractDirect/extractCompressed now take std::ostream& so both the
ofstream and ostringstream paths share the same implementation.
libbsarch: shim rewrite on Linux to map the Windows API surface onto
bsatk::Archive, enabling preview_bsa to open vanilla BSAs via the
same interface as other plugins.
preview_bsa: double-click on a file in the BSA tree now extracts the
bytes via extractToMemory() and hands them to IOrganizer::previewFileData
for nested preview rendering (e.g. NIF inside a BSA).
uibase + organizerproxy + organizercore: new IOrganizer::previewFileData
virtual that routes in-memory file bytes through the first registered
IPluginPreview matching the extension. Uses heap + WA_DeleteOnClose +
open() to avoid nested exec() softlock from mod info filetree ->
preview_bsa tree -> nif preview.
docker/build-inner.sh: stage preview_nif shaders into plugin_data/shaders
so ShaderManager can load them via IOrganizer::getPluginDataPath().
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Vendored from github.com/Parapets/mo2-preview_nif and adapted for Linux:
- BSA texture resolution via MO2 virtual tree with case-insensitive
fallback walk (Linux FS is case-sensitive; NIFs reference textures
with arbitrary case against files on disk)
- Process-wide cached BSA candidate list, keyed by profile path,
priority-sorted, filtering disabled mods
- Path backslash normalization before loose-file lookup
- Polygon offset on SLSF1_Decal-flagged shapes to break z-ties with
coincident base meshes (road decals, moss overlays)
- Opaque framebuffer via glColorMask alpha lock so alpha-blended shapes
can't bleed the dialog background through the preview area
- QMouseEvent::globalPosition().toPoint() for Qt6 compatibility
Preview dialog behavior (organizercore.cpp):
- Switch previewFile and previewFileWithAlternatives from stack-allocated
exec() to heap + ApplicationModal + show() + WA_DeleteOnClose.
- Parent passed as nullptr so preview lifetime is decoupled from the
stack-allocated ModInfoDialog that spawns it. Nested exec() inside
the enclosing dialog's exec() was softlocking on close.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The sort button was already hidden on Linux (setVisible(false)) and
the LOOT feedback path (LootDialog -> addLootReport) has always been
Windows-only, meaning the dirty/incompatibilities/missingMasters/messages
tooltip bullets and the dirty-plugin icon could never fire on Linux.
MO2 shipped lootcli + libloot for a load-order sort mechanism that was
unreachable from the UI.
Users who want LOOT can download it from https://github.com/loot/loot
and run it against their instance directly.
- Drop libs/lootcli/, libloot Dockerfile build step, and lootcli
staging/patchelf in build-inner.sh
- Drop src/src/loot.{cpp,h}, src/src/lootdialog.{cpp,h,ui}
- Extract MarkdownDocument / MarkdownPage (used by UpdateDialog for its
changelog view) into src/src/markdowndocument.{h,cpp}
- Strip addLootReport, makeLootTooltip, Loot::Plugin field, LOOT icon
checks, and LOOT tooltip/isProblematic/hasInfo paths from pluginlist
- Delete sortButton widget, updateSortButton(), onSortButtonClicked(),
and all m_didUpdateMasterList wiring
- Delete lootLogLevel setting, combo box, and "Integrated LOOT" group
from the diagnostics settings tab
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Linux port uses its own in-process FUSE VFS (FuseConnector), so
usvfs was dead weight: libs/usvfs/ was never built, usvfsconnector.cpp
was explicitly REMOVE_ITEM'd from the source list, and every call site
was already guarded by #ifdef _WIN32 with a working Linux branch (the
Windows build is also broken on this branch, since FUSE3 is a hard
find_package dependency).
- Drop libs/usvfs/ and src/src/usvfsconnector.{cpp,h}
- Collapse #ifdef _WIN32 / usvfs include branches in organizercore,
mainwindow, settings, util, spawn
- Remove getUsvfsVersionString() and the "usvfs: ..." log field
- Remove vfs32/64DLLName APPPARAMs (nothing reads them)
- Relabel About dialog "usvfs:" row to "VFS: FUSE 3"
|
| |
|
|
|
|
|
|
|
|
|
| |
backslash
The checkGameRegistryKey map still had "Installed Path" (capitalized) and
the winePath was missing the trailing backslash. When the user clicked "Yes"
on the mismatch dialog it would overwrite the correct lowercase key with
the wrong casing and strip the trailing separator.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
| |
Wine registry value names are case-sensitive. Game launchers (FNV, Oblivion,
etc.) look up "installed path" (lowercase) but Fluorine was writing
"Installed Path" (capitalized), causing "Unable to find an INI file" errors.
Also adds trailing backslash to match Steam's path format and fixes the
registry mismatch dialog to ignore trailing separator differences.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
xEdit (and similar tools) rename game files to create backups before
saving (e.g. DLCHorseArmor.esp -> TES4Edit Backups/...backup...).
OverwriteManager::rename() only searches staging and overwrite
directories, so renames of backing (game) or mod files failed with
EACCES since those files live in the real game/mod directories.
Now when OverwriteManager::rename() can't find the source file, the
VFS copies it to staging at the destination path via COW instead of
failing. The file disappears from its old path and appears at the new
path in the virtual view, without modifying the real game/mod dirs.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
| |
The readyRead handler consumed data via readAll() but discarded it when
no destination file was set. The final readAll() then returned empty
since the buffer was already drained, causing all in-memory HTTP
responses (like BUILD_ID.txt) to appear empty.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
| |
- FO4VR: Add Daytripper4.dll to ESL support detection
- Remove empty categories tooltip when mod has no categories
- Fix QDialogButtonBox/QMessageBox enum mismatch in profile dialogs
- Use refreshESPList(false) after LOOT sort (no full re-read needed)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Wine uses renameat2(RENAME_NOREPLACE) when translating MoveFileW() calls
where ReplaceIfExists is FALSE (e.g. xEdit saving plugins). The VFS was
rejecting all non-zero flags with EINVAL, which a real filesystem like
ext4 handles correctly. This caused "Could not rename" errors in xEdit
and similar tools.
Now properly handles RENAME_NOREPLACE: checks if the destination exists
in the VFS tree and returns EEXIST (matching ext4 behavior), otherwise
proceeds with the rename. Adds diagnostic logging on EACCES failures.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Replace NaK Rust crate and nak_ffi with native C++ implementations:
- Game detection (Steam, Heroic, Bottles), VDF parser, icon extraction
- Prefix symlinks, SLR manager, Steam path detection
- Known games database
Add FUSE VFS optimizations: zero-copy reads via fuse_reply_data,
default_permissions mount option, FUSE_CAP_ASYNC_DIO, FUSE_CAP_EXPIRE_ONLY,
lookup cache parent index, cached mode bits, increased I/O buffer sizes.
Update build system, prefix setup, and various fixes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
| |
XDG_RUNTIME_DIR may point to a read-only location on Steam Deck
(SteamOS has a read-only root filesystem). Use ~/.local/share/fluorine/tmp/
instead, which is always writable.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
| |
XDG_RUNTIME_DIR may point to a non-existent path on some
configurations, causing QLocalServer::listen to fail with
"Name error". Ensure the parent directory exists before
attempting to create the socket.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
| |
ENOENT from process execution was always shown as "No Wine Prefix"
when useProton was set, even when the prefix was properly configured.
Now checks FluorineConfig::isSetup() first — if the prefix exists,
falls through to the generic spawnFailed dialog which shows the
actual missing binary path.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
- Add inode→VfsNode* cache to skip splitPath + full tree walks on
lookup/getattr/open (O(1) hash lookup instead of tree traversal)
- Combine canonicalChildName + snapshotForPath into single lookupChild
function (one tree lock, one parent resolution, one child hash probe)
- Add userspace lookup cache keyed by (parent_ino, normalized_name) to
absorb Wine's case-variant probes that bypass the kernel FUSE dcache
- Add mo2_access handler to avoid kernel fallback to lookup+getattr
- Invalidate all caches correctly on tree mutations (create/unlink/rename/mkdir)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
- Merge all DirectX June 2010 extractions into single "DirectX Runtimes"
step (d3dcompiler_42/43/47, d3dx9, d3dx10, d3dx11_42/43, xinput, xact)
- Merge .NET 6/7/8/9 runtimes into single "NET Runtimes (6-9)" step
- Replace .NET 9 SDK with .NET 9 Runtime (9.0.14)
- Add .NET 10 SDK (10.0.201)
- Remove .NET Desktop Runtime 10
- Add xinput, d3dcompiler_42, d3dx10, d3dx11_42 extractions
- Move .NET Desktop Runtime 6 before runtime installs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Direct DLL extraction and installer execution instead of shelling out
to winetricks for all prefix dependencies. Fixes d3dcompiler_47 being
sourced from wrong SDK cabs (broke CommunityShaders), and eliminates
the fragile cabextract-in-pressure-vessel PATH issue.
- d3dcompiler_47: direct download from Mozilla fxc2 repo
- d3dcompiler_43, d3dx9, d3dx11_43: cabextract from DirectX June 2010
redist with correct system32/syswow64 mapping
- xact/xact_x64: cab extraction + batched regsvr32 registration
- vcrun2022: direct installer with msvcp140.dll workaround (Wine #57518)
- dotnet6/7/8, dotnetdesktop6: direct x86+x64 installer pairs
- Game detection: batched into single .reg file import
- All DLL overrides added to WINE_SETTINGS_REG
- SHA256 verification for downloaded files
- Winetricks kept only for win11 mode step
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Handle modl:// links alongside nxm:// for downloading mods from
mod.pub. mod.pub uses a different URL format than Nexus — the download
URL is embedded as a query parameter: modl://game/?url=<encoded-url>
- Register x-scheme-handler/modl in desktop files and MIME lists
- Accept both nxm and modl schemes in URL parsing (NxmLink, NXMUrl)
- Parse modl:// direct download links and route to startDownloadURLs
- Make key/expires optional in NxmLink::parse (not used by mod.pub)
- Add debug logging for received URLs to aid troubleshooting
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Move Wine prefix creation from NaK Rust FFI to native C++ with a new
modal dialog showing per-step progress, retry support, and proper
failure cleanup.
- PrefixSetupRunner: runs each step (proton init, drive cleanup, 11
individual winetricks verbs, .NET installers, game detection, wine
registry, win11 mode, post-setup) as discrete retryable operations
- PrefixSetupDialog: step list with status icons, filtered log output,
cancel/retry-failed/delete-prefix buttons
- Downloads use QNetworkAccessManager (no host curl dependency)
- Zip extraction uses downloaded 7zz binary (no host unzip dependency)
- Winetricks runs on host (not in pressure-vessel) so cabextract is
always findable via PATH
- Wine/winetricks debug spam filtered from log output
- Skip Python basic_games proxy since native equivalent exists
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
| |
New executables added via the Modify Executables dialog now have
"Use Proton" checked by default, matching user expectations for
a Linux mod manager.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Replace SHA256SUMS verification with Content-Length check for SLR
downloads. Valve's CDN frequently serves stale SHA256SUMS that don't
match the current archive, causing persistent checksum mismatches.
Fix CMake DEPENDS for nak_ffi to glob all .rs sources instead of
hardcoding a few files — previously changes to slr.rs (and most other
Rust files) were silently ignored by ninja.
Also wrap prefix init and runtime commands through SLR pressure-vessel
container, and expose Proton install directories to the container.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
| |
|
|
|
|
|
|
| |
The downloadView widget was missing selectionMode and selectionBehavior
properties, defaulting to SingleSelection. All other list views in the
UI already had ExtendedSelection set.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|