Luks2.unlock() ran 'cryptsetup open' with no error handling, so a failure raised a bare CalledProcessError. Python renders that exception with only the exit status and discards the captured output, so cryptsetup's stderr (merged into stdout by run()) never reached the install log.
encrypt() already wraps its cryptsetup call and raises a DiskError that includes the captured output. Mirror that for unlock() so a failure reports the actual cryptsetup message instead of an opaque traceback.
Reported in #4327, where the underlying 'device-mapper: crypt: unknown table type' error was hidden from the log for this reason.
* fix: restrict EFI partition permissions with fmask/dmask=0077
Mount the ESP with fmask=0077 and dmask=0077 to prevent world-readable
files like /efi/loader/random-seed.
Closes#4241
* fix(efi): collapse fmask/dmask dedup to dict.fromkeys one-liner
Per @Torxed's review feedback. Same semantics as the previous loop
(dedupe by exact-string match) but shorter. dict.fromkeys preserves
insertion order, where set() would not.
* fix(efi): drop defensive list wrap per review
The list() copy on line 378 was load-bearing only if options were
mutated downstream, but the EFI branch reassigns options via
dict.fromkeys() (line 381) and the non-EFI branch passes through
to mount() without mutating. Drop the copy.
* Fix broken localization: tr(f-string) never matches translation catalog
tr(f'Invalid configuration: {error}') evaluates the f-string before
tr() runs, so xgettext extracts the literal placeholder as the msgid
while runtime passes the formatted string - the two never match.
Switch to tr('...{}').format(...) and update msgid in base.pot.
* Add CI validation for translations and pot_tools dev utility
Add translation-check workflow with two jobs:
- validate-po: msgfmt --check on changed .po files, .mo sync warning,
tr(f-string) anti-pattern grep on changed .py files
- validate-pot: verify all tr() strings exist in base.pot when .py
files change
Workflow only triggers on .py/.po/.pot file changes.
Add scripts/pot_tools.py developer utility (stats, list, add_missing)
for managing base.pot.
* Fix code style: use tabs and reformat xgettext arguments
Align check_pot_freshness.py and pot_tools.py with project
indentation (tabs) and ruff format requirements.
Sorry :-)
* Replace custom PO parser with msgcmp, drop pot_tools.py
Address review feedback: use standard gettext msgcmp instead of
hand-rolled parser for base.pot freshness check. Remove pot_tools.py
that duplicated locales_generator.sh functionality.
* Move translation checks into locales_generator.sh, simplify CI workflow
Use msgcmp instead of diff for base.pot validation to avoid failing on
legacy stale entries - the same cascading breakage that killed the
original workflow (disabled 2023, removed in #4483).
* Fix broken .po files: duplicate msgid in Hindi, missing format args in Finnish
* Add iwd standalone option to network configuration
Adds NicType.IWD as a third option alongside NM and NM_IWD: install
iwd, write /etc/iwd/main.conf with EnableNetworkConfiguration=true and
NameResolvingService=systemd, enable iwd.service + systemd-resolved.service.
iwd handles DHCP itself and resolved picks up its DNS via the symlink, so
no NetworkManager pulled in.
Also fills in parse_arg cases for NM_IWD and IWD so config files
round-trip both nic types.
Assisted-By: Flint
* wire up resolv stub and networkd service
* exclude virtual devices, dont harcode iface name instead match 'ether'
similar pattern to .network files shipped in /etc/systemd/network
* use dedent and rename menu option
I think it was a mistake to have made the previous changes to KDE Apps. In retrospect, it seemed like a good idea since the Budgie developer had done it that way in Fedora, which is the distro the developer uses as a reference for Budgie. When you start using it, you realize there's no bridge between the desktop and the KDE Apps, and things like Dolphin recognizes icons and themes, requiring some rather annoying manual configurations. Then, if you want to change them again, you have to change those configurations again. Files don't open by default with the apps either; you have to configure them for that to work.
At first, I thought the Budgie packager for Arch had forgotten some stray dependency, but with some free time, I tested it with Fedora 44, and to my surprise, it has exactly the same problems, which is completely unacceptable for a final stable release. I suppose he'll make the necessary changes in the near future, but right now, it's a disaster.
* Add --share-log flag to upload install.log to paste.rs
* Apply ruff-format to share_log.py
* Rework share-log per review: subcommand, TUI confirmation, truncate large logs
* Replace curl/SysCommand with urllib.request, remove defensive try/except
Per review: use stdlib urllib instead of shelling out to curl,
drop unnecessary try/except around TUI confirmation,
remove tempfile (content is passed directly as bytes).
* Fix TUI imports after ui module was pulled one level up (#4515)
* Decouple share_install_log from TUI module
* Add unit tests for share_install_log function
* Update docs to use share-log subcommand syntax
* Fix truncated package metadata in additional packages preview (#3580)
* Simplify package info fix: use rstrip and CSS wrap instead of env var plumbing
* Apply preview text wrap conditionally via wrap_preview parameter
* Add missing wrap-preview CSS to OptionListScreen and pass parameter through SelectMenu
* Keep standalone initramfs for grub-btrfs when UKI is enabled
Fixes#4505
* Move keep_initramfs logic into add_bootloader()
The grub-btrfs snapshot detection was in guided.py, forcing custom
script authors to replicate it. Since Installer already holds
disk_config, the check belongs inside add_bootloader().
The cutefish package was never in the official Arch repos (AUR only)
and the upstream project is abandoned. Installing this profile would
always fail with a pacman "not found" error.
renovate[bot]
Clinton Phillips
Akilesh A S
Franco Castillo
BB
Anton Hvornum
utuhiro78
Daniel Girtler
Softer
Matteo
HADEON
stinga11
Lena Pastwa
aronmr-1
correctmost
Steen Rabol