Every Nix build of main and of each release tag now goes to
proton-drive-linux.cachix.org, so installing the flake downloads the
client instead of compiling it. The flake names the cache in nixConfig
for nix run and nix shell, and the install guide shows the nix.settings
a NixOS system build needs, since it does not read an input's
nixConfig. A fork's pull request has no token and only pulls.
A flake input pinned to vX.Y.Z is the Nix release, so the release
workflow now builds the flake at the tag and checks its version before
the GitHub release is published. The install guide shows how to pin the
input to a release tag, and the release checklist names the files that
carry the version.
The release workflow created each GitHub release without a body, so 3.1.2,
3.2.0 and 3.2.1 shipped with empty notes. It now copies the version's
section of docs/CHANGELOG.md into the release, and fails when the section
is missing.
A package upgrade replaces the binaries but does not restart the user
service, so the daemon keeps the old release and its bugs until the next
login. On 3.2.0 the reporter of #33 still saw "Couldn't sync Documents:
read .../snapshots/37: Permission denied", which only a daemon older than
3.2.0 can produce: since then the scan leaves an unreadable subfolder out
and only an unreadable sync root fails the pass (B206).
The daemon now reports its release in Status; one from before the field
omits it, which reads as older. When the daemon is older than the
front-end, the tray says "Update not finished - restart Proton Drive" and
offers a restart, the sync card shows "Restart to finish updating" with a
Restart button, and pdfs status and pdfs diagnose print the systemctl
command. A newer daemon is not flagged, since then the front-end is the
stale one.
Refs #33
Every SDK error that was not an API reply was classified as Offline.
Only a transport failure is. An upload on an account without a photo
library fails in the SDK before any request, and the app said "You're
offline" while every other call worked (issue #38).
Classify by variant: Transport is Offline, InvalidOperation is Invalid,
Serialization is Remote, Crypto and KeysLocked are Internal. The daemon
logs a failed photo upload, and the Photos page greys out Upload when
the account has no photo library.
docs/BUGS.md B205.
The CPU spin while offline, the expired session shown as offline, the
lost refresh tokens and the keyring-refused rotation were fixed without
a changelog entry.
A single timed run lost the CPU to the rest of the suite running
alongside it and read as the full-scan regression (171 ms against
701 ms). The fastest of five runs still catches the 6x scaling of the
old path.
GTK ships only a few stock icons itself. On a NixOS session without
Adwaita installed, such as Hyprland, the Photos and Computers sidebar
entries and about twenty other icons showed as missing (#37).
adwaita-icon-theme in buildInputs puts it on the wrapper's XDG_DATA_DIRS.
On NixOS the CAPTCHA window failed with "TLS support is not available":
WebKitGTK's GIO had no TLS backend. With glib-networking in buildInputs,
wrapGAppsHook4 puts its GIO modules on GIO_EXTRA_MODULES in every
wrapper.
Refs #37
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EuyxByBpK3fSWtyPkugeHz
Builds the four binaries with rustPlatform.buildRustPackage from
Cargo.lock, installs the desktop entries, icon, user service and
translations as the other packages do, and wraps the binaries with
exiftool and ffmpeg on PATH. The NixOS module installs the package,
pulls proton-drive.service into the graphical session with the
fusermount3 wrapper on its PATH, and enables FUSE and GNOME Keyring.
A Nix workflow builds the flake when its recipe, the lock files or the
code change.
Refs #37
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EuyxByBpK3fSWtyPkugeHz
A partial write whose base was not in the content cache queued an
incomplete revision, and every write open over it was refused with EIO
until the drain uploaded it. CI's one-client-lan simulation hit this on
seed 1: a create over a file whose partial edit had not drained yet.
The write open now copies the blob and fills its gaps from the remote
base with Core::fill_gaps, as the drain does, and fails only when that
fill fails (B204).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01EuyxByBpK3fSWtyPkugeHz
Issue #35: a refused session showed as offline forever, the drain
workers spun while offline, and refreshed tokens could be lost or
overwritten (B200-B203, B61). Needs proton-sdk 0.8.2.
A rotation the keyring refused was retried and kept in the kernel
keyring, but only the log said so, and a reboot before the keyring took
it still signed the user out.
`Response::Status` carries `tokens_unsaved` while one waits. The tray
says "Sign-in not saved — unlock your keyring" and needs attention, the
app's status says the same, `pdfs status` prints a Keyring line and
`pdfs diagnose` warns. New strings are translated in every catalog
(B61).
A rotation the Secret Service refused lived only in the daemon's
memory, so a crash or SIGKILL before the retry landed lost the only live
refresh token.
Keep it in the kernel keyring as well until it is written. The next
resume, by the daemon or `pdfs unlock`, puts it back in when the stored
blob still holds the refresh token it spent, and writes it; one left
over from an older login is dropped. `kernelkey` now holds named keys,
the passphrases and the unsaved tokens (B61).
`pdfs unlock` resumes the daemon's session and may refresh it. The
daemon then holds a spent refresh token, and since B202 a refusal of it
read as an expired session although the keyring held a live one.
Before calling a refused session expired, compare the keyring with what
this process last stored. When it holds a newer rotation or another
login, end the mount with `MountOutcome::ResumeStored` and mount again
from the keyring. A rotation another process already wrote counts as
this session's, so the daemon keeps writing its later ones (B203).
A failed keyring write after a refresh was only a warning, and the one
live refresh token then lived in memory only (B61). The sink now keeps
the rotation as unsaved and retries it in the background after 1 s, 5 s,
30 s and then every 5 minutes, until it is written or the blob has moved
on. The refusal is logged as an error naming `pdfs login`.
docs/BUGS.md B61.
A refresh token Proton refused, or a deleted or disabled account, was
treated as a lost connection: the mount came up "offline", the probe
retried forever and logged only at debug, and nothing told the user that
only signing in again helps (issue #35).
session_revoked() names those errors. The root fetch, probe, drain,
event poll, event seed and photos root now set session_expired, log one
warning naming `pdfs login`, and stop retrying. Cached files stay
readable and changes stay queued.
Response::Status carries session_expired (serde default false). The
tray shows "Session expired — sign in again" with a Sign In Again…
item that opens the app's sign-in page; the app's status and Sync card
say the same with a button. `pdfs status` and `pdfs diagnose` report it.
New strings are translated in every catalog.
docs/BUGS.md B202.
Proton refresh tokens are single-use. store_tokens replaced the tokens of
whatever blob was stored, so after `pdfs login` restarted the daemon, the
old daemon's shutdown persist wrote its spent tokens into the new
session's blob, and the new daemon failed with InvalidRefreshToken
(issue #35). `pdfs unlock` and the legacy migration had the same shape.
Each session now writes through a TokenSink that only touches a blob
with its session id that still holds the refresh token it last left
there. Shutdown only flushes a rotation the keyring refused. unlock and
migrate_legacy re-read the blob after verifying.
docs/BUGS.md B201.
Offline or paused, a drain worker claims nothing and then slept until the
earliest queued op fell due. Every queued op was already due and nothing
attempted it, so the wait was zero and all 16 workers spun on the queue
for the whole outage (issue #35: 17 min of CPU in 2 min 16 s).
A worker that claims nothing now sleeps the full idle poll; a reconnect
or resume wakes it. earliest_due_at also uses the claim's exclusions, so
an op held behind a claimed op on the same node no longer counts as due.
docs/BUGS.md B200.
A popover is sized and placed again when its parent's layout manager
allocates it. Context menus were parented to the outermost list view,
and the keyboard menu to the ScrolledWindow; neither has a layout
manager, so a menu kept the size it was first given and often opened
too short, with a scrollbar.
The host is now the first widget from the list view up that has a
layout manager, which still keeps menus out of list items and the theme
rules that style their insides.
Send Invitations and Create Public Link sat alone in boxed-list rows,
which read as broken settings. They now sit under their section at its
right edge. The role picker moves into the address row it applies to,
people get avatars, a member needs no subtitle, and invitations say they
are pending. The People and Public link sections show a loading row
until the share is read, and one line instead of nothing when it can't
be. The page is an AdwPreferencesPage, for the standard spacing.
Translated in every catalog.
The sidebar header had a hamburger beside the brand, and the account row
in the footer had its own ⋮ menu; both offered Preferences. The header
button sat next to the page's own tools and read as one of them.
The account row is now the main menu's button, as in Proton's own apps:
Preferences, Keyboard Shortcuts and About, then Sign Out. The whole row
opens it, upwards, and F10 still does as the primary menu.
Stop Sharing removes the link, invitations and members but cannot delete
the share itself, and Drive keeps listing the empty share until its
backend removes it, minutes later.
The daemon now leaves a share with no member, invitation or link out of
the listing (unless a lookup failed, so a live share is never hidden),
and the app drops stopped items from the page as soon as the requests
succeed instead of waiting for the slow reload.
Every mirror pass, the two-minute poll included, reported a "Checking
{folder}" job, and the tray took any job for syncing, so its icon spun
every two minutes over folders with nothing to do.
JobItem gains a background flag (absent from older daemons, so false)
for housekeeping that changes nothing in the user's files: a pass still
scanning, the home index rebuild and the cache integrity check. The tray
keeps them in its menu line but only turns busy for transfers or a job
that is not background; a pass that queues work becomes a Syncing job
and still spins the icon.
The watcher also reconciles only the folder an event lies in, instead of
walking every mirror folder over one edit.
The current folder in the breadcrumb trail was a bare label, so the "›"
before it hugged it while the ones between buttons sat in their padding.
One helper now builds the separator and the current segment for My Files,
Shared with me and Computers, with a flat button's side padding.
Grid tile names broke mid-word with a hyphen that is not in the name
("SteamSetup.e-xe"); turn Pango's inserted hyphens off.
The timing ratio between a near-empty and a 2500-entry index measured the
CI runner's load and flaked on main. Plant an oversized index row behind
the running total instead: any budget check that consults the index would
evict and reseed, so the test fails without depending on wall-clock time.
An unreadable subfolder, such as a container's root-owned cache, failed
every pass of its synced folder, so nothing else in the tree reached
Drive (#33, B196). The scan now leaves such a path out, together with an
entry that vanished mid-scan and a name that is not UTF-8, and the pass
drops everything under it from classification, the wipe guard and the
push pass's deletions. It is never read as deleted and never downloaded
into. An unreadable or absent root still fails the pass. The activity
feed names unreadable paths once each time the set changes.
Two related bugs on the way to online-only:
- The push pass read deletions from the raw baseline, so a synced path
that later became ignored was trashed on Drive (B197).
- The switch evicted everything once the folder was idle, including
ignored paths such as .git/, symlinks, unreadable folders and files
changed after the pass, none of which are on Drive (B198). It now
refuses and names the path.
Closes#33
The verification window now runs in its own process, so the app survives
when WebKit cannot start its sandbox on secureblue, but the CAPTCHA still
could not be solved there (#29). WebKit binds the journal socket with
`--ro-bind-try`, bubblewrap only skips a missing path, not an unreadable
one, and WebKit has no switch to leave the socket out.
Proton Mail Bridge never uses a web view: it opens verify.proton.me in the
system browser and retries with the original challenge token, typed with
the offered methods. Do the same as a fallback.
- core: `browser_verification_url` and `browser_verified` build the page
URL and the retry credential from the challenge
- gui: when the verification window fails, a dialog opens the page in the
browser and continues the sign-in once the user is done
- cli: `pdfs login` handles a CAPTCHA the same way instead of failing
- i18n: translate the new dialog in every catalog
- docs: B195, TROUBLESHOOTING.md, ARCHITECTURE.md, changelog
Fixes#29
A user switching from a Windows PC to Linux had two half answers:
adopting the old device makes this machine back up as it, and restoring
its folders keeps them in the old device's tree. Neither puts the data
under this machine's own device.
Device roots are separate shares on the main volume, so each top-level
folder of the other device can move under this device's root in one
server-side move. Nothing is uploaded again, and the other device stays
registered, without those folders.
- core: `Request::MigrateDevice { uid }` and
`Db::sync_folder_set_share`
- fuse: `Core::migrate_device` moves every folder, suffixes a name
already taken here with the other device's name, keeps going past a
failed folder, and re-points sync folders that already tracked a
moved folder at the new share
- cli: `pdfs devices migrate UID [--yes]` confirms, moves, then runs
the ordinary restore picker
- gui: "Move to This Computer…" on another computer's row, followed by
the restore picker
- i18n: translate the new strings in every catalog
- docs: CLI.md, USER_GUIDE.md, RECOVERY.md, ARCHITECTURE.md, changelog
On secureblue, WebKit cannot start the sandbox for the verification page:
bubblewrap is kept from the journal socket, xdg-dbus-proxy fails, and
WebKit ends its process with g_error. The webview lived in pdfs-app, so
the whole app crashed after the sign-in (#29).
- gui: `pdfs-app --human-verification` shows the page in its own window,
reads the URL on stdin and prints the token on stdout; its logs go to
stderr
- gui: a child that dies or fails without a token ends the sign-in with a
message that the page could not be opened, instead of "cancelled"
- i18n: translate the new message in every catalog
- docs: B195 in the bug ledger, the cause and a one-time workaround in
TROUBLESHOOTING.md, the child process in ARCHITECTURE.md, changelog
On a fresh install only the daemon and the CLI created
~/.local/state/proton-drive-linux. The app starts the tray before the
first sign-in, so the tray failed to bind its single-instance socket and
exited, and the Takeout page could not remember staged archives (#29).
- gui: pdfs-tray and pdfs-app call AppDirs::ensure at start, which also
makes the directories owner-only
- docs: B194 in the bug ledger, changelog entry
nvrl
Nils Pukropp
Claude