chokidar stats are captured at poll time and may not reflect the file's
final byte length by the time vault.read() is called. The downstream
integrity check compares stat.size to content length; a mismatch causes
other LiveSync clients to reject the file as corrupted.
Fix by updating file.stat.size from the actual content in read() and
readBinary().
Co-authored-by: Joysimple <Joysimple@users.noreply.github.com>
The config mismatch dialog's defaultAction is "Dismiss" which blocks
replication. Since the daemon cannot resolve mismatches interactively,
skip the check entirely and accept the remote configuration as-is.
IgnoreRules (src/apps/cli/serviceModules/IgnoreRules.ts):
- Reads .livesync/ignore for user-defined glob patterns
- Applies gitignore matchBase semantics: patterns without / get **/ prefix,
patterns ending with / get ** appended for directory contents
- Supports `import: .gitignore` directive to merge gitignore patterns
- Rejects negation patterns with a warning (not fully supportable)
- Integrated into both daemon and mirror commands via isTargetFile handler
Wiring:
- IgnoreRules loaded before LiveSyncBaseCore construction so beginWatch()
receives rules when it fires during onLoad/onFirstInitialise
- Passed through initialiseServiceModulesCLI -> StorageEventManagerCLI ->
CLIStorageEventManagerAdapter -> CLIWatchAdapter
Deployment:
- src/apps/cli/deploy/livesync-cli.service - systemd unit template
- src/apps/cli/deploy/install.sh - user/system install script
Testing:
- src/apps/cli/test/test-daemon-linux.sh - e2e tests for ignore rules
- src/apps/cli/serviceModules/IgnoreRules.unit.spec.ts - 15 unit tests
- src/apps/cli/commands/daemonCommand.unit.spec.ts - 7 unit tests
- Add chokidar ^4.0.0 as dependency (root package.json, runtime-package.json)
- Mark chokidar as external in vite.config.ts (not bundled, loaded at runtime)
- Implement CLIWatchAdapter.beginWatch() with chokidar:
- ignoreInitial: true (startup files handled by mirror scan)
- awaitWriteFinish to prevent partial-write events
- Excludes dotfiles and .livesync/ directory at watcher level
- Maps add/change/unlink/addDir/unlinkDir to IStorageEventWatchHandlers
- Fatal error handler: logs clearly and releases watcher resources
- Add close() to CLIWatchAdapter, StorageEventManagerCLI for clean shutdown
- Register onUnload hook in CLIServiceModules to close watcher on shutdown
- Add daemon command to help text and --interval/-i flag for polling mode
- Capture original sync settings before suspendAllSync() clobbers them
- Implement daemon startup: mirror scan → restore settings → applySettings()
which triggers the full suspend/resume lifecycle and starts the _changes feed
- Guard processSynchroniseResult no-op to non-daemon commands so default
handler writes incoming CouchDB changes to the local filesystem
- Polling mode: restore settings + clearInterval-safe try/catch error handling
- Warn when both liveSync and syncOnStart are false after restore (no-op config)
- Fix: only block indefinitely if daemon startup succeeded
Add prev/next buttons to jump between diff blocks in the
Document History view. Includes position indicator and
auto-scroll with visual focus highlighting.
- Now larger settings can be exported and imported via QR code without issues. (#595)
- Fixed some errors during serialisation and deserialisation of the settings, which caused issues in some cases when importing/exporting settings via QR code.
Co-authored-by: Copilot <copilot@github.com>
- `ls` and `mirror` commands now provide informative feedback when no documents are found or filters skip all files, resolving the issue where they would exit silently (#860).
- The command-line argument `vault` has been renamed to a more appropriate name, `databaseDir`.
- The `mirror` command now accepts a `vault` directory, which specifies the location where the actual files are stored. For compatibility reasons, the previous behaviour is still supported.
Co-authored-by: Copilot <copilot@github.com>
When couchDB_URI ends with a trailing slash (e.g. https://host/), the
database name concatenation produces a double-slash path
(https://host//obsidiannotes), which causes CouchDB to reject requests
with 401 "Name or password is incorrect".
Strip trailing slashes from couchDB_URI / baseUri at the path
concatenation sites in:
- src/common/utils.ts (_requestToCouchDBFetch, _requestToCouchDB)
- src/features/LocalDatabaseMainte/CmdLocalDatabaseMainte.ts
The companion fix for the replication path is in the livesync-commonlib
submodule.
Ref: #859
- Improved background worker crash cleanup so pending split/encryption tasks are released cleanly instead of being left in a waiting state (related: #855).
- On start-up, the selected remote configuration is now applied to runtime connection fields as well, reducing intermittent authentication failures caused by stale runtime settings (related: #855).
Co-authored-by: Copilot <copilot@github.com>
- Remote configuration URIs are now correctly encrypted when saved after editing in the settings dialogue.
- Fixed an issue where devices could no longer upload after another device performed 'Fresh Start Wipe' and 'Overwrite remote' in Object Storage mode (#848).
- Now surely remote configurations are editable in the settings dialogue.
- We can fetch remote settings from the remote and apply them to the local settings for each remote configuration entry.
- No longer layout breaking occurs when the description of a remote configuration entry is too long.
- No unexpected error (about a replicator) during early stage of initialisation.
### New features
- Now we can configure multiple Remote Databases of the same type, e.g, multiple CouchDBs or S3 remotes.
- We can switch between multiple Remote Databases in the settings dialogue.
- Add inline TODO comment in runCommand.ts about standardising
replication failure cause identification logic.
- Add test-sync-locked-remote-linux.sh that verifies:
1. sync succeeds when the remote milestone is not locked.
2. sync fails with an actionable error when the remote milestone
has locked=true and accepted_nodes is empty.
When the remote database is locked and the CLI device is not in the
accepted_nodes list, openReplication returns false with no CLI-specific
guidance. The existing log message ('Fetch rebuilt DB, explicit
unlocking or chunk clean-up is required') is aimed at the Obsidian
plugin UI.
Check the replicator's remoteLockedAndDeviceNotAccepted flag after
sync failure and print a clear message directing the user to unlock
from the Obsidian plugin.
Ref: #832
Node.js v25 provides a built-in localStorage on globalThis, but without
`--localstorage-file` it is an empty object lacking getItem/setItem.
The existing check `!("localStorage" in globalThis)` passes, so the
polyfill is skipped and the CLI crashes with:
TypeError: localStorage.getItem is not a function
Check for getItem as well so the polyfill is applied when the native
implementation is incomplete.
