The folder automation re-admitted a file every minute when the user had
removed its jobs while a hoster was still missing, because nothing
recorded the removal. Removing automatic jobs (selected, all, per hoster
or through MCP) now stores the file path and identity in
automation-dismissals.json. The automation skips such files until the
file changes or the user adds it manually again.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Electron 45 deprecates the synchronous safeStorage methods and Electron 46
removes them. Credential encryption now uses encryptStringAsync,
decryptStringAsync and isAsyncEncryptionAvailable.
- secret-store: async encryptField/decryptField/encryptCredentials/
decryptCredentials, decryptFieldWithStatus exposes shouldReEncrypt.
Unavailable or failing encryption still rejects with the existing
SECRET_STORE_* codes; legacy plaintext stays readable.
- config-store: the first load runs asynchronously through init();
writes encrypt before the atomic write. A synchronous load() of an
encrypted config before init() fails with SECRET_STORE_NOT_READY and
schedules a refresh. Values flagged with shouldReEncrypt are
re-encrypted once after loading.
- main: await configStore.init() after app ready, before the first load.
- MCP access store and online backup keyring await the async helpers;
MCP saves keep their call order.
Existing enc:v1 values written with encryptString stay readable: on
Windows both APIs use the same v10 format and Local State key, verified
in both directions with real Electron 41 and 45.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Clipboard writes are awaited because clipboard.writeText returns a Promise
since Electron 44. Preloads only load electron, so the removed sandbox shims
in Electron 45 do not apply; the synchronous safeStorage API is deprecated
but still available.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
With full access the installation grants every capability, destructive and
conditionally confirmed actions (queue retry/remove with source deletion,
history clear, app restart, ...) run without the local dialog, and the new
mhu_files_grant tool can add or widen directory grants. Narrowing or removing
grants and changing permissions stay local. Enabling full access in the
settings asks for local confirmation once.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Previously sources were only removed when the whole upload batch ended. In
automation mode a batch keeps growing for hours, so hundreds of finished
files stayed on disk. A file is now removed once every required hoster
succeeded, its links are flushed to the upload log, the automation ledger
is written and the renderer confirmed the queue save. Without that evidence
the file still waits for the batch end as before.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The shared upload limiter started with a full second of tokens and refilled
up to one second during short idle gaps, so new uploads could briefly send
almost twice the limit. In a parallel-upload test it released 3.87 MB in one
second at a 2 MB/s limit. Uploads now reserve send time on one shared
schedule with at most a 200 ms burst and exact waits.
Speeds and limits used 1024-based kB and MB while being shown as kB/s and
MB/s, so a limit of 100 meant 104.9 MB/s. Limits, measured speeds, ETA and
the speed graph now use 1 MB = 1,000,000 bytes. Saved limits are converted
once so the displayed value stays the same.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
mhu_queue_resume only lifts the shared pause and continues the queue
with the settings that were active before mhu_queue_pause, so it no
longer requires the local confirmation dialog when source deletion or
shutdown-after-finish is enabled. Starting, retrying and reconciling
uploads via MCP still require confirmation in that case.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Setting UV_THREADPOOL_SIZE in main.js had no effect: Electron
initialises libuv before the main script runs, so the uploader kept the
default four threads for upload reads, logs and queue saves. Packaged
builds now relaunch once at startup with UV_THREADPOOL_SIZE=32 before
the single-instance lock. Portable and development runs are unchanged.
App info and the startup marker report the effective threadpool size.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On remote desktop sessions the web contents could appear offset with white borders until the next start. After the window is shown, compare the renderer viewport with the content size and force one relayout on a mismatch or in an RDP session. The measured values are logged.
Coalesce concurrent and repeated Doodstream login checks by credential identity, keep the challenged cookie session for OTP verification, and rate-limit explicit code resends. Prefer an existing Doodstream API key during health checks and add regression coverage for imports with duplicate accounts, concurrent checks, expired challenges, and session reuse.
Load encrypted account credentials only after Electron reaches its ready state so Windows DPAPI is available during startup. Retry transient secure-storage discovery instead of caching an unavailable result for the process lifetime. Add hidden two-process DPAPI regression coverage for the real renderer IPC path and bump the public version to 2.1.40.
Persist successful watched-file uploads in a dedicated fsync-backed ledger before exposing completion to the renderer. Match entries by normalized full path, hoster, size, and modification time so restart reconciliation skips unchanged completed files while changed files and explicit manual retries remain available.
Remove restored queue ghosts from the ledger even when history and user upload logging are unavailable. Preserve per-hoster partial completion, capture missing file metadata asynchronously, fail closed on corrupted or unwritable evidence, and keep local persistence failures outside automatic upload retries.
Stream managed upload logs with bounded lines, bytes, files, directories, and result counts. Include numbered rotations, reject unconfirmed rows, share concurrent scans through a generation-safe cache, invalidate after successful appends, avoid synchronous configuration and directory reads, and close streams on every path.
Discover base, daily, session, and fallback upload logs across configured, Desktop, and application-data locations without blocking the Electron main process.
Treat in-session completion keys as authoritative automation evidence so removing finished rows cannot cause periodic reconciliation to enqueue and upload the same file again.
Add real handler and done-remove-reconcile regressions covering custom extensions, mixed-case log names, fallback paths, stale evidence, and asynchronous log reads.
Move atomic configuration writes off the Electron main event loop and preserve concurrent queue and folder monitor state.
Make finish-and-pause interrupt queued admission waits without aborting active uploads, then resume only persistently marked automation jobs across active, idle, restart, and finalization races.
Reuse bounded automation evidence snapshots, refresh them during long drains, and cover high-load persistence, pause, resume, and hidden Electron behavior with regression tests.
Bind deferred startup reconciliation to an explicit FolderMonitor lifecycle generation and invalidate pending work on stop, pause, resume, import replacement, shutdown, and every non-deferred restart. This prevents an old startup request from surviving into a later productive monitor lifecycle.
Challenge each renderer document with a monotonic generation before accepting readiness, reject stale or foreign acknowledgements, and preserve exactly-once release across reloads and renderer initialization recovery. Add dynamic hidden Electron and lifecycle harness coverage for stale readiness, restart, close, shutdown, and ARIA regressions.
Defer the productive FolderMonitor startup scan until the renderer has registered its candidate listener and acknowledged readiness. Keep the handshake idempotent across duplicate signals, reloads, window closure, shutdown, paused startup, and renderer initialization failure without delaying unrelated startup work.
Preserve the real over-limit queue count while clamping finite progressbar ARIA values, and cover both regressions with hidden Electron and preload ordering tests.
Keep full folder scans recoverable under duplicate protection and classify every discovered file once before automatic admission.
Separate manual previews from automatic capacity limits, serialize renderer intake, normalize automation counters and intervals, and make Main authoritative for runtime timestamps.
Enforce persistent pause across startup, import, close recovery, resume failures, and reconciliation while preserving read-only paused scans and exactly one activation reconciliation.
Restore failed asynchronous rotation chunks, complete automation localization and unlimited queue accessibility, and tighten hidden integration cleanup coverage.
Route synchronous account-rotation shutdown flushes through the internal log writer so failures under userData/logs retry userData/internal-logs in the same order as runtime writes.
Keep buffered lines until a target accepts the full chunk, report the successful fallback as the active path, and retain the buffer when every internal target fails.
Cover the real EISDIR primary-path case, complete fallback failure, and main-process quit delegation without launching Electron.
Route upload audit and account rotation output through one internal userData log resolver with a contained fallback directory.
Report active writer paths to diagnostics and log reveal actions while preserving the configurable fileuploader.log fallback contract and leaving existing Desktop files untouched.
Cover primary paths, fallback containment, rotation, active path reporting, log opening, and upload log isolation with focused tests.