Commit Graph
497 Commits
Author SHA1 Message Date
bincxzandClaude Opus 4.6 fb5c4aaa58 fix(update): update latestRelease on version mismatch, surface check failures
- Replace stale latestRelease when electron-updater reports a different
  version than the cached GitHub API result
- Surface checkForUpdate() failures by setting autoDownloadStatus to
  error instead of silently dropping them

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-11 18:15:34 +08:00
bincxzandClaude Opus 4.6 c6e3989a1b fix(auto-update): restrict error broadcasts to download phase only
- Remove hasUpdate gate from ready-to-install toast so dismissing
  availability notification doesn't prevent restart prompt
- Only open releases page on platforms without auto-download
- Increase startup check delay to 8s to let electron-updater fire first

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-11 17:57:35 +08:00
bincxzandClaude Opus 4.6 ace081414f fix(review): skip redundant startup check, honor dismissed version in toasts
- Skip renderer's GitHub API startup check if electron-updater's
  auto-download has already started, preventing duplicate toast
  notifications for the same release
- Set hasUpdate in onUpdateAvailable IPC handler, checking dismissed
  version so that dismissed releases don't trigger the persistent
  "restart now" toast after auto-download completes
- Guard "ready to install" toast with hasUpdate check

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-11 17:49:25 +08:00
bincxzandClaude Opus 4.6 049a609bca fix(review): fix retry path, consolidate useUpdateCheck, show startup updates
- Fix autoDownloadStatusRef stale read during checkNow retry: eagerly
  sync the ref when resetting error->idle so checkForUpdate() fires
- Refactor SettingsApplicationTab to accept update props instead of
  creating its own useUpdateCheck instance, preventing duplicate checks
  and inconsistent state between Application and System tabs
- Show startup-detected updates (hasUpdate) in System tab, not only
  manualCheckStatus=available, so Linux/unsupported platforms see the
  update and manual download button

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-11 17:43:34 +08:00
bincxzandClaude Opus 4.6 44409e6d32 fix(review): hydrate update status for late-opening windows, fix toast race
- Add getUpdateStatus IPC handler so windows opened after download started
  can immediately reflect the current state instead of showing stale 'idle'
- Track _lastStatus in main process across all updater events
- Hydrate autoDownloadStatus on useUpdateCheck mount via getUpdateStatus()
- Fix toast race: use ref to track previous autoDownloadStatus so ready/error
  toasts only fire on actual status transitions, not when unrelated callback
  references change

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-11 17:37:15 +08:00
penguinway c7d2482996 fix(update): classify checkNow error when result.error is set
performCheck returns a non-null UpdateCheckResult with error populated
on GitHub API/network failures. Extend the status derivation to treat
result.error as an error state instead of falling through to up-to-date.
2026-03-11 17:08:31 +08:00
penguinway d2391f5472 fix(update): restore checkNow return type and add error state retry
P1: change checkNow return type from Promise<null> to Promise<UpdateCheckResult | null>
and return actual result so callers can read hasUpdate/latestRelease.

P2: reset autoDownloadStatus from 'error' to 'idle' when user triggers manual
check, enabling a retry path; also show Check for Updates button in error state.
2026-03-11 16:58:21 +08:00
penguinway 9be84c71f5 feat(auto-update): improve UX — auto-reset badge, trigger download, show last checked time
- Fix 1: when manual check finds update and electron-updater hasn't started
  downloading yet (autoDownloadStatus=idle), fire-and-forget checkForUpdate()
  to kick off the auto-download pipeline without blocking the UI

- Fix 2: manualCheckStatus='up-to-date' now auto-resets to 'idle' after 5s
  so the badge doesn't stay stale until the next check; any new check cancels
  the pending timer first

- Fix 3: SettingsSystemTab shows "last checked: X min ago" below the update
  section using lastCheckedAt from updateState; new i18n keys added for both
  zh-CN and en locales (lastCheckedJustNow, lastCheckedMinutesAgo,
  lastCheckedHoursAgo, lastCheckedPrefix)

Internal: add autoDownloadStatusRef and manualCheckResetTimeoutRef to
useUpdateCheck for reliable cross-closure state access and timer lifecycle.
2026-03-11 16:08:32 +08:00
陈大猫andClaude Opus 4.6 77fd7a42a8 fix(sftp): drag-upload goes to wrong directory after navigation (#311)
* fix(sftp): update currentPath immediately on navigation to prevent stale upload target

When navigating directories without a cache hit, currentPath was only
updated after the async file listing completed. If a drag-and-drop upload
occurred during or shortly after this window, getActivePane would return
the old currentPath, causing files to upload to the previous directory.

Now currentPath is updated immediately when loading begins, ensuring
upload operations always target the correct directory.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): revert currentPath to previous value when navigation fails

Address review feedback: if the directory listing throws a non-session
error, restore currentPath to its previous value so later operations
(e.g. uploads) don't target a path that was never successfully loaded.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): clear files when entering loading state to prevent stale interactions

Address P1 review: the loading overlay is pointer-events-none, so users
could still interact with old files during navigation. Since currentPath
is now updated immediately, actions like delete/rename would resolve
against the new path but display old files. Clear files and selection
when loading begins to eliminate this inconsistency.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): restore previous files when reverting path on navigation error

Address P2 review: since files are now cleared when loading begins,
a failed navigation would leave the pane with the old path but an
empty file list. Save and restore the previous files alongside the
previous path in the error handler.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): restore selected files when reverting on navigation error

Address P2 review: save and restore selectedFiles alongside path and
files in the error handler so users don't lose their selection when
a navigation attempt fails.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): restore tab state when navigation is superseded by another request

Address P1 review: navSeqRef is tracked per-side not per-tab, so a
navigation from a different tab on the same side can invalidate this
request. When the sequence check causes an early return, restore this
tab's previous path, files, and selection instead of leaving it with
cleared files and a stale loading state.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): avoid overwriting newer navigation state when superseded

When a navigation request is superseded by a newer one on the same tab
(e.g., fast A→B→C), the completing request should not blindly restore
its previous state, as that would overwrite the latest navigation's
optimistic update. Now we check if the tab's current path still matches
what this request set before restoring.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): use per-tab request ID to guard superseded navigation restores

Replace the ambiguous currentPath equality check with a per-tab
navigation request ID (tabNavSeqRef). The old check failed when
refresh() triggered a navigation to the same path — the stale request
would incorrectly match and restore previous state.

The new approach tracks the latest requestId per tab, so:
- Same-tab superseded navigations (including same-path refreshes)
  correctly skip the restore.
- Cross-tab superseded navigations (different tab on the same side)
  correctly restore the orphaned tab's state.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): track per-tab nav sequence to prevent cache-hit state overwrite

When a cache-miss request (A) is pending and a cache-hit request (B) runs
on the same tab, A's superseded handler could overwrite B's result because
it only checked path equality. Add tabNavSeqRef to track the latest
requestId per tab, so superseded requests correctly skip restore when
a newer navigation (including cache hits) has already handled the tab.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix: remove leftover merge conflict markers

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): restore to last confirmed state instead of optimistic state

When multiple navigations are in flight (A→B→C), the second navigation
would snapshot the optimistic state (path=B, files=[]) as its "previous"
state. If it then failed or was superseded, it would restore to an empty
file list instead of the last successfully loaded directory.

Introduce lastConfirmedRef to track the last known-good state per tab,
updated only on successful navigation (cache hit or listing success).
Restore-on-error and restore-on-supersede now always revert to this
confirmed state.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): guard restores against stale connection after reconnect/disconnect

connect() and disconnect() reuse the same tab ID but bump navSeqRef
without updating tabNavSeqRef, so a pending navigation could restore
stale state from a previous host into a freshly reconnected tab.

Fix by:
- Capturing connectionId at navigation start and checking it in every
  updateTab restore callback (prev.connection?.id !== connectionId)
- Storing connectionId in lastConfirmedRef and re-seeding confirmed
  state when the connection changes, preventing old host data from
  being used as the restore target

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): keep files visible during loading and re-seed confirmed state

Two UI regressions fixed:

1. After a file mutation (delete/create/rename), lastConfirmedRef still
   held the pre-mutation snapshot. If the subsequent refresh failed, the
   error handler would restore stale files (e.g. resurrecting deleted
   items). Fix: re-seed confirmed state from the pane whenever it is
   settled (not loading), capturing any optimistic mutation updates.

2. Clearing files to [] on navigation start left a tab blank when
   superseded by another tab navigating on the same side. Fix: keep
   existing files visible during loading — the loading overlay already
   has pointer-events-none to prevent interaction. Files are replaced
   on success or restored from lastConfirmedRef on error/supersede.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(sftp): block interaction with stale files during directory loading

The loading overlay used pointer-events-none, allowing clicks to pass
through to stale file rows underneath. Since currentPath is updated
immediately on navigation, interacting with old filenames during a slow
load would resolve paths against the new directory.

Remove pointer-events-none from the loading overlay so it properly
blocks all interaction with the stale file list while loading.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* chore: ignore .claude/ directory in eslint config

The .claude/worktrees/ directory contains full repo copies from agent
worktrees. ESLint was scanning these, causing 621 pre-existing errors
(no-undef for Node.js globals in .cjs files) that blocked npm run dev.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-11 16:05:41 +08:00
penguinway a86a5e6839 fix(auto-update): revert checkNow to use performCheck (GitHub API) instead of electron-updater IPC
checkNow was calling bridge.checkForUpdate() which invokes updater.checkForUpdates()
via IPC. startAutoCheck() in the main process already calls checkForUpdates() on a
5s timer, and if that network request is still pending, the concurrent IPC call from
checkNow hangs indefinitely, causing the UI to be stuck in "checking" state forever.

Per the original design spec, checkNow should use performCheck() (GitHub API) directly.
This is completely independent of electron-updater's internal state machine, so it
never conflicts with the background startAutoCheck(). performCheck handles isCheckingRef,
isChecking, hasUpdate, and latestRelease; checkNow only manages manualCheckStatus.
2026-03-11 15:46:05 +08:00
penguinway 4c08888b60 fix(auto-update): fix isCheckingRef conflict in checkNow fallback path and stale version closure
- Critical: In Linux fallback path, temporarily reset isCheckingRef before calling
  performCheck so its own guard can run (was silently returning null due to double-set)
- Critical: Replace updateState.currentVersion closure in checkNow with currentVersionRef
  to avoid reading stale '' version on early user click; remove from useCallback deps
- Important: Add explicit !result guard when bridge is unavailable, returning 'error'
  status instead of silently falling through to 'up-to-date'
2026-03-11 15:20:27 +08:00
penguinway 2ea4c88680 feat(auto-update): add manualCheckStatus to UpdateState, rewrite checkNow to use electron-updater IPC 2026-03-11 15:13:54 +08:00
penguinway 0883585704 feat(auto-update): add autoDownloadStatus state and IPC subscriptions to useUpdateCheck 2026-03-11 02:35:45 +08:00
bincxz f4d833497d fix: ref-count singleton effects and use only stopPortForwardByRuleId for cleanup\n\n1. Replace boolean guard flags (reconnectCancelListenerActive,\n heartbeatActive) with ref-counting. Resources are created when\n count goes 0→1 and destroyed when count goes 1→0. The previous\n boolean approach broke when React ran child effects before parent\n ones: opening the Port Forwarding page let the child register\n the listener/heartbeat, but navigating away tore them down even\n though the App instance was still mounted.\n\n2. stopAndCleanupRule now uses only stopPortForwardByRuleId (which\n sets tunnel.cancelled = true before conn.end()). The old code\n called stopPortForward(tunnelId) first, which deletes the\n main-process tunnel entry immediately — making the cancelled\n flag invisible to conn.on('close') and causing intentional\n deletions to surface as error toasts. 2026-03-10 02:50:58 +08:00
bincxz 2ee7781b82 fix: reconnect stuck state, side-effect guards, and syncWithBackend status\n\n1. scheduleReconnectIfNeeded now returns false when the\n activeConnections entry is missing (deleted by stopAndCleanupRule\n while handshake was in-flight). Previously it returned true\n but never set the timeout, leaving reconnect-enabled rules\n stuck in 'connecting' permanently.\n\n2. Module-level guards (reconnectCancelListenerActive,\n heartbeatActive) prevent duplicate initReconnectCancelListener\n and reconcile heartbeat instances. The hook mounts from both\n App.tsx and PortForwardingNew.tsx, so without guards each\n window gets double listeners and double backend polling.\n\n3. syncWithBackend now uses tunnel.status from the backend\n (connecting or active) instead of hardcoding 'active',\n matching the reconcileWithBackend fix from the previous commit. 2026-03-10 02:09:03 +08:00
bincxz 95780a29dc fix: strip lastUsedAt from sync fallback and use real tunnel status in reconciliation\n\n1. useAutoSync localStorage fallback now also strips lastUsedAt\n (alongside status/error). Without this, the hash computed\n before async init (with lastUsedAt) differs from the hash\n after init (App.tsx strips it), causing a needless sync upload\n on every launch.\n\n2. reconcileWithBackend now uses tunnel.status from the backend\n (connecting or active) instead of hardcoding 'active' when\n seeding activeConnections. This prevents falsely marking a\n handshaking tunnel as active in the renderer. 2026-03-10 02:01:05 +08:00
bincxz 060c35f66a fix: auto-sync localStorage fallback for PF rules and settled Promise in bridge\n\n1. useAutoSync buildPayload/getDataHash now fall back to localStorage\n when portForwardingRules is empty (async init not complete).\n Previously, clicking sync immediately after launch would upload\n portForwardingRules: [] and overwrite the cloud snapshot.\n\n2. SettingsSyncTab localStorage fallback strips transient per-device\n fields (status, error) before building the sync payload.\n\n3. startPortForward Promise now tracks a settled flag across all\n resolve/reject paths. conn.on('close') rejects the Promise when\n it hasn't been settled yet (tunnel killed during SSH handshake\n by stopPortForwardByRuleId), preventing callers from hanging\n indefinitely in pendingOperations. 2026-03-10 01:52:48 +08:00
bincxz ee5d3827d5 fix: reconnect cancel on clear-all, strip transient sync fields, tunnel connecting status\n\n1. importRules([]) now iterates stored rules calling\n stopAndCleanupRule() for each one, broadcasting per-rule reconnect\n cancellation to other windows. Previously only called\n stopAllPortForwards() which doesn't signal reconnect cancel.\n\n2. SettingsSyncTab localStorage fallback strips transient per-device\n fields (status, error) before feeding rules to buildSyncPayload.\n This prevents uploading stale connection state to the cloud.\n\n3. portForwardingBridge tunnel entries now track status explicitly:\n 'connecting' on early registration, 'active' after server.listen\n or forwardIn succeeds. listPortForwards and getPortForwardStatus\n report the actual status instead of hardcoding 'active'. 2026-03-10 01:42:01 +08:00
bincxz a07c644ec8 fix: add stopPortForwardByRuleId IPC and fix uninitialized diff baseline\n\nTwo issues:\n\n1. Cross-window cleanup couldn't stop tunnels still in SSH handshake\n because listPortForwards doesn't list them. New approach:\n stopPortForwardByRuleId IPC directly iterates the main process\n portForwardingTunnels map matching by rule ID in the tunnel ID\n string, catching tunnels in ANY state.\n\n - portForwardingBridge.cjs: new stopPortForwardByRuleId function\n - preload.cjs + global.d.ts: expose the new IPC\n - portForwardingService.ts: stopAndCleanupRule and\n initReconnectCancelListener now use stopPortForwardByRuleId\n instead of fragile listPortForwards + match\n\n2. importRules diff loop missed removed/changed rules in a freshly\n opened settings window where globalRules was still empty (async\n initializeStore hadn't finished). Now falls back to reading from\n localStorage as the diff baseline. 2026-03-10 01:28:06 +08:00
bincxz ab0c4ede7e fix: handle settings window initialization timing for sync and cleanup\n\nTwo race conditions in the settings window when hooks haven't finished\nasync initialization:\n\n1. clearAllLocalData calls importRules([]) but globalRules is still\n empty, so no stopAndCleanupRule calls are made. Fix: when\n importRules receives an empty array, call stopAllPortForwards()\n on the backend as a safety net.\n\n2. onBuildPayload reads portForwardingRules from hook state which\n starts as [] until initializeStore finishes. Fix: fall back to\n reading directly from localStorage when hook state is empty,\n preventing empty-array upload that would overwrite remote data. 2026-03-10 01:10:07 +08:00
bincxz 686a707fef fix: address Codex round-3 reviews (legacy payload, heartbeat, cross-window reconnect)\n\n1. Preserve local state for legacy payloads: use !== undefined\n checks instead of ?? [] so older cloud snapshots that omit\n knownHosts/portForwardingRules don't wipe local data.\n\n2. Skip connecting tunnels during heartbeat eviction: the backend\n only lists tunnels after SSH handshake completes, so slow\n connections would be falsely evicted.\n\n3. Cross-window reconnect cancellation: stopAndCleanupRule now\n broadcasts via localStorage so other windows cancel pending\n reconnect timers. initReconnectCancelListener listens for\n these events and clears timers + activeConnections entries. 2026-03-10 00:55:07 +08:00
bincxz 8a6e915dd7 fix: address Codex review P1 (stale tunnel on config change) and P2 (additive-only sync)\n\nP1: importRules now compares 6 connection-relevant fields\n(type, localPort, remoteHost, remotePort, bindAddress, hostId)\nbetween existing and incoming rules. If any differ, the old\ntunnel is torn down so the UI no longer shows 'active' for\na tunnel pointing at stale parameters.\n\nP2: applySyncPayload now uses ?? [] fallback for\nportForwardingRules and knownHosts. This ensures 'download\nand replace' truly replaces all data, even when the payload\nwas created by an older client that didn't emit these fields. 2026-03-10 00:36:49 +08:00
bincxz 474a8bae87 chore: reduce heartbeat interval from 30s to 4s 2026-03-10 00:23:55 +08:00
bincxz 6c2e902007 feat: add periodic heartbeat to reconcile port forwarding state\n\nAdd a 30-second heartbeat that queries the main process for actual\nactive tunnels and reconciles with the renderer's state. This\nprevents state drift caused by:\n- Tunnel dying without IPC notification reaching renderer\n- Status callbacks being unsubscribed after page navigation\n- Any other edge case where renderer and backend disagree\n\nChanges:\n- Add reconcileWithBackend() to portForwardingService that detects\n gone (renderer has it, backend doesn't) and appeared (backend\n has it, renderer doesn't) tunnels\n- Add 30s heartbeat useEffect in usePortForwardingState that\n auto-corrects rule statuses when drift is detected 2026-03-10 00:18:17 +08:00
bincxz 0e61262bc0 fix: stop active tunnels when rules are deleted or replaced\n\nPreviously, deleteRule() and importRules() only removed port\nforwarding rules from state/UI without stopping the backend SSH\ntunnels. This left orphaned tunnels listening on ports with no\nUI control to shut them down.\n\nChanges:\n- Add stopAndCleanupRule() to portForwardingService for fire-and-\n forget tunnel teardown (clears reconnect timers, unsubscribes\n status events, sends IPC stop to main process)\n- deleteRule() now calls stopAndCleanupRule() before removing\n- importRules() now diffs old vs new rule IDs and stops tunnels\n for any rules being removed (covers cloud sync download and\n Clear Local Data scenarios) 2026-03-10 00:12:18 +08:00
bincxz a7873fc457 fix: unify sync payload build/apply logic to prevent data loss\n\nThe settings window was building sync payloads with customGroups\nhardcoded to [] and missing portForwardingRules/knownHosts entirely.\nThis caused data loss when syncing from the settings window.\n\nChanges:\n- Add domain/syncPayload.ts with buildSyncPayload/applySyncPayload\n pure functions as the single source of truth\n- Update App.tsx to use applySyncPayload instead of inline logic\n- Rewrite SettingsSyncTab.tsx to use unified domain functions\n- Wire portForwardingRules through SettingsPage.tsx to the sync tab\n- Fix useAutoSync getDataHash to include customGroups and knownHosts\n so their changes trigger auto-sync 2026-03-09 23:40:07 +08:00
陈大猫 1286975a4b fix: improve URL highlighting precision (#302)
* fix: improve URL highlighting precision

* fix: tighten ipv4 highlight boundaries

* fix: narrow version prefix exclusion

* fix: trim trailing URL delimiters

* fix: preserve bracketed ipv6 urls
2026-03-09 23:07:10 +08:00
陈大猫 2933e108bc feat: support system theme auto-switching (#301)
* feat: support system theme auto-switching\n\nAdd 'system' as a third theme option alongside 'light' and 'dark'.\nWhen set to 'system', the UI theme automatically follows the OS\ncolor scheme preference and switches in real-time when the system\nappearance changes.\n\nChanges:\n- useSettingsState.ts: Add resolvedTheme state derived from\n  matchMedia('prefers-color-scheme: dark'), add listener for\n  system preference changes, update applyThemeTokens to use\n  resolvedTheme instead of theme directly\n- SettingsAppearanceTab.tsx: Replace dark mode Toggle with\n  3-segment selector (Light / System / Dark) using Sun/Monitor/Moon\n  icons\n- en.ts/zh-CN.ts: Replace darkMode i18n keys with new theme keys\n  including 'system' option\n- Default theme changed from 'light' to 'system' for new users\n\nPartially addresses #294

* fix: derive resolvedTheme synchronously and guard matchMedia\n\nAddress Codex review feedback:\n1. Replace resolvedTheme useState+useEffect with synchronous\n   derivation from systemPreference state. This eliminates the\n   one-frame stale render where useLayoutEffect could apply\n   tokens from the old palette before useEffect updated\n   resolvedTheme.\n2. Add window.matchMedia guard in the system preference listener\n   to prevent crashes in jsdom tests or constrained webviews.\n3. Make the matchMedia listener unconditional (always tracks OS\n   preference) to avoid setup/teardown churn when toggling modes.

* fix: resolve 'system' theme in pre-hydration bootstrap to prevent flash

The index.html bootstrap script only handled 'dark'/'light' stored
values. Since DEFAULT_THEME is now 'system', new users (or users who
chose system mode) would get a wrong-theme first paint until React
mounted. Now resolve 'system' via matchMedia('(prefers-color-scheme:
dark)') before applying the CSS class, eliminating the visible flash.

Also use the resolved theme (not raw stored value) for accent foreground
calculation to ensure correct contrast on first paint.

Addresses Codex review on PR #301.

* fix: use resolvedTheme for top-bar toggle to avoid no-op in system mode

When theme preference is 'system' and the OS is dark, the toggle button
showed a moon icon and clicking it just switched from 'system' to 'dark'
— visually a no-op. Now we:

1. Pass resolvedTheme (always 'light'|'dark') to TopTabs for icon display
2. Toggle based on resolvedTheme so the first click always produces a
   visible change (e.g. system+dark → light, system+light → dark)

Addresses Codex review on PR #301.
2026-03-09 21:49:00 +08:00
陈大猫 8d014193ca Remove dead code and unused components (#288) 2026-03-08 10:55:17 +08:00
f3676734a7 feat(sftp): show download progress for "Open With" temp file downloads (#283)
* feat(sftp): show download progress for "Open With" temp file downloads

When opening remote files via "Open With" or double-click, the download
to a temp directory now displays real-time progress (bar, speed, ETA) in
the transfer overlay instead of silently blocking until completion.

Reuses the existing transferBridge infrastructure (fastGet with throttled
IPC progress events) and the SftpTransferItem UI. Cancellation is handled
gracefully — the task transitions to "cancelled" status, the partial temp
file is cleaned up, and the file is not opened in the external application.
The original downloadSftpToTemp path is preserved as a fallback for
contexts without a transfer queue.

* fix(sftp): harden temp download transfer state

---------

Co-authored-by: midasgao <midasgao@distinctclinic.com>
Co-authored-by: bincxz <16399091+binaricat@users.noreply.github.com>
2026-03-07 10:14:30 +08:00
陈大猫 35f531bb55 Fix SFTP folder copy into symlinked directories (#282)
* Fix SFTP directory copy into symlinked folders

* Honor SFTP directory drop targets

* Limit SFTP drop targeting to symlink directories

* Bind SFTP drops to the visible target pane

* Revert "Bind SFTP drops to the visible target pane"

This reverts commit d1bad223ff.

* Revert "Limit SFTP drop targeting to symlink directories"

This reverts commit edc67ed4a2.

* Revert "Honor SFTP directory drop targets"

This reverts commit fed0d7bdd0.

* Stabilize SFTP directory transfer progress

* Enable compressed uploads in SFTP view

* Fix directory transfer cancellation and total growth

* Keep prescan cancellation in transfer cleanup

* Sync compressed uploads and persistent cancellation

* Tighten SFTP cancellation cleanup

* Handle Windows SFTP directory paths
2026-03-06 17:07:18 +08:00
陈大猫 71ff9953bd Fix issue #278 identity refresh and session log autosave (#281)
* Fix issue #278 identity refresh and session log autosave

* Sync session log settings across windows
2026-03-06 15:12:38 +08:00
bincxzandClaude Opus 4.6 7342b4a872 feat: add terminal encoding switcher for SSH sessions (#261)
Allow users to switch between UTF-8 and GB18030 encoding mid-session
via a toolbar popover, fixing garbled output when viewing mixed-encoding
logs on remote servers.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-05 01:17:05 +08:00
bincxz 0a43220057 Merge remote-tracking branch 'origin/main' into fix/sftp-stale-channel-recovery
# Conflicts:
#	components/sftp-modal/hooks/useSftpModalSession.ts
#	electron/bridges/transferBridge.cjs
2026-03-04 22:47:05 +08:00
bincxz 9ca7e39748 chore(sftp): remove dead isFatalUploadError function
The function was exported but never imported anywhere in the codebase.
2026-03-04 22:13:07 +08:00
bincxz cf352502f8 fix(sftp): deep review fixes for channel recovery
- Fix per-client dedup: store _reopeningPromise on client object
  instead of module-level global to prevent cross-session confusion
- Narrow isSessionError patterns: replace overly broad "not found"
  and "closed" with specific "channel closed"/"connection closed",
  add "timed out" for channel open timeout errors
- Prevent channel leak on timeout: close orphaned SFTP channel
  when tryOpenSftpChannel callback fires after timeout
- Auto-reload directory listing after successful reconnect in
  SFTP modal to avoid stale UI state
2026-03-04 22:07:51 +08:00
bincxz 72d270580f fix(sftp): harden channel recovery across all operations
P1 fixes:
- Add requireSftpChannel() to all SFTP operations: readSftp,
  readSftpBinary, writeSftp, writeSftpBinary,
  writeSftpBinaryWithProgress, renameSftp, statSftp, chmodSftp,
  and deleteSftp UTF-8 branch
- Add 10s timeout to tryOpenSftpChannel to prevent hang when
  SSH connection is half-dead

P2 fixes:
- Deduplicate concurrent getSftpChannel calls to avoid redundant
  channel re-opens
- Refactor isFatalUploadError to compose with isSessionError,
  eliminating pattern duplication and drift risk
2026-03-04 22:01:44 +08:00
bincxz f0cfcbc560 refactor(sftp): consolidate duplicate isSessionError logic
- Add "write after end" and "no response" patterns to the shared
  isSessionError() in errors.ts
- Replace inline duplicate in useSftpModalSession with an import
  of the shared function
- Remove stale isSessionError from useCallback dependency array
2026-03-04 21:53:44 +08:00
rorychou f8262a64ab fix(sftp): recover stale channel after reconnect 2026-03-04 21:37:31 +08:00
bincxz 5d75f1acd4 fix: resolve multiple SFTP bugs (#254)
- Fix new folder input not resetting after deletion (SftpPaneToolbar/View)
- Fix folder download stuck at 95% by replacing simulated progress with real child-file progress tracking (useSftpTransfers)
- Add download menu item for directories in SFTP modal context menu (SftpModalFileList)
- Implement recursive folder download in SFTP modal with real-time progress (useSftpModalTransfers, SFTPModal)
- Fix mkdirLocal EEXIST error when target path is an existing file (localFsBridge)
- Close settings window when main window is minimized to tray (windowManager)

Closes #254
2026-03-04 11:04:34 +08:00
陈大猫 10ede85ae3 feat: Custom terminal themes with .itermcolors import (#245)
* feat: implement custom terminal themes with .itermcolors import (#228)\n\n- Add customThemeStore with CRUD operations and localStorage persistence\n- Create .itermcolors parser with RGB-to-hex conversion and auto theme type detection\n- Add CustomThemeEditor component with inline color pickers\n- Refactor ThemeCustomizeModal with Custom tab for create/import/edit/delete\n- Update all theme consumers (Terminal, TerminalLayer, LogView, ThemeSelectPanel,\n  SettingsTerminalTab, useSettingsState) to resolve custom themes\n- Add i18n keys for custom theme features (en + zh-CN)\n- Add isCustom flag to TerminalTheme model and STORAGE_KEY_CUSTOM_THEMES constant

* feat: add import .itermcolors button to Settings Terminal tab\n\nAllows importing .itermcolors files directly from Settings → Terminal\nwithout opening the theme modal first.

* fix: move delete button from editor to modal footer for clean layout\n\nThe delete button was rendering inside the CustomThemeEditor left panel,\ncausing misalignment with the full-width Cancel/Save footer. Now the\nfooter shows: [Delete] (left) | [Cancel] [Save] (right) when editing\nan existing custom theme.

* refactor: extract custom theme editor into standalone modal\n\nThe inline CustomThemeEditor was causing layout conflicts in the\nThemeCustomizeModal (editor + footer overlapping). Extracted into\na dedicated CustomThemeModal with:\n- Two-column layout: editor panel (left) + terminal preview (right)\n- Own footer: Delete (left) | Cancel + Save (right)\n- z-index 300 layering above the main theme modal\n- Proper scroll containment for the color editor

* fix: correct z-index stacking for custom theme modal\n\nRemoved inline style zIndex: 99999 from ThemeCustomizeModal that was\npushing it above CustomThemeModal. Now uses Tailwind z-[200] for the\nmain modal and z-[300] for the custom theme editor modal.

* feat: add new custom theme button to settings terminal tab\n\nReuses CustomThemeModal from the settings page. Creates a new theme\nbased on the currently selected theme, opens the editor modal, and\nautomatically selects the new theme on save.

* feat: add cross-window IPC sync for custom themes\n\nCustom themes created/imported/deleted in the Settings window are now\nimmediately synced to the main window (and vice versa) using the\nexisting netcatty:settings:changed IPC channel. Each mutation\nbroadcasts the change, and each window listens for incoming changes\nand reloads themes from localStorage.

* fix: show custom themes in ThemeSelectModal\n\nThemeSelectModal was only displaying built-in TERMINAL_THEMES.\nNow imports useCustomThemes hook and renders custom themes in a\nseparate section at the bottom of the theme list.

* feat: add edit/delete buttons for custom themes in settings\n\nWhen a custom theme is selected, Edit and Delete buttons appear next\nto the New/Import buttons. Edit opens the CustomThemeModal in edit\nmode, Delete removes the theme and falls back to the default theme.

* refactor: remove redundant header from CustomThemeEditor\n\nThe inner header with back arrow and title was duplicating the\nparent CustomThemeModal header. Removed the header block,\nArrowLeft import, and prefixed unused props with underscore.

* fix: add missing common.edit i18n key\n\nAdded 'Edit' / '编辑' translations for the common.edit key\nthat was showing as raw key text in the Settings page.

* fix: add error feedback for .itermcolors import in settings\n\nAdded step-by-step console logging for debugging import issues.\nShows user-visible alert on parse failure with localized message.\nAlso added terminal.customTheme.importError i18n keys.

* fix: handle extra keys in .itermcolors color dicts\n\nThe parseColorDict function assumed keys[i] aligned with reals[i],\nbut .itermcolors files with extra keys like 'Alpha Component' (real)\nand 'Color Space' (string) broke the index mapping.\n\nNow iterates through dict children properly, pairing each <key>\nwith its next sibling and skipping non-<real> values.

* fix: subscribe to custom theme store for reactive re-renders\n\nReplaced imperative customThemeStore.getThemeById() calls with reactive\nuseCustomThemes() hook in useMemo dependencies across 5 files:\n- useSettingsState.ts (currentTerminalTheme)\n- Terminal.tsx (effectiveTheme for host-override)\n- TerminalLayer.tsx (composeBarThemeColors)\n- LogView.tsx (currentTheme for log replay)\n- SettingsTerminalTab.tsx (currentTheme)\n\nThis ensures editing a custom theme in-place (same ID) triggers\nre-renders in all consuming components, instead of showing stale colors\nuntil the user switches theme IDs or reloads.

* fix: theme editor hex validation, import error feedback, and Escape propagation\n\n1. ColorInput: Use local state for text field so partial hex values\n   (#1, #abc) are held locally while typing. Only complete #rgb (auto-\n   normalized to #rrggbb) or #rrggbb values are committed to the theme.\n   On blur, partial values revert to the last valid color.\n\n2. ThemeCustomizeModal handleFileSelected: Added error feedback via\n   window.alert when .itermcolors parsing fails, reusing the existing\n   terminal.customTheme.importError i18n key. Also extended filename\n   regex to strip .xml extension.\n\n3. ThemeCustomizeModal Escape handler: Skip parent modal cancelation\n   when editingTheme is active, so pressing Escape only closes the\n   child CustomThemeModal without reverting the parent dialog.

* fix: backdrop click closes CustomThemeModal + remove nested buttons in ThemeItem\n\n1. CustomThemeModal: Attach onClick={onCancel} directly to the backdrop\n   div instead of checking e.target === e.currentTarget on the container.\n   The modal content div now stops event propagation to prevent\n   accidental dismissal when clicking inside the dialog.\n\n2. ThemeItem: Replace outer <button> with <div role=\"button\"> and inner\n   edit <button> with <div role=\"button\"> to eliminate invalid nested\n   interactive elements. Added keyboard handlers (Enter/Space) for\n   accessibility parity.

* fix: restore Escape key in CustomThemeModal + stabilize store snapshots\n\n1. CustomThemeModal: Add Escape key handler (capture phase) so pressing\n   Escape dismisses the child editor. Fixes regression where parent\n   ThemeCustomizeModal skips Escape when editingTheme is active but\n   the child had no handler of its own.\n\n2. customThemeStore: Cache the merged allThemes array (built-in +\n   custom) and only rebuild it when the store is mutated. The previous\n   getAllThemes() created a new array every call, violating the\n   useSyncExternalStore contract that getSnapshot must return a stable\n   reference between mutations.

* fix: accept <integer> plist nodes and guard NaN in itermcolors parser\n\nparseColorDict now accepts both <real> (float 0.0-1.0) and <integer>\n(0-255) plist value types for RGB components. Integer values are\nnormalized by dividing by 255. Also added isNaN guard on parseFloat\nresults to prevent malformed '#NaNNaNNaN' color strings from being\npersisted as custom themes.

* fix: use customThemeStore.getThemeById in HostDetailsPanel\n\nHostDetailsPanel used TERMINAL_THEMES.find() for both SSH and Telnet\ntheme previews, which only searched built-in themes. When a custom\ntheme was selected for a host, the preview fell back to Flexoki Dark\ndefaults. Now uses customThemeStore.getThemeById() which searches\nboth built-in and custom themes.

* chore: remove fake user counts from ThemeSelectPanel\n\nRemoved Math.random() generated fake user counts for Kanagawa and\nHacker themes, 'new' badges for Flexoki themes, and the Users icon.\nOnly meaningful labels remain: 'Default' for netcatty-dark and\n'Light mode' for netcatty-light.
2026-03-03 19:27:37 +08:00
陈大猫 afcc33b7fb fix: add missing passphrase to SFTP dual-pane credentials (#238) (#239)
useSftpHostCredentials.ts omitted `passphrase` when building the
credentials object for the target host, causing SFTP connections with
passphrase-protected private keys to fail with:

  Error: Cannot parse privateKey: Encrypted private OpenSSH key
  detected, but no passphrase given

The jump host path (L50) already included passphrase correctly.
This adds the same pattern to the main host credentials.
2026-03-03 13:59:02 +08:00
陈大猫 b58120998f fix: improve SFTP transfer speed with parallel requests and accurate progress (#226) (#231)
- Replace sequential stream piping with ssh2 fastPut/fastGet (64 concurrent SFTP requests)
- Use 512KB chunk size instead of default 32KB for better throughput
- Fix speed calculation with sliding window to prevent inflated burst speeds
- Throttle progress IPC to 100ms intervals to reduce event loop contention
- Simplify frontend speed display by removing ref-based smoothing layer
- Update memo comparison for smoother progress bar re-renders
2026-03-03 11:27:42 +08:00
2215d52b09 feat: credential protection guards for enc:v1: placeholders (#212)
* feat: add credential protection guards for enc:v1: placeholders

Prevent encrypted credential placeholders from being sent as
actual passwords when safeStorage decryption is unavailable
(e.g. different device/user profile). Adds guards at terminal
connection, cloud sync, and settings boundaries with user-facing
warnings and i18n support.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix: validate base64 format in encrypted credential detection

Only treat values as encrypted placeholders when the content after
the enc:v1: prefix is valid base64. Prevents false positives if a
real password happens to start with the prefix literal.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix: resolve regressions in master-key change flow and credential placeholder detection

- Make ensureSyncablePayload non-blocking in changeMasterKey handler so
  success toast and dialog close always fire after a successful key change,
  even when the payload contains unresolved enc:v1: placeholders
- Add MIN_CIPHERTEXT_BASE64_LENGTH (32) threshold to
  isEncryptedCredentialPlaceholder to avoid false-positive matches on
  plaintext credentials that happen to start with enc:v1: (e.g. enc:v1:hello)

* fix: clean up chain-progress listener on credential reentry and gate proxy check on auth usage

- Unsubscribe onChainProgress before returning in needsCredentialReentry
  branch to prevent listener leaks across connection attempts
- Only block connection for undecryptable proxy password when proxy
  authentication is actually in use (has a username)

* fix: reduce enc:v1 placeholder false positives

* fix: require syncable payload before master key rotation

---------

Co-authored-by: rorychou <roryechou@gmail.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: bincxz <16399091+binaricat@users.noreply.github.com>
2026-02-25 15:32:24 +08:00
Rory Chou 4445bf578c security: harden external navigation / window.open (#209) 2026-02-14 16:21:11 +08:00
bincxzandClaude Opus 4.6 d267c4b6fc fix: prevent stale cross-window writes and deferred-read init races
- CloudSyncManager: bump providerWriteSeq on storage events so an
  in-flight local save is discarded when newer cross-window data arrives
- useVaultState: defer reads of keys/identities/groups/snippets to just
  before their processing stage instead of reading all upfront, so data
  updated during a prior async decrypt gap is not overwritten by a stale
  pre-await snapshot

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-13 23:20:46 +08:00
bincxzandClaude Opus 4.6 1a1da02e92 fix: guard storage decrypt callbacks against local writes and sync updates
- useVaultState: storage-event decrypt callbacks now also check
  writeVersion so a local edit during the decrypt gap causes the stale
  cross-window result to be discarded
- CloudSyncManager: bump providerDecryptSeq in uploadToProvider before
  mutating lastSync/lastSyncVersion so a pending cross-window decrypt
  cannot overwrite the newer sync metadata

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-13 22:54:28 +08:00
bincxzandClaude Opus 4.6 5e753334ed fix: capture write version before async init decryption to prevent startup race
Move hostsWriteVersion/keysWriteVersion/identitiesWriteVersion increments
to before the await decryptHosts/Keys/Identities calls, and guard both
setstate and re-encrypt with the version check. This prevents a write
that occurs during the decryption await (storage event, user edit) from
being overwritten by stale init data.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-13 22:23:28 +08:00
bincxzandClaude Opus 4.6 a488bc466b fix: invalidate in-flight async operations on local and cross-window state changes
- useVaultState: bump writeVersion counters on storage events so pending
  local encrypts are discarded when newer cross-window data arrives
- CloudSyncManager: bump providerSeq on all local provider mutations
  (connect, disconnect, status updates, save) so in-flight decrypt
  callbacks from startup or storage events cannot revert newer state

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-13 22:14:52 +08:00
bincxzandClaude Opus 4.6 2748cd5363 fix: add sequence guards to all async decrypt paths
Prevent out-of-order async decrypt results from overwriting newer state:

- useVaultState: add per-key readSeq counters for cross-window storage
  event decrypt callbacks (hosts, keys, identities)
- CloudSyncManager: add per-provider sequence counters shared between
  initProviderDecryption and cross-window storage handler, so stale
  decrypt results are discarded in both paths

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-13 22:04:59 +08:00