* fix(tun): restart core when tun mode is toggled to true on linux
* fix(changelog): add note for Linux TUN not taking effect immediately
---------
Co-authored-by: Tunglies <77394545+Tunglies@users.noreply.github.com>
Transitioned panic strategy from 'abort' to 'unwind' and integrated a
global panic hook into the logging framework.
This change allows the application to capture critical failure metadata—
including specific file locations, line numbers, and panic payloads—
persisting them to local logs before exit.
Note: We prioritize troubleshooting capability and long-term stability
at this stage. Reverting to 'panic = abort', `debug = false`,
`strip = true`, `remove split-debuginfo` for peak performance and
minimal binary size will only be considered once the project reaches
a mature state with near-zero community-reported crashes.
Backend:
- Move resolve_done() from sync setup() to async task after futures::join!
so Timer waits for actual init completion instead of firing immediately
- Replace std::thread::sleep(50ms) with tokio::time::sleep in async context
- Remove duplicate refresh_tray_menu in tray_init (keep post-join call only)
- Delete dead code reset_resolve_done (process restarts, static is destroyed)
- Rename create_window(is_show) → create_window(should_create) for clarity
Frontend:
- Remove duplicate verge://refresh-clash-config listener from AppDataProvider
(useLayoutEvents handles it via invalidateQueries — single consumer path)
- Stabilize useEffect deps with useRef for TQ refetch references
- Simplify AppDataProvider event listener setup (profile-changed + proxy only)
- Change TQ_MIHOMO retryDelay from fixed 2000ms to exponential backoff
(200ms → 400ms → 800ms, cap 3s) so core-dependent queries retry faster
- Expose isCoreDataPending from AppDataProvider to distinguish between
data still loading vs actual errors
- ClashModeCard: show placeholder instead of "communication error" while
core data is pending
- CurrentProxyCard: show empty space instead of "no active node" while
core data is pending
- Remove Rust-side `eval(INITIAL_LOADING_OVERLAY)` that prematurely
dismissed the overlay before React/MUI theme was ready
- Defer `window.show()` from Rust `activate_window` to an inline
`<script>` in index.html, executed after the themed overlay is in DOM
- Remove `useAppInitialization` hook (duplicate of `useLoadingOverlay`
with no themeReady gate)
- Simplify overlay to pure theme-colored background — no spinner or
loading text — so fast startup feels instant
- Simplify `hideInitialOverlay` API and reduce overlay fade to 0.2s
- Clean up unused CSS variables (spinner-track, spinner-top, etc.)
- Dispose Monaco editor instances on dialog close to prevent cycle leak
- Replace gcTime: Infinity with finite TTLs and evict orphaned subscription queryKeys
- Add missing useEffect cleanup for timers, move setTimeout out of useMemo
- Add 10s startup grace period before TUN auto-disable logic activates;
service IPC may not be ready when the frontend first queries, causing
a transient isServiceOk=false that incorrectly persists
- Replace placeholderData (which set isLoading=false with stale data)
with a proper isStartingUp guard; query now polls every 2s during
startup to catch service readiness quickly
- Add 'getSystemState' to refresh-verge-config invalidation keys to
fix key mismatch that prevented event-driven refetches from working
URLs with percent-encoded characters in credentials (e.g. %40 for @) were
being double-encoded after Url::parse() + as_str() serialization, causing
the constructed Basic Auth header to contain the wrong credentials and
resulting in 401 Unauthorized errors.
Replace swr with @tanstack/react-query v5 across all hooks, providers,
and components. Introduce singleton QueryClient, WS subscription pattern
via useQuery+useEffect, and enforce component-layer cache access contract.
Preserve SWR cache in onConnected to avoid replacing accumulated logs
with kernel buffer on reconnect. Add KeepAlive for the logs page so
its DOM stays mounted across route changes, removing the visible blank
window when navigating back.
Add leading-edge throttle to useMihomoWsSubscription, reduce SWR retry
aggressiveness, and increase WebSocket reconnect delay to prevent event
storms when switching profiles under poor network conditions.
