* fix(core): put the user's node choices back when the core starts
mihomo restores the node picked in each group from the `cache.db` it keeps
in the directory it was started against. The app relied on that entirely,
and it does not hold: `profile.store-selected` comes from a merge template
a user is free to replace, and a service older than the durable runtime
generation starts every core in a directory nothing has ever run in. Both
end the same way — every `select` group falls back to the first entry in
its `proxies:` list.
The app already knows better. It records each selection in the profile,
and `activate_selected_nodes` already reconciles that record against the
running core — it was only ever called when switching profiles, never when
a core came up. Call it from the three places a core process becomes
ready. Repeat calls supersede each other, so no coordination is needed
between the start paths.
That makes the profile authoritative, which turns any selection it does
not know about into one the next start silently undoes. The chain proxy
moved a group without recording it, so recording is now one shared hook
that both it and the proxy list go through.
Restoration is spawned, not awaited: it polls the core until its groups
are readable, and a start must not wait for that.
* fix(tray): record the node the tray switched to
Restoring the profile's selections when a core starts made the profile
authoritative, which turns any selection it never learned about into one
the next start silently undoes. The proxy list and the chain proxy both
record; the tray did not.
So switching node from the tray survived only until the next restart, and
then reverted to whatever the profile last knew — a regression the
restoration itself introduced, and only for the tray.
`record_selected_node` is the backend counterpart of the frontend's
`useRecordSelection`. Failing to record does not fail the switch: the core
is already on the new node by then, and reporting the switch as failed
would be wrong. It is logged instead.
* fix(core): restoring selections must not prune what the core has not loaded
Putting the profile's selections back when a core starts reused the
reconciliation written for profile switches, and that one also *repairs*:
a record whose group is absent from two snapshots a second apart is
dropped, and the reduced list is written to profiles.yaml.
After a profile switch the configuration is loaded, so absence is
evidence. After a core start it is not — a provider-backed group can be
empty or missing for seconds — and a record pruned on that evidence is
gone for good, because the provider finishing later cannot bring it back.
Restoring now keeps the records and only applies what it can match;
pruning stays with the caller entitled to conclude it.
The restore is also asynchronous: it reads the profile, polls the core
until its groups are readable, then puts its selections. A choice the user
makes inside that window is newer than what it captured, so letting it
finish pushed the core back to the older node while the profile already
held the newer one. Recording a selection now supersedes any activation in
flight — from the tray, and from the UI and chain proxy through
`patch_profile`.
* fix(core): activate the selections of the profile that is actually running
The activation worker read the profiles draft. A profile switch stages its
target there before validating it, so a switch that fails — leaving the
previous profile running — could still have handed an activation the
rejected profile's selections, which it would then apply to whichever
same-named groups the running profile has.
Reading the committed state removes the window. The switch that succeeds
commits before it activates, so nothing changes for that caller; what
changes is that a restore running at core start can no longer observe a
switch that never happened.
* fix(core): stop a restoration from undoing a choice made while it runs
Two gaps in making the profile authoritative, both found by adversarial
review of the commits that introduced it.
A restoration reads the profile's selections once, then polls the core
until its groups can be read — seconds, on a cold start. A selection made
inside that window is newer than what the restoration is carrying, and
nothing told the restoration so: it went on to apply the older choice over
the user's. The UI accepted the new node and traffic snapped back to the
old one, with the profile and the running core then disagreeing. Recording
a selection now abandons any restoration still in flight, through the same
generation counter every other superseding caller uses.
The other gap is the home card in global mode, which passed
`skipConfigSave` and so recorded nothing. That was harmless while nothing
re-applied the profile; now it means the next start undoes the choice. It
was the only caller that ever set the flag, so the flag is gone rather than
flipped — a parameter whose one use was a bug is not worth keeping.
* fix(core): wait for the selections that can be put back, keep trying for the rest
Two gaps in restoring the profile's selections when a core starts, both
found by adversarial review.
Restoring re-checked once, a second later, and stopped. That is enough
after a profile switch, which happens with the configuration already
loaded. A core start is the opposite case: a provider-backed group is
present but empty until its provider finishes loading, and a `select` can
fail while the core is still settling. Neither produces an activation, so
the restore reported success having applied nothing and the group stayed
on the first entry of its `proxies:` list — the outcome restoring exists
to prevent. It now keeps looking, on a deadline, at what the core is
actually on rather than at what it managed to send.
Restoring was also spawned, and the caller enables the system proxy the
moment the start returns — so traffic could be pointed at a core still on
the wrong node. The first pass is now awaited, bounded, and the rest keeps
settling in the background: a core that will not answer delays a start
rather than blocking it, because being proxied through the wrong node
beats not being proxied at all.
TUN is deliberately not covered, and cannot be by anything here: it is
configured in the core's own file, so it carries traffic from the moment
the process starts. What puts a TUN user on the right node from the first
packet is the core's own `cache.db`, which is why the service keeps an
owner's runtime generation across restarts.
`supersede_selected_activation` and `cancel_selected_node_activation` were
the same function under two names, added independently for the same
reason, and `patch_profile` called both. One survives, called once, after
the profile is written — the only placement that needs to exist, since an
activation that read the older profile is cancelled by it and one starting
afterwards reads the newer one.
* fix(core): record a selection by naming it, not by resending the list
Recording a node selection sent the whole selection list, built from the
profile the frontend had rendered. Two selections made before that list
refreshed were therefore built from the same stale snapshot, and the later
write dropped the earlier group's choice. Since a core start re-applies
whatever the profile holds, the dropped one came back on the next restart
— the profile and the running core disagreeing, in the direction the user
would least expect.
The tray already avoided this, because it went through
`record_selected_node`, which merges one group into the profile inside
`with_data_modify`. Exposing that as a command and pointing the frontend
at it leaves one way to record a selection instead of two, and removes the
list from the wire: what a caller knows is the group and the node, and
that is now all it says.
`useRecordSelection` no longer reads the profile at all, which also makes
it stable across renders rather than changing identity whenever the
profile does.
* test(core): make every way of declining to stage answer for itself
`CoreRestarted` was added to `StageRejection` upstream and reached
production without a test. It inherited the right verdict — every
`RestartRequired` is matched together — but nothing said so, and nothing
would have noticed if it had inherited the wrong one.
The list of rejections now goes through a wrapper whose `match` is
exhaustive, so the next variant added upstream fails to compile here. That
is the part worth having: it does not prove the list is complete, but it
does mean nobody can add a way to decline without being sent to this test
to say what it should lead to.
* build(deps): pin the service protocol to v2.6.1
The pin tracked `feat/stage-runtime` because the protocol change it
carried had no tag yet. It has one now, and a branch pin makes the build
irreproducible: an upstream force-push silently changes what this
compiles against.
v2.6.1 carries the runtime generation an owner keeps across restarts,
which is what lets the core hold on to the `cache.db` its node selections
and fake-ip leases live in. The protocol itself is untouched — epoch,
revision and both compatibility floors are byte-identical to v2.6.0 — so
this is a fix release in both directions: an older service still works
with this app, and this app still works with an older service, only
without the durable generation.
The lockfile change is the two lines that name it. Reaching for
`cargo update -p` instead re-resolved a subtree and quietly downgraded
`windows-sys`, `windows-core` and `getrandom` along the way.
* feat(service): isolate macOS launch and owner runtime
* ci(windows): run owner security tests
* chore(deps): use kode-bridge v0.4.1 tag
* fix(macos): classify installed service state
* fix(macos): classify marked unavailable service
* fix(macos): gate core startup on service migration
* fix(macos): scope service init to supported platforms
* feat(macos): prompt before migrating installed service
* fix(macos): shorten sidecar IPC path
* build(service): pin helper protocol and artifacts
* fix(core): require confirmed service and sidecar readiness
* fix(macos): harden launch roots and singleton state
* fix(ui): refresh migration state before restart
* fix(macos): linearize service migration lifecycle
* fix(ui): keep service migration state fresh
* feat(service): track authenticated owner sessions
* fix(service): preserve newer owner sessions
* feat(macos): route service proxy changes through helper
* fix(macos): serialize service proxy guard
* fix(macos): close service proxy owner lifecycle
* fix(macos): close remaining proxy stop races
* fix(app): do not cancel controlled service stop
* fix(lifecycle): keep controlled updates cancellation-safe
* fix(macos): stage dev core for service
* test(macos): exercise staged core cleanup
* fix(dev): await controlled macOS exit
* chore: update changelog
* fix(macos): recover sidecar after service install failure
Allow the existing sidecar to continue when service installation leaves the manager in InstallRequired, then settle the session state as SidecarAllowed. Show neutral status text while the migration dialog waits for its final state.
* fix(macos): stage dev service tools before elevation
* fix(macos): localize launch guard prompts
* fix(service): address cross-platform integration issues
* feat: support seamless service migration
* fix: pass Windows service installer path safely
* fix(prebuild): derive service release from Cargo dependency
* fix(dev): preserve service installation state
* fix: resolve clippy warnings in service lifecycle
- Replace boolean/string validation results with structured ValidationOutcome values across config validation, profile switching, DNS validation, and script validation. Callers can now distinguish valid, invalid, skipped, and busy states, with invalid results carrying an error kind and message.
- Route core config updates through structured outcomes and checked forced updates. Generation and validation failures now discard runtime drafts and surface validation failures instead of relying on empty success messages.
- Preserve existing frontend ergonomics by keeping boolean service wrappers where callers only need success, while direct DNS and script validation consumers receive ValidationOutcome. Failed profile editor saves reload the backend-restored file content and synchronize Monaco models so reopened editors show the restored state.
* refactor: replace handle::Handle::get_window() with WindowManager::get_main_window() in multiple files
* refactor: enhance WindowManager to return window state alongside the main window instance
* refactor: update useProfiles and ProfilePage to support profile overrides and improve patchProfilesConfig return type
* refactor: enhance handle_success to check main window existence before notifying profile changes
* refactor: simplify get_main_window_with_state by using pattern matching and improve window state handling
* refactor: fix window activation by removing unnecessary reference in activate_existing_main_window
* refactor: remove redundant macOS conditional for window_manager import
* perf: improve config processing
* perf: enhance profile reordering logic and adjust logging level
* perf: add PartialEq derive to PrfSelected and PrfExtra structs for improved comparison
* perf: refactor PrfOption merge logic and streamline update_item method in IProfiles
* perf: simplify current_mapping and profiles_preview methods in IProfiles for improved readability
* perf: optimize filename matching logic in IProfiles by using a static regex
refactor(merge): simplify deep_merge function signature
chore: remove unused fmt_bytes function and related tests
chore: clean up help module by removing unused macros
chore: remove format module from utils
* perf(draft): optimize memory layout by removing double indirection
- Replace `Arc<Box<T>>` with `Arc<T>` to reduce pointer chasing and memory overhead.
- Leverage `Arc::from(Box<T>)` in async modify path for efficient ownership transfer.
- Fix race conditions in `edit_draft` by ensuring atomicity under write lock.
- Performance improved by ~16-24% across all operations (based on Criterion bench).
Benchmarks:
- latest_arc: 41.1ns (-24.2%)
- edit_draft: 92.2ns (-17.6%)
- apply: 89.8ns (-17.7%)
- async_modify: 66.0ns (-16.6%)
* perf(draft): implemented optimistic locking via Arc::ptr_eq for with_data_modify
Benchmarks confirm only a negligible ~2% (1.3ns) overhead for async operations, ensuring total data integrity during concurrent updates.
* feat(logging): introduce clash-verge-logging crate for management
- Added a new crate `clash-verge-logging` with dependencies on `log`, `tokio`, `compact_str`, and `flexi_logger`.
- Implemented logging types and macros for structured logging across the application.
- Replaced existing logging imports with the new `clash_verge_logging` crate in various modules.
- Updated logging functionality to support different logging types and error handling.
- Refactored code to improve logging consistency and maintainability.
* fix(logging): update import paths for clash_verge_logging in linux.rs and dns.rs
* fix(logging): update import statement for clash_verge_logging in windows.rs
fix: improve error handling in patch_profiles_config to prevent requests during profile switching
fix: change error handling in patch_profiles_config to return false when a switch is in progress
fix: ensure CURRENT_SWITCHING_PROFILE is reset after config updates in perform_config_update and patch_profiles_config
- Updated multiple instances in the codebase to replace calls to latest_ref() with latest_arc() for improved performance and memory management.
- This change affects various modules including validate, enhance, feat (backup, clash, config, profile, proxy, window), utils (draft, i18n, init, network, resolve, server).
- Ensured that all references to configuration data are now using the new arc-based approach to enhance concurrency and reduce cloning overhead.
refactor: update imports to explicitly include ClashInfo and Config in command files
- Replaced direct log calls with a new logging macro that includes a logging type for better categorization.
- Updated logging in various modules including `merge.rs`, `mod.rs`, `tun.rs`, `clash.rs`, `profile.rs`, `proxy.rs`, `window.rs`, `lightweight.rs`, `guard.rs`, `autostart.rs`, `dirs.rs`, `dns.rs`, `scheme.rs`, `server.rs`, and `window_manager.rs`.
- Introduced logging types such as `Core`, `Network`, `ProxyMode`, `Window`, `Lightweight`, `Service`, and `File` to enhance log clarity and filtering.
* refactor: optimize item handling and improve profile management
* refactor: update IVerge references to use references instead of owned values
* refactor: update patch_verge to use data_ref for improved data handling
* refactor: move handle_copy function to improve resource initialization logic
* refactor: update profile handling to use references for improved memory efficiency
* refactor: simplify get_item method and update profile item retrieval to use string slices
* refactor: update profile validation and patching to use references for improved performance
* refactor: update profile functions to use references for improved performance and memory efficiency
* refactor: update profile patching functions to use references for improved memory efficiency
* refactor: simplify merge function in PrfOption to enhance readability
* refactor: update change_core function to accept a reference for improved memory efficiency
* refactor: update PrfItem and profile functions to use references for improved memory efficiency
* refactor: update resolve_scheme function to accept a reference for improved memory efficiency
* refactor: update resolve_scheme function to accept a string slice for improved flexibility
* refactor: simplify update_profile parameters and logic
* refactor: proxy refresh
* fix(proxy-store): properly hydrate and filter backend provider snapshots
* fix(proxy-store): add monotonic fetch guard and event bridge cleanup
* fix(proxy-store): tweak fetch sequencing guard to prevent snapshot invalidation from wiping fast responses
* docs: UPDATELOG.md
* fix(proxy-snapshot, proxy-groups): restore last-selected proxy and group info
* fix(proxy): merge static and provider entries in snapshot; fix Virtuoso viewport height
* fix(proxy-groups): restrict reduced-height viewport to chain-mode column
* refactor(profiles): introduce a state machine
* refactor:replace state machine with reducer
* refactor:introduce a profile switch worker
* refactor: hooked up a backend-driven profile switch flow
* refactor(profile-switch): serialize switches with async queue and enrich frontend events
* feat(profiles): centralize profile switching with reducer/driver queue to fix stuck UI on rapid toggles
* chore: translate comments and log messages to English to avoid encoding issues
* refactor: migrate backend queue to SwitchDriver actor
* fix(profile): unify error string types in validation helper
* refactor(profile): make switch driver fully async and handle panics safely
* refactor(cmd): move switch-validation helper into new profile_switch module
* refactor(profile): modularize switch logic into profile_switch.rs
* refactor(profile_switch): modularize switch handler
- Break monolithic switch handler into proper module hierarchy
- Move shared globals, constants, and SwitchScope guard to state.rs
- Isolate queue orchestration and async task spawning in driver.rs
- Consolidate switch pipeline and config patching in workflow.rs
- Extract request pre-checks/YAML validation into validation.rs
* refactor(profile_switch): centralize state management and add cancellation flow
- Introduced SwitchManager in state.rs to unify mutex, sequencing, and SwitchScope handling.
- Added SwitchCancellation and SwitchRequest wrappers to encapsulate cancel tokens and notifications.
- Updated driver to allocate task IDs via SwitchManager, cancel old tokens, and queue next jobs in order.
- Updated workflow to check cancellation and sequence at each phase, replacing global flags with manager APIs.
* feat(profile_switch): integrate explicit state machine for profile switching
- workflow.rs:24 now delegates each switch to SwitchStateMachine, passing an owned SwitchRequest.
Queue cancellation and state-sequence checks are centralized inside the machine instead of scattered guards.
- workflow.rs:176 replaces the old helper with `SwitchStateMachine::new(manager(), None, profiles).run().await`,
ensuring manual profile patches follow the same workflow (locking, validation, rollback) as queued switches.
- workflow.rs:180 & 275 expose `validate_profile_yaml` and `restore_previous_profile` for reuse inside the state machine.
- workflow/state_machine.rs:1 introduces a dedicated state machine module.
It manages global mutex acquisition, request/cancellation state, YAML validation, draft patching,
`CoreManager::update_config`, failure rollback, and tray/notification side-effects.
Transitions check for cancellations and stale sequences; completions release guards via `SwitchScope` drop.
* refactor(profile-switch): integrate stage-aware panic handling
- src-tauri/src/cmd/profile_switch/workflow/state_machine.rs:1
Defines SwitchStage and SwitchPanicInfo as crate-visible, wraps each transition in with_stage(...) with catch_unwind, and propagates CmdResult<bool> to distinguish validation failures from panics while keeping cancellation semantics.
- src-tauri/src/cmd/profile_switch/workflow.rs:25
Updates run_switch_job to return Result<bool, SwitchPanicInfo>, routing timeout, validation, config, and stage panic cases separately. Reuses SwitchPanicInfo for logging/UI notifications; patch_profiles_config maps state-machine panics into user-facing error strings.
- src-tauri/src/cmd/profile_switch/driver.rs:1
Adds SwitchJobOutcome to unify workflow results: normal completions carry bool, and panics propagate SwitchPanicInfo. The driver loop now logs panics explicitly and uses AssertUnwindSafe(...).catch_unwind() to guard setup-phase panics.
* refactor(profile-switch): add watchdog, heartbeat, and async timeout guards
- Introduce SwitchHeartbeat for stage tracking and timing; log stage transitions with elapsed durations.
- Add watchdog in driver to cancel stalled switches (5s heartbeat timeout).
- Wrap blocking ops (Config::apply, tray updates, profiles_save_file_safe, etc.) with time::timeout to prevent async stalls.
- Improve logs for stage transitions and watchdog timeouts to clarify cancellation points.
* refactor(profile-switch): async post-switch tasks, early lock release, and spawn_blocking for IO
* feat(profile-switch): track cleanup and coordinate pipeline
- Add explicit cleanup tracking in the driver (`cleanup_profiles` map + `CleanupDone` messages) to know when background post-switch work is still running before starting a new workflow. (driver.rs:29-50)
- Update `handle_enqueue` to detect “cleanup in progress”: same-profile retries are short-circuited; other requests collapse the pending queue, cancelling old tokens so only the latest intent survives. (driver.rs:176-247)
- Rework scheduling helpers: `start_next_job` refuses to start while cleanup is outstanding; discarded requests release cancellation tokens; cleanup completion explicitly restarts the pipeline. (driver.rs:258-442)
* feat(profile-switch): unify post-switch cleanup handling
- workflow.rs (25-427) returns `SwitchWorkflowResult` (success + CleanupHandle) or `SwitchWorkflowError`.
All failure/timeout paths stash post-switch work into a single CleanupHandle.
Cleanup helpers (`notify_profile_switch_finished` and `close_connections_after_switch`) run inside that task for proper lifetime handling.
- driver.rs (29-439) propagates CleanupHandle through `SwitchJobOutcome`, spawns a bridge to wait for completion, and blocks `start_next_job` until done.
Direct driver-side panics now schedule failure cleanup via the shared helper.
* tmp
* Revert "tmp"
This reverts commit e582cf4a65.
* refactor: queue frontend events through async dispatcher
* refactor: queue frontend switch/proxy events and throttle notices
* chore: frontend debug log
* fix: re-enable only ProfileSwitchFinished events - keep others suppressed for crash isolation
- Re-enabled only ProfileSwitchFinished events; RefreshClash, RefreshProxy, and ProfileChanged remain suppressed (they log suppression messages)
- Allows frontend to receive task completion notifications for UI feedback while crash isolation continues
- src-tauri/src/core/handle.rs now only suppresses notify_profile_changed
- Serialized emitter, frontend logging bridge, and other diagnostics unchanged
* refactor: refreshClashData
* refactor(proxy): stabilize proxy switch pipeline and rendering
- Add coalescing buffer in notification.rs to emit only the latest proxies-updated snapshot
- Replace nextTick with queueMicrotask in asyncQueue.ts for same-frame hydration
- Hide auto-generated GLOBAL snapshot and preserve optional metadata in proxy-snapshot.ts
- Introduce stable proxy rendering state in AppDataProvider (proxyTargetProfileId, proxyDisplayProfileId, isProxyRefreshPending)
- Update proxy page to fade content during refresh and overlay status banner instead of showing incomplete snapshot
* refactor(profiles): move manual activating logic to reducer for deterministic queue tracking
* refactor: replace proxy-data event bridge with pure polling and simplify proxy store
- Replaced the proxy-data event bridge with pure polling: AppDataProvider now fetches the initial snapshot and drives refreshes from the polled switchStatus, removing verge://refresh-* listeners (src/providers/app-data-provider.tsx).
- Simplified proxy-store by dropping the proxies-updated listener queue and unused payload/normalizer helpers; relies on SWR/provider fetch path + calcuProxies for live updates (src/stores/proxy-store.ts).
- Trimmed layout-level event wiring to keep only notice/show/hide subscriptions, removing obsolete refresh listeners (src/pages/_layout/useLayoutEvents.ts).
* refactor(proxy): streamline proxies-updated handling and store event flow
- AppDataProvider now treats `proxies-updated` as the fast path: the listener
calls `applyLiveProxyPayload` immediately and schedules only a single fallback
`fetchLiveProxies` ~600 ms later (replacing the old 0/250/1000/2000 cascade).
Expensive provider/rule refreshes run in parallel via `Promise.allSettled`, and
the multi-stage queue on profile updates completion was removed
(src/providers/app-data-provider.tsx).
- Rebuilt proxy-store to support the event flow: restored `setLive`, provider
normalization, and an animation-frame + async queue that applies payloads without
blocking. Exposed `applyLiveProxyPayload` so providers can push events directly
into the store (src/stores/proxy-store.ts).
* refactor: switch delay
* refactor(app-data-provider): trigger getProfileSwitchStatus revalidation on profile-switch-finished
- AppDataProvider now listens to `profile-switch-finished` and calls `mutate("getProfileSwitchStatus")` to immediately update state and unlock buttons (src/providers/app-data-provider.tsx).
- Retain existing detailed timing logs for monitoring other stages.
- Frontend success notifications remain instant; background refreshes continue asynchronously.
* fix(profiles): prevent duplicate toast on page remount
* refactor(profile-switch): make active switches preemptible and prevent queue piling
- Add notify mechanism to SwitchCancellation to await cancellation without busy-waiting (state.rs:82)
- Collapse pending queue to a single entry in the driver; cancel in-flight task on newer request (driver.rs:232)
- Update handle_update_core to watch cancel token and 30s timeout; release locks, discard draft, and exit early if canceled (state_machine.rs:301)
- Providers revalidate status immediately on profile-switch-finished events (app-data-provider.tsx:208)
* refactor(core): make core reload phase controllable, reduce 0xcfffffff risk
- CoreManager::apply_config now calls `reload_config_with_retry`, each attempt waits up to 5s, retries 3 times; on failure, returns error with duration logged and triggers core restart if needed (src-tauri/src/core/manager/config.rs:175, 205)
- `reload_config_with_retry` logs attempt info on timeout or error; if error is a Mihomo connection issue, fallback to original restart logic (src-tauri/src/core/manager/config.rs:211)
- `reload_config_once` retains original Mihomo call for retry wrapper usage (src-tauri/src/core/manager/config.rs:247)
* chore(frontend-logs): downgrade routine event logs from info to debug
- Logs like `emit_via_app entering spawn_blocking`, `Async emit…`, `Buffered proxies…` are now debug-level (src-tauri/src/core/notification.rs:155, :265, :309…)
- Genuine warnings/errors (failures/timeouts) remain at warn/error
- Core stage logs remain info to keep backend tracking visible
* refactor(frontend-emit): make emit_via_app fire-and-forget async
- `emit_via_app` now a regular function; spawns with `tokio::spawn` and logs a warn if `emit_to` fails, caller returns immediately (src-tauri/src/core/notification.rs:269)
- Removed `.await` at Async emit and flush_proxies calls; only record dispatch duration and warn on failure (src-tauri/src/core/notification.rs:211, :329)
* refactor(ui): restructure profile switch for event-driven speed + polling stability
- Backend
- SwitchManager maintains a lightweight event queue: added `event_sequence`, `recent_events`, and `SwitchResultEvent`; provides `push_event` / `events_after` (state.rs)
- `handle_completion` pushes events on success/failure and keeps `last_result` (driver.rs) for frontend incremental fetch
- New Tauri command `get_profile_switch_events(after_sequence)` exposes `events_after` (profile_switch/mod.rs → profile.rs → lib.rs)
- Notification system
- `NotificationSystem::process_event` only logs debug, disables WebView `emit_to`, fixes 0xcfffffff
- Related emit/buffer functions now safe no-op, removed unused structures and warnings (notification.rs)
- Frontend
- services/cmds.ts defines `SwitchResultEvent` and `getProfileSwitchEvents`
- `AppDataProvider` holds `switchEventSeqRef`, polls incremental events every 0.25s (busy) / 1s (idle); each event triggers:
- immediate `globalMutate("getProfiles")` to refresh current profile
- background refresh of proxies/providers/rules via `Promise.allSettled` (failures logged, non-blocking)
- forced `mutateSwitchStatus` to correct state
- original switchStatus effect calls `handleSwitchResult` as fallback; other toast/activation logic handled in profiles.tsx
- Commands / API cleanup
- removed `pub use profile_switch::*;` in cmd::mod.rs to avoid conflicts; frontend uses new command polling
* refactor(frontend): optimize profile switch with optimistic updates
* refactor(profile-switch): switch to event-driven flow with Profile Store
- SwitchManager pushes events; frontend polls get_profile_switch_events
- Zustand store handles optimistic profiles; AppDataProvider applies updates and background-fetches
- UI flicker removed
* fix(app-data): re-hook profile store updates during switch hydration
* fix(notification): restore frontend event dispatch and non-blocking emits
* fix(app-data-provider): restore proxy refresh and seed snapshot after refactor
* fix: ensure switch completion events are received and handle proxies-updated
* fix(app-data-provider): dedupe switch results by taskId and fix stale profile state
* fix(profile-switch): ensure patch_profiles_config_by_profile_index waits for real completion and handle join failures in apply_config_with_timeout
* docs: UPDATELOG.md
* chore: add necessary comments
* fix(core): always dispatch async proxy snapshot after RefreshClash event
* fix(proxy-store, provider): handle pending snapshots and proxy profiles
- Added pending snapshot tracking in proxy-store so `lastAppliedFetchId` no longer jumps on seed. Profile adoption is deferred until a qualifying fetch completes. Exposed `clearPendingProfile` for rollback support.
- Cleared pending snapshot state whenever live payloads apply or the store resets, preventing stale optimistic profile IDs after failures.
- In provider integration, subscribed to the pending proxy profile and fed it into target-profile derivation. Cleared it on failed switch results so hydration can advance and UI status remains accurate.
* fix(proxy): re-hook tray refresh events into proxy refresh queue
- Reattached listen("verge://refresh-proxy-config", …) at src/providers/app-data-provider.tsx:402 and registered it for cleanup.
- Added matching window fallback handler at src/providers/app-data-provider.tsx:430 so in-app dispatches share the same refresh path.
* fix(proxy-snapshot/proxy-groups): address review findings on snapshot placeholders
- src/utils/proxy-snapshot.ts:72-95 now derives snapshot group members solely from proxy-groups.proxies, so provider ids under `use` no longer generate placeholder proxy items.
- src/components/proxy/proxy-groups.tsx:665-677 lets the hydration overlay capture pointer events (and shows a wait cursor) so users can’t interact with snapshot-only placeholders before live data is ready.
* fix(profile-switch): preserve queued requests and avoid stale connection teardown
- Keep earlier queued switches intact by dropping the blanket “collapse” call: after removing duplicates for the same profile, new requests are simply appended, leaving other profiles pending (driver.rs:376). Resolves queue-loss scenario.
- Gate connection cleanup on real successes so cancelled/stale runs no longer tear down Mihomo connections; success handler now skips close_connections_after_switch when success == false (workflow.rs:419).
* fix(profile-switch, layout): improve profile validation and restore backend refresh
- Hardened profile validation using `tokio::fs` with a 5s timeout and offloading YAML parsing to `AsyncHandler::spawn_blocking`, preventing slow disks or malformed files from freezing the runtime (src-tauri/src/cmd/profile_switch/validation.rs:9, 71).
- Restored backend-triggered refresh handling by listening for `verge://refresh-clash-config` / `verge://refresh-verge-config` and invoking shared refresh services so SWR caches stay in sync with core events (src/pages/_layout/useLayoutEvents.ts:6, 45, 55).
* feat(profile-switch): handle cancellations for superseded requests
- Added a `cancelled` flag and constructor so superseded requests publish an explicit cancellation instead of a failure (src-tauri/src/cmd/profile_switch/state.rs:249, src-tauri/src/cmd/profile_switch/driver.rs:482)
- Updated the profile switch effect to log cancellations as info, retain the shared `mutate` call, and skip emitting error toasts while still refreshing follow-up work (src/pages/profiles.tsx:554, src/pages/profiles.tsx:581)
- Exposed the new flag on the TypeScript contract to keep downstream consumers type-safe (src/services/cmds.ts:20)
* fix(profiles): wrap logging payload for Tauri frontend_log
* fix(profile-switch): add rollback and error propagation for failed persistence
- Added rollback on apply failure so Mihomo restores to the previous profile
before exiting the success path early (state_machine.rs:474).
- Reworked persist_profiles_with_timeout to surface timeout/join/save errors,
convert them into CmdResult failures, and trigger rollback + error propagation
when persistence fails (state_machine.rs:703).
* fix(profile-switch): prevent mid-finalize reentrancy and lingering tasks
* fix(profile-switch): preserve pending queue and surface discarded switches
* fix(profile-switch): avoid draining Mihomo sockets on failed/cancelled switches
* fix(app-data-provider): restore backend-driven refresh and reattach fallbacks
* fix(profile-switch): queue concurrent updates and add bounded wait/backoff
* fix(proxy): trigger live refresh on app start for proxy snapshot
* refactor(profile-switch): split flow into layers and centralize async cleanup
- Introduced `SwitchDriver` to encapsulate queue and driver logic while keeping the public Tauri command API.
- Added workflow/cleanup helpers for notification dispatch and Mihomo connection draining, re-exported for API consistency.
- Replaced monolithic state machine with `core.rs`, `context.rs`, and `stages.rs`, plus a thin `mod.rs` re-export layer; stage methods are now individually testable.
- Removed legacy `workflow/state_machine.rs` and adjusted visibility on re-exported types/constants to ensure compilation.
* feat(config): enhance configuration initialization and validation process
* refactor(profile): streamline profile update logic and enhance error handling
* refactor(config): simplify profile item checks and streamline update flag processing
* refactor(disney_plus): add cognitive complexity allowance for check_disney_plus function
* refactor(enhance): restructure configuration and profile item handling for improved clarity and maintainability
* refactor(tray): add cognitive complexity allowance for create_tray_menu function
* refactor(config): add cognitive complexity allowance for patch_config function
* refactor(profiles): simplify item removal logic by introducing take_item_file_by_uid helper function
* refactor(profile): add new validation logic for profile configuration syntax
* refactor(profiles): improve formatting and readability of take_item_file_by_uid function
* refactor(cargo): change cognitive complexity level from warn to deny
* refactor(cargo): ensure cognitive complexity is denied in Cargo.toml
* refactor(i18n): clean up imports and improve code readability
refactor(proxy): simplify system proxy toggle logic
refactor(service): remove unnecessary `as_str()` conversion in error handling
refactor(tray): modularize tray menu creation for better maintainability
* refactor(tray): update menu item text handling to use references for improved performance
* perf: utilize smartstring for string handling
- Updated various modules to replace standard String with smartstring::alias::String for improved performance and memory efficiency.
- Adjusted string manipulations and conversions throughout the codebase to ensure compatibility with the new smartstring type.
- Enhanced readability and maintainability by using `.into()` for conversions where applicable.
- Ensured that all instances of string handling in configuration, logging, and network management leverage the benefits of smartstring.
* fix: replace wrap_err with stringify_err for better error handling in UWP tool invocation
* refactor: update import path for StringifyErr and adjust string handling in sysopt
* fix: correct import path for CmdResult in UWP module
* fix: update argument type for execute_sysproxy_command to use std::string::String
* fix: add missing CmdResult import in UWP platform module
* fix: improve string handling and error messaging across multiple files
* style: format code for improved readability and consistency across multiple files
* fix: remove unused file