In rule mode, picking a node in a url-test or fallback group also points the first select group listing it at that group. Toggle in Misc settings, on by default.
* perf(enhance): cache chain items by file identity
Every config generation re-read and re-parsed all seven chain files.
Cache the parsed ChainItem keyed by (path, mtime, length) and clone on
hit: measured 486us -> 20us per generation (-95.9%) on a seven-file
chain set (merge/script/rules fixtures). Profile saves go through
tmp+rename so mtime always changes.
* perf(config): guard runtime yaml re-serialization
runtime_config_yaml re-serialized the multi-megabyte runtime Mapping on
every call even when unchanged. Hash the Mapping tree (Value: Hash) and
return the cached Arc<String> handle on hit: 4.10ms -> ~0.39ms (hash
290us, ~-90%) for a 3000-node config on the unchanged path, and the
handle keeps both hit and miss at a refcount clone instead of copying
the yaml body. The 64-bit SipHash key is sound here: a birthday
collision needs ~2^32 distinct config states at a volume of a few
hundred generations per day, and a colliding pair only serves one stale
yaml until the next config change re-hashes.
* build(deps): bump tauri-plugin-mihomo to 0.7.1
Pins the revision that derives Clone on ProxyType and DelayHistory,
which the upcoming proxy-view cache needs to return cached views
without changing the IPC contract.
* refactor(backend): right-size the new caches' storage
The chain cache never shares its stored ChainItem -- every hit
deep-clones it -- so its Arc wrapper only added indirection; store it
directly and const-shape it behind a LazyLock static instead of a
OnceLock accessor. The yaml cache keeps Arc<String> because callers
receive the stored handle, making hit and miss clones O(1).
* style(config): hoist the runtime yaml cache to module scope
Match the module-level placement of the chain cache and the other
named statics; the closure-local declaration shadowed nothing but read
worse and forced rustfmt to split the line.
* refactor(config): adopt tristr for config strings
Wire the tristr git dependency (pinned rev) and switch the config and
enhance modules from smartstring::alias::String to tristr::alias::String.
Drop the smartstring-era per-field specta overrides except the three
webdav fields required by serialize_with and the ipc::JsonValue bridge:
tristr maps to Primitive::str natively and the generated bindings stay
byte-identical (gen-contract --check passes).
Call sites in not-yet-migrated modules (cmd/core/feat/utils) are bridged
with explicit tristr::alias::String paths and as_str().into() conversions;
each bridge is absorbed when its module migrates.
* refactor(cmd): adopt tristr for command strings
Switch the cmd module from smartstring::alias::String to
tristr::alias::String and drop its smartstring-era per-field specta
overrides (ProfileSelection/ProfileCreate/ProfilePatch, RuntimeLogs);
bindings stay byte-identical. Calls into not-yet-migrated feat/core
functions keep smartstring-typed locals via full paths until those
modules migrate.
* refactor(core): adopt tristr for core strings
Switch the core module (validate, timer, sysopt, backup, hotkey, tray,
manager, notify) from smartstring::alias::String to tristr::alias::String
and drop the smartstring-era specta overrides on ValidationOutcome and
the notify payload structs; bindings stay byte-identical. WebDAV calls
from the not-yet-migrated feat module are bridged with as_str().into().
* refactor(feat): adopt tristr for feature strings
Switch the feat module (backup, icon, dns, profile, clash) from
smartstring::alias::String to tristr::alias::String, drop its
smartstring-era specta overrides (IconInfo, backup listings,
DnsOverrideOutcome), and absorb the temporary smartstring-typed bridge
locals the cmd module used to call into feat.
* refactor(utils): adopt tristr for utility strings
Switch utils/network.rs and utils/resolve/scheme.rs from
smartstring::alias::String to tristr::alias::String and absorb the
temporary user_agent and profile-name bridge conversions.
* refactor(deps): drop smartstring in favor of tristr
Remove the smartstring workspace and member dependencies and its
Cargo.lock entry; tristr is the only string representation left.
* build(deps): bump tristr to fuzz-harness fix rev
Pin 3738d8c renames the stale vstr path dependency in the upstream
fuzz manifest so its seven targets build again; library code is
unchanged from 3756ed1a.
Every menu API call runs on the main thread; invoked from a worker,
each call pays a blocking round-trip there and back (channel, event
loop hop, recv). Rebuilding the tray menu with proxy groups enabled
issued one round-trip per menu item — 200+ for a typical subscription —
and the per-call wait grows whenever the event loop is busy with
webview work.
Gather the async inputs (verge/clash/profiles snapshots, proxies and
runtime group order) on the caller's task, then build the menu and
attach it in a single run_on_main_thread closure where every item
constructs inline. Measured on a standalone tauri bench (5 groups x
40 nodes): a rebuild drops from ~6.2ms to ~1.6ms on an idle event
loop and from ~11.4ms to ~1.6ms with the loop busy (8ms of work per
16ms frame), with ~450 fewer allocations (~155KB) per rebuild.
* fix: mechanical string conversions
- &String parameters become &str (event payload fields borrow &str; wire format unchanged)
- Cow/std-string double conversions collapse to a single Into
- sha256 digests format through a shared hex lookup table instead of one
format! allocation per byte
- i18n placeholder replacement only allocates when the placeholder exists
* perf(enhance): marshal scripts by value and diff fields by fingerprint
- scripts cross into boa via JsValue::from_json/to_json instead of the
JSON text round-trip (config clone + json string + UTF-16 JsString +
stringify + re-parse all disappear); size guards use an allocation-free
approximate walk
- merge documents and tun/dns patches move values instead of deep-cloning
whole sub-trees per pass
- authoritative-field snapshots hash dns/hosts/tun values instead of
cloning them around every chain item (9 snapshots per enhance)
- proxy-name cleanup sets build from &str in one conversion
* perf(config): write yaml headers in place and record selections without copying
- save_yaml_str takes an optional header prefix written as separate chunks,
so the MB-scale serialized payload is never duplicated by format! just
to prepend a comment (profiles/verge/clash/runtime yaml all benefit;
file bytes unchanged)
- the runtime yaml string no longer round-trips through a prefixed copy;
write_runtime_file adds the header when persisting
- record/forget selected nodes mutate the stored selection list in place
instead of cloning it per click
* perf(tray): drop the redundant full rebuild in update_part
update_part rebuilt the menu twice back to back: once without proxy groups,
then a spawned full update_menu. Keep the partial rebuild only.
* fix(service): detect the service on OpenRC
* chore(deps): bump Service IPC and bundle to 2.8.1
* docs(changelog): note OpenRC service fixes
* docs(changelog): shorten OpenRC fix entry
---------
Co-authored-by: Tunglies <77394545+Tunglies@users.noreply.github.com>
Route frontend and tray node selections through the same backend flow for
selection, persistence, state refresh, and connection cleanup. Read the
group's previous selection before switching and preserve the original
rule of closing only connections whose chains contain that selection.
Run optional cleanup in the background and keep its failures separate
from successful selection. Remove redundant frontend effects and unused
IPC commands, and refresh the proxy view for every successful switch.
Refs: 32095daf90
Refs: bea0dde074
Quote plain hex short-id values before YAML parsing, since a value like 1234e123 becomes a float that cannot be restored. Keep the rewrite only when nothing else changes, and let the editor save check a quoted copy.
- Drop tauri-plugin-process: no frontend or Rust consumer; exit and
restart already go through the exit_app/restart_app commands backed
by AppHandle::exit/restart. Remove the registration and the unused
process:allow-exit capability.
- Replace the glob dev dependency with Node's built-in fs.globSync in
prebuild's service permission step (Node baseline is 26.13.0).
- Remove unused direct Cargo declarations: bitflags and regex (source
and workspace entries), and src-tauri's criterion dev-dependency;
clash-verge-draft keeps criterion for its bench.
- Narrow sysinfo to the network/system features actually consumed.
- Use futures::StreamExt for DelayQueue in the timer and drop the
tokio-stream direct declaration (still transitively present via
clash-verge-service-ipc).
Refresh the tray menu, icon, and tooltip after service operations so restored TUN capability is reflected without changing the saved TUN setting.
Ref: e60285b4ae
* fix(backup): serialize restore with config write lock
Acquire the configuration write lock at the entry of
finalize_restored_verge_config so the clash, profiles and verge draft
applies all happen under it, instead of only the verge layer. A
concurrent patch (patch_clash / patch_verge via try_lock_config_write)
could otherwise stage a draft that the unlocked restore applies
overwrite, or interleave its persistence with the restore.
Also drop the pointless clones in the clash and profiles edit_draft
closures; edit_draft takes FnOnce so the restored values can be moved
in. The lock is released before patch_verge, which re-acquires it via
try_lock_config_write.
* perf(backup): extract archive off the runtime
Run File::open, ZipArchive::new and the extraction of the backup
archive inside spawn_blocking. Extracting a multi-file backup blocked
a tokio worker for hundreds of milliseconds to seconds while the
runtime kept serving other work on the remaining workers.
Move the PROFILE_WRITE_LOCK acquisition after the extraction, right
before finalize_restored_verge_config: the lock protects the
config-data transition (extracted files reloaded into drafts), not the
plain filesystem writes, so profile operations no longer queue behind
the blocking unzip.
* fix(clash): take config write lock when changing mode
change_clash_mode committed the clash draft and saved config.yaml without
claiming CONFIG_WRITE_LOCK, so a concurrent patch_clash transaction could
have its staged draft committed here or its file-snapshot rollback swallow
the mode change. Take lock_config_write (waiting) rather than
try_lock_config_write: two of the three callers (hotkey, tray) discard the
result, so fail-fast would silently drop the switch, and the lock is only
attempted after the mihomo patch already applied, where an Err would roll
the frontend back while the core runs the new mode.
* refactor(utils): drop dead reset helper and tidy init_resources
- remove resolve_reset_async, which had no callers
- drop the duplicated time.*.com entry from the default fake-ip-filter
- init_resources no longer recreates app_home (ensure_directories during
init_config already created it) and now propagates res_dir creation errors
* refactor(profile): remove identical blind retry when switching node
The second select_node_for_group call in switch_proxy_node was a
zero-interval retry with identical arguments. Its success path was a
degraded copy of the first (missing the Refresh::Proxies announcement
ever since the old force-refresh emit was removed), and its distinct
"回退/最终失败" log wording misled maintainers. Keep a single call,
the primary success aftermath, and one error log.
* refactor(enhance): simplify tun mode check and lowercase helpers
- Drop the always-true second clause in use_tun's enhanced-mode check:
the key is absent only when the unwrap_or default already makes
current_mode == "fake-ip" true.
- Collapse use_lowercase into a delegation to use_lowercase_owned.
* refactor(core): drop unreachable hotkey dtor and dedupe tray predicate
Remove impl Drop for Hotkey: the instance lives in a OnceLock static via
the singleton! macro and is never dropped; exit-time unregistration is
already handled by Hotkey::reset().
Spell the include_proxy_groups/display-mode predicate identically at
both evaluation sites in tray menu building; behavior is unchanged
(Option != Some("disable") is equivalent to unwrap_or("default") != "disable").
* refactor(config): fold port guards into single helper
The five IClashTemp port guards duplicated the same read/parse/default
logic with drifting defaults: some hardcoded 7895/7898/7899 while others
referenced network::ports constants. Route them all through one
guard_port_value helper and use the constants everywhere, keeping each
guard's name, key and cfg attributes.
* refactor(config): dedupe prfitem factories and companion seeding
from_rules/from_proxies/from_groups differed only in uid prefix, type
string and template, so they now delegate to a shared from_seq_item.
The five current_* accessors collapse into a local macro. The identical
five-block companion seeding in from_local and from_url moves into
ensure_companion_items, preserving seeding order and error propagation.
* refactor(config): share patch macro and profiles helpers
IVerge::patch_config, IProfiles::patch_item and PrfOption::merge each
hand-rolled the same copy-Some-overwrite pattern; they now share one
patch_opt_field macro (the or-else merge form is field-for-field
equivalent for Option fields). The four profiles *_safe wrappers that
inlined Config::profiles().with_data_modify boilerplate delegate to a
with_profiles_edit helper, and record/forget_selected_node share the
current-item lookup prelude via current_item_mut.
* refactor(enhance): merge chain processing and drop unused builtin js payload
- collapse the identical process_global_items/process_profile_items into
one parameterized process_override_items (entry uid sources unchanged)
- fold the identical rules/proxies/groups arms of ChainItem::from_async
into a single arm with an inner dispatch
- stop embedding builtin JS payloads never executed by any path: the
enhance loop dispatches builtin items by uid to native implementations,
so ChainItem::builtin now carries empty script data (uids unchanged)
- source ProfileItems fallback literals from Default instead of a second
hand-written copy in collect_profile_items
* refactor(net): reuse webpki tls config and fallback
Extract build_webpki_rooted_tls_config so the delay-test TLS connector
and the reqwest clients share one static-webpki-roots builder; ALPN
stays caller-specific (only reqwest offers h2).
Collapse the duplicated platform-verifier -> static-webpki-roots retry
in get/get_bytes into request_with_static_webpki_fallback, gated by the
caller's certificate-verification setting.
* refactor(app): extract common exit cleanup
restart_app and quit duplicate the whole exit preparation: set the
exiting flag, persist configs, run interactive cleanup, and on a failed
core stop clear the flag, restore public DNS under the lifecycle lock
on macOS, and send the stop-failed notice.
Move that flow into run_exit_cleanup_flow next to clean_async, taking
the notice variant as a parameter and returning the CleanupResult so
each caller keeps its own tail: restart aborts before shutdown, quit
always shuts down and exits with the cleanup code.
* refactor(backup): share restore finalize path
restore_webdav_backup and restore_local_backup duplicated the
credential extraction (backup creation strips the WebDAV credentials,
so both re-inject the live ones), the archive extraction and the
finalize tail. Move that tail into a private restore_backup_archive
helper so the two entries keep only their download/locate logic; the
config-write lock scope and the off-runtime extraction land in one
place. The webdav entry keeps its downloaded-file cleanup.
* refactor(sysopt): drop redundant lock and busy-wait
The proxy target pair sat behind an Arc<RwLock> with no readers and an
Arc that was never cloned; both writers already serialize on
update_lock, so store the pair inside that mutex and delete the RwLock.
refresh_guard waited for the guard to become ready with a yield_now spin
at full CPU while holding guard_operation_lock; a 10 ms sleep keeps the
same wait condition and lock scope without the spin.
* refactor(sysopt): share macos bypass formatting
The macOS default bypass list and its three-way composition (empty ->
default, default + custom, custom only) were maintained twice: in
sysopt's format_bypass and inline in proxy_control's service_bypass.
Keep the literal once in constants and compose both paths through
format_bypass; service_bypass keeps its NUL check and UTF-8 truncation.
* refactor(core): dedupe window listeners, elevated runners, backup settings
- Collapse the close/focus window listener pairs in lightweight.rs onto a shared WindowListener with per-listener event and log label.
- Extract the windows runas-or-direct execution into service::run_maybe_elevated, shared by win_uwp and the service uninstaller, and factor the common uninstaller path resolution and status check shared across platforms.
- Make the auto-backup watch channel the single settings source, replacing the duplicate Arc<RwLock> copy with a stored receiver.
* refactor(enhance): fold lowercase helpers into single owned function
use_lowercase_owned disappears; the surviving use_lowercase takes the
owned Mapping and borrowed callers clone at the call site.
* refactor(enhance): rename ChainSupport variants to Mihomo names
Clash is a legacy name; the variants match the verge-mihomo and
verge-mihomo-alpha cores they gate.
* refactor(enhance): lowercase takes a reference and chain support is copy
use_lowercase borrows the mapping and clones only the values; call
sites no longer clone whole configs. ChainSupport is a fieldless enum
and now Copy; builtin() registers each support pair without clones.
* refactor(config): name prfitem constructor for what it builds
new_seq_item states that a fresh item is built from a template; the
from_ prefix wrongly implied an existing source. append_companion_item
returns the uid directly: append_item bails on a missing uid, so the
Option was always Some on the Ok path.
* refactor(sysopt): name the proxy state lock for its full role
The lock now owns the (Sysproxy, Autoproxy) state, so proxy_state_lock
says what it guards; the guard locals follow. constants::proxy held a
single sysproxy constant and is renamed sysproxy.
* refactor: replace remaining clash meta naming with mihomo
Doc comment, generated bindings and the prebuild sidecar helpers stop
referring to the legacy Clash Meta name; installer legacy-process
cleanup keeps the old executable names on purpose.
* refactor(enhance): drop builtin core filtering that never discriminates
Both valid cores matched a different ChainSupport variant, yet each
variant registered the same uid pair, so the filtered builtin list was
identical for stable and alpha. Invalid cores are coerced at startup
(validate_and_fix_config / get_valid_clash_core) and the None arm of
is_support was unreachable. builtin() now returns the items directly
and the clash_core threading through ConfigValues goes away; re-add
core discrimination when a builtin actually differs per core.
* refactor(enhance): drop unreachable mode script guard
merge_default_config overwrites the profile's mode with the verge-owned
clash config, whose mode is always present (template seeds rule and
patch semantics never clear it) and restricted to rule/global/direct;
change_clash_mode is rejected by the core before persisting anything
else. mode:script can therefore only arrive through a hand-crafted
patch_clash call, and for that path an explicit invalid-mode validation
error is more honest than silently rewriting to rule. mihomo rejects
mode:script outright (invalid mode), and user merge/script chains run
after the builtins, so the guard never covered them anyway.
* refactor(enhance): call the hysteria alpn builtin directly
builtin() existed to feed the uid-dispatch loop; with a single
builtin the vec and match are ceremony. apply_builtin_scripts calls
builtin_hy_alpn directly and to_script goes away; tests construct
ChainItem literals the way the merge items already did.
The service core runs with its own home directory, so relative and absolute
private-key paths no longer resolved. Stage them as runtime assets under an
opaque name; UNC and device paths are left to the core.
Refs #8445, #8229
This reverts commit 12f842d038.
Restore the Accessory activation policy when closing the main window or
using Cmd+W so the Dock icon disappears again. Keeping the Regular policy
avoided menu bar jitter but caused the Dock visibility regression in #8513.
Hiding the application before switching to Accessory did not reliably
eliminate the jitter. An isolated Tao 0.37.2 checkout with a native muda
menu reproduced it on macOS 27.2 without CVR or WebView. The responsible
layer remains undetermined and needs separate investigation.
The menu bar jitter reported in #8319 can still occur; this revert restores
Dock behavior and does not claim to fix that issue.
Closes#8513
Serialize Verge patches behind the active configuration writer so toggling the system proxy after TUN does not fail immediately on lock contention.
CLose: #6613, #8481, #8357
`watch::Sender::send` drops the value when there are no receivers at
startup, leaving the scheduler on the default disabled settings.
`send_replace` stores it for the runner that subscribes right after.
Support the dns-mode field introduced in mihomo v1.19.33. A
button group below the stack selector offers
Auto/Disabled/Native/Hijack; Auto writes an empty string, which
the core treats as the legacy behavior. Adding dns-mode to
GUI_KEYS makes the value GUI-authoritative once saved, matching
the other TUN dialog fields.
This reverts commit cb751052a4.
Remove the tray latency test and its dashboard URL persistence and cache
changes because the two entry points have diverging behavior.
Preserve the subsequent IPC binding migration and remove generated fields,
input projections, and locale entries that depended on the reverted feature.
The switching flag was released by a defer registered inside
perform_config_update, so a panic during the write-lock wait or the
draft edit could leave it stuck set. Register the defer right after the
CAS. Skip the generic validation notice for busy switch outcomes; the
frontend already shows its own busy toast.
Every traffic refresh rebuilt the attributed-string dictionary (font,
color, paragraph style, baseline number, dict) and ran a CoreText layout
via NSAttributedString::size() just to size the status item, although
both depend only on the button height and the glyph advance.
The dictionary and the monospaced per-glyph advance (measured once
through the same layout engine with a "0" probe) are now cached keyed by
button height, and the width is computed as longest-line length times
the advance. The speed charset is pure ASCII on a monospaced font, so
the arithmetic width matched the layout width exactly (max deviation
0.000pt over 2000 realistic samples).
Standalone A/B over 43,200 updates (12h at 1 Hz), same hot path:
| variant | mean | p95 | phys_footprint after |
| --- | ---: | ---: | ---: |
| rebuild + layout (old) | 6.89us | 7.8us | 19.8MB |
| cached attrs | 4.71us | 5.0us | 7.8MB |
| cached + arithmetic width | 0.27us | 0.3us | 6.0MB |
The retained-memory delta is malloc-zone fragmentation from the churn,
which this removes at the source.
The cap of 8 was a conservative pick from 8ccb0b9a; that commit's own
setup benchmark had the 4-worker / 8-blocking row as the best config:
639.3ms avg / 645.5ms max setup time and 22.1 avg threads, ahead of the
adopted 8-worker row (650.1ms / 662.4ms, 26.0 threads).
Startup parallelism is unaffected in practice: the setup join! polls
cooperatively on one task, and CPU work (validation, script eval, YAML
parse) runs on the spawn_blocking pool, which keeps its 8-thread cap
through the existing 2x invariant. A 12h live session showed the 8 idle
workers accumulating under 0.5s CPU total, so four parked stacks
(~0.4MB dirty) are pure overhead.
* feat!: remove signal crate
* fix: clippy
* feat(signal): limit the shutdown signal crate to Linux
Upstream now covers termination and session-end events on Windows and
macOS, but Linux has no equivalent, so pre-shutdown cleanup there still
depends on listening for SIGTERM/SIGINT/SIGHUP (#7827). Restore
clash-verge-signal behind a target_os = "linux" gate instead of dropping
it: the dependency lives in a linux-only target section, the crate root
carries an English Linux-only rationale and compiles to nothing
elsewhere, and the Windows backend is gone.
register() now takes the quit flow's unit future: the latch serializes
in-flight cleanups and releases after each attempt, leaving the
exit-or-cancel decision to the quit flow itself.
Verified with cargo check and clippy for aarch64-unknown-linux-gnu, and
cargo check on the host where the crate compiles empty.
tauri-cli's get_binaries() adds every file in src-tauri/src/bin/ as a
bundle binary, and the deb/rpm packagers copy each one to /usr/bin, so
the dev-only bindings generator ended up as /usr/bin/frontend-contract
(1.0 MB) in release packages and broke downstream repackaging (Copr:
"Installed but unpackaged file(s) found"). There is no bundler option
to exclude a binary, so move the generator to an example target, which
cargo keeps dev-only and get_binaries() never scans, and invoke it via
'cargo run --example'.
Ref https://github.com/clash-verge-rev/clash-verge-rev/pull/8131#issuecomment-6137907930
Hostnames such as mi13.local are resolved by mDNS on the local network,
which the proxy core cannot do, so they failed with 'couldn't find ip'
while the system proxy was on. The macOS default already bypasses
*.local.
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
* feat(tray): add a per-group latency test
The tray's proxy group submenus offered no way to test a group: refreshing
the delays shown there meant opening the dashboard and running a test from
the proxy page.
Add a "Test Latency" entry at the head of each group submenu. It measures
the group through the proxy page's own view, so member resolution and
provider attribution match the page exactly: provider nodes are health
checked on their provider, anything else goes through the proxy delay
endpoint, and members the page cannot resolve are skipped. Nested groups
are therefore tested through their currently selected node, as the page
does. The menu is rebuilt when the run finishes.
The test URL follows the page's chain -- the group's own test URL, then the
configured default, then the built-in one -- including the page's validity
check, so a group with a malformed URL falls back where the page falls back.
A group test started from the app rebuilds the tray as well, so the two
views do not drift.
Co-Authored-By: Claude <noreply@anthropic.com>
* fix(tray): keep latency tests and displayed results consistent
* fix(tray): isolate latency histories and preserve usable menus
* fix(tray): keep settings usable when proxy reads fail
* fix(tray): show selected node latency for proxy groups and display timeouts as T/O
Co-authored-by: wonfen <eiwonfen@gmail.com>
* fix(tray): refresh proxy latency on menu interaction
---------
Co-authored-by: Claude <noreply@anthropic.com>
Co-authored-by: wonfen <eiwonfen@gmail.com>
* refactor(updater): drop write-only pending_* slots
* refactor(notify): fold sequence counter into entries lock
* refactor(port): fold startup-core-blocked flag into reason slot
* refactor(service): drop redundant provider sync serial lock
* refactor(port): store startup block reason in a OnceLock
* refactor(core): replace macOS network monitor slot with armed latch
* feat(contract): export wire contract from backend and generate frontend types
Adds a zero-dependency `frontend-contract` bin that serializes the real
NoticeStatus variants and FrontendEvent payloads, scripts/gen-contract.mjs
which emits the committed src/services/contract.ts, and a CI job that
regenerates it and fails on a non-empty diff. events.ts now derives its
event map from the generated contract instead of hand-typed shapes.
* feat(events): route all backend events through a single event bus and store
Adds src/services/bus (the only listen/emit site: one handler per contract
event plus the frontend-internal test-all event) and the React
context/reducer app store. Migrates the five scattered subscriptions
(layout events, pending failures, proxy refresh, profile update/timer,
test-all), deletes the old glue (events.ts, use-listen, use-layout-events,
LoadingCacheProvider), moves run state into the store, and adds an
architecture gate script wired into CI. Behavior is unchanged; the
run-state query's poll/focus safety net is preserved by the store
provider.
* refactor(notice): make the notice table exhaustive and drop dead wire entries
D5 cleanup, frontend and backend in one commit: removes the six
config_validate::* statuses that never receive a backend send and the
set_config::ok send (enum variants, golden list 41→34, send point in
restart_clash_core), their frontend handlers, and the five i18n keys only
those handlers used. The notice table is now typed as a Record over the
generated NoticeStatus union, so a status without a frontend entry fails
to compile. The info status (sent by the silent updater for breaking
change notices, previously dropped with a console warning) now shows an
info toast — the one deliberate behavior delta.
* feat(mutate): single funnel for user-intent commands and invoke gate
Adds src/services/mutate.ts: one path for optimistic apply, services-layer
invoke, Busy handling, failure rollback, and the error toast, with busy
bookkeeping mirrored into the app store. The patch hooks (verge, clash,
profiles) now route through it. Moves the remaining direct invokes
(unlock ×3, DNS ×4) into cmds.ts and turns on the invoke-only-in-services
architecture gate in CI.
* refactor(settings): route settings-domain commands through the mutate funnel
Ports port saving, webdav/local backup operations, core change/restart/
upgrade, service install/uninstall/repair flows, the system-proxy toggle
queue, the DNS override toggle, and app lifecycle commands onto the single
funnel. Bespoke flows keep their own error handling via errorNotice:false;
simple flows gain the funnel's default error toast and revalidation.
* refactor(profiles): route profile operations through the mutate funnel
Ports profile create/import/reorder/delete/update/patch/save-file and
enhance calls onto the funnel; retry and editor flows keep their own
error handling with errorNotice:false. Per-uid ids keep the update
worker pool concurrent under the funnel's in-flight dedupe.
* refactor(proxies): route proxy/home/misc commands through the mutate funnel
Ports mode switching (page and home card), chain runtime updates, node
selection with tray sync, selection recording, provider updates and
syncs, and lightweight-mode entry onto the funnel; every site keeps its
existing optimistic/revert and error-handling behavior.
* docs: add frontend architecture guardrails for src
Layering rules for the contract flow: generated contract as the only
wire-typing source, event bus as the only listen/emit site, pure-reducer
store for event-derived state, the mutate funnel for user intent, and
invoke only in the services layer — each backed by the CI gates.
* fix(events): close run-state race and restore single error toasts
Self-review findings: the bus now re-reads the run-state snapshot in
onSubscribed so an event racing the store's initial read cannot leave a
stale snapshot (the ordering the old layout listener guaranteed), and the
patch hooks pass errorNotice:false so GuardState/page handlers remain the
only error toast, matching the pre-migration behavior.
* fix(frontend): preserve queued intents and fresh runtime snapshots
* refactor(frontend): trim unused mutation and busy state APIs
* docs(frontend): focus guardrails on architecture ownership