A group invite makes QQ emit a content-less c2c message (msgType=166/0, body
richText.elems=[]) right after the invite card. The TRACE log from #102's
diagnostic confirmed the body is truly empty — no commonElem, no msgContent,
no ptt/file — so it's a phantom control push, not an element type we fail to
decode. It was reaching OneBot clients as a confusing "[空消息]".
parseMsgPush now drops a message-kind event that decodes to zero elements when
its body carries nothing decodable. When the body *did* carry content but still
decoded to nothing (a real missing decoder), the event is kept and a warning
fires so the gap gets noticed rather than silently swallowed — replacing the
temporary #102 diagnostic with the actual fix.
A group invite surfaces a content-less "[空消息]" friend message (#102). Before
deciding whether to drop or decode it, we need to know what its body actually
carries — SnowLuma's decoder covers fewer element types than Lagrange, so an
empty decode may be a missing decoder (real content) rather than a genuinely
empty control push.
parseMsgPush now warns, on any message that decodes to zero elements, with a
summary of the body: each element's field names, every commonElem's
serviceType/businessType + payload hex, and any msgContent. One log line is
enough to identify the culprit element offline.
No behavior change — the message still dispatches as before. Temporary; remove
once the #102 culprit is identified and handled.
A group invite push carries no group name (confirmed across SnowLuma / Lagrange
/ NapCat — all three invite protos only have groupUin + invitorUid), and
get_group_info only resolved joined groups (fetchGroupList → roster), so an
invited group rendered "未知群名". Add a by-id server lookup so the name
resolves even when the bot isn't a member.
- proto-defs: 0x88D_0 request (field1 + config.uin + a request-mask `flags`
block) + response (groupInfo.results: name / member counts / owner / …),
field tags transcribed from
dev/Lagrange.Core/.../OidbSvcTrpcTcp0x88D_0{,Response}.cs.
- protocol: FetchGroupDetail oidb-service. Envelope reserved=0 (NO uinForm) —
Lagrange's GetGroupInfoService builds it with isUid=false.
- core: ContactsApi.fetchGroupDetail(gid) → QQGroupInfo | null; deliberately
NOT remembered into the joined-groups roster (mustn't pollute get_group_list).
- onebot: get_group_info falls back to fetchGroupDetail for non-member groups,
behind a 5-min per-id TTL cache so a burst of invites for one group can't
hammer 0x88D_0 (rate-limit / kick risk). noCache bypasses it.
- tests: request/envelope round-trip (reserved=0, uin, command), response
decode; + non-member resolve / cache-hit / noCache-bypass at the onebot layer.
Live note: byte-identical to Lagrange's 0x88D_0, but not verified against a real
account on this host.
When an incoming reply quotes a message SnowLuma's store doesn't have (e.g. one
from before it was running, or never observed), the receive pipeline now fetches
just that message from the server and persists it before dispatch — so a
consumer's get_msg on the quote resolves instead of returning "message not
found". Works for both group and private (c2c).
- core: MessageApi.getGroupMessageBySeq / getC2cMessageBySeq — fetch a single
message over a 1-wide [seq,seq] range, through the shared history throttle
gate (same rate-limit as the bulk history APIs — no flood / kick risk).
- onebot: backfillReplyTarget — on a group/friend reply whose target isn't in
the store, fetch (group SsoGetGroupMsg / c2c SsoGetC2cMsg, uid resolved for
c2c) and persist under the EXACT id the reply resolves to, so get_msg hits
regardless of who sent the quoted message (a bot self-sent c2c message
converts with user_id=self but is keyed under the peer session). Awaited
before dispatch so the consumer sees it; best-effort — a back-fill failure
never blocks delivery of the live message.
- tests: gating (already-stored / no-reply / nothing-on-server) + keying
(group, c2c, c2c self-sent under peer session).
Only fires on a store miss, so the common path (no reply, or quote already
stored) stays zero-cost.
Replying to a bot self-sent private message still failed get_msg on the quoted
id ("无法解析申请信息"). Root cause is the reply DECODE, not the keying: for a
c2c (friend) reply the canonical replied-to sequence is carried in
srcMsg.pbReserve.friendSequence (tag 8), while origSeqs[0] holds the per-sender
clientSequence. The decoder read origSeqs[0], so the reply resolved to a
sequence that never matches how the original message is keyed (by its server /
private sequence) — get_msg missed. Mirrors Lagrange's incoming-reply rule
`Sequence = reserve.FriendSequence ?? OrigSeqs[0]` (ForwardEntity.cs +
SourceMsg.PbPreserve.cs ProtoMember(8)).
- proto-defs: +SrcMsgPbReserve { receiverUid@7, friendSequence@8 }
- protocol: decodeRichBody reads friendSequence for c2c replies (group keeps
origSeqs[0], the shared group seq); fall back to origSeqs[0] when absent
- tests: c2c→friendSequence, group→origSeqs[0], no-reserve→fallback
This supersedes the keying approach in 4571864 (self-sent c2c keyed by
clientSequence), which is fully reverted here: the message is keyed by its
private sequence as before, and the fix is purely on the decode side so it also
covers replies to *received* old c2c messages, not just self-sent ones.
A friend replying to a bot self-sent private message → bot get_msg(quoted id)
returned "message not found" (bot then answered "无法解析申请信息"). c2c replies
reference a message by the sender's clientSequence, not the server sequence
(Lagrange: OrigSeqs = ClientSequence || Sequence), but self-sent private
messages were keyed by the server (private) sequence — so the incoming reply's
resolved id hash(clientSequence, peer, private) never matched the stored event.
Reply-SENDING still worked because it only needs the meta row (resolveReplySequence
queries by hash, no data), not the full event (findEvent requires data IS NOT
NULL) — which is exactly why get_msg / "引用解析" failed but replying didn't.
Key self-sent c2c messages by clientSequence (fallback to sequence, mirroring
Lagrange's `||`) via a shared c2cSelfMessageId helper, across all three c2c
self-sent paths (send_private_msg, private forward, forward-single-to-user).
Group messages unchanged (their replies reference the shared group sequence).
message_id is a signed int32 hash (hashMessageIdInt32) and is frequently
NEGATIVE, but both history actions declared it as `f.int({min:0}).default(0)`
— so a real anchor like message_id=-497311472 was rejected at param
validation (retcode 1400, 0ms) before run() fired, and the server-side
history fetch never happened (the client then reports "未找到足够的群聊记录").
Drop the min:0 bound → `f.int().default(0)`: negatives pass, absent/present-0
still mean "fetch latest", count stays ≥0. Matches every other action, which
uses f.messageId() for signed message IDs.
Regression tests: negative group/friend anchors now reach the handler;
absent message_id still defaults to 0.
Market faces (商城表情) now decode to an OneBot `image` segment carrying
NapCat-style markers (emoji_id/emoji_package_id/key/summary) plus a gxh
gif URL, so clients that don't special-case `mface` still render the
sticker while aware clients can reproduce it.
Adds the previously-missing send path: a dedicated `mface` segment OR an
`image` segment carrying `emoji_id` rebuilds the wire `marketFace`
(faceId=hex(emoji_id), itemType 6 / subType 3 / 300×300 /
pbReserve.field8=1) — byte-matched against NapCat and Lagrange.Core.
- proto-defs: MarketFace.key bytes→string; +MarketFacePbReserve{field8@8}
- protocol: decode faceId→emojiId(hex)/tabId/key; makeMarketFaceElem +
buildSendElems 'mface' case; format preview keeps the sticker name
- onebot: to-segment mface→image; message-parser mface + image(emoji_id)
- tests: wire decode + full proton round-trip; parse→build; segment shape
Reply/forward round-trips preserved (element keeps type 'mface' + text).
Mirror the group-history fetch for private chats via SsoGetC2cMsg
(trpc.msg.register_proxy.RegisterProxy.SsoGetC2cMsg, fields per Lagrange:
flat friendUid/start/end request, messages@7 response). get_friend_msg_history
now resolves the anchor sequence + the friend's UID, fetches real history from
the server, converts + persists each message, and falls back to the local
store on unknown anchor / no uid / empty / error.
- protocol: fetchC2cMessageRange (decodes via the friend decoder).
- core: the group + c2c fetch now share a generic walkBackHistory loop
(same throttle gate + caps); added MessageApi.getC2cHistory.
- onebot: getFriendHistory + a generalized persistHistoryEvent (keys private
by sender uin, exactly like the live path — so reply to old private
messages resolves too).
Tests: SsoGetC2cMsg request/response round-trip + friend-decode reuse.
get_group_msg_history was served only from the local observed-message store,
so it could never return more than what SnowLuma saw live (the "only 100
messages" symptom). It now fetches real history from the server.
- protocol: new SsoGetGroupMsg packet (trpc.msg.register_proxy.RegisterProxy
.SsoGetGroupMsg, sequence-range fetch; fields per Lagrange). Each returned
PushMsgBody is decoded by the existing group decoder via a factored-out
buildContextFromMessage.
- core: MessageApi.getGroupHistory walks backward in small sequence windows
with rate-limiting + data-volume guards — a process-wide throttle (≥300ms
between sends), 30-seq chunks, ≤12 requests, ≤200 messages per call — so a
hammering client can't make us flood the server and get kicked.
- onebot: get_group_msg_history resolves the anchor sequence (requested
message_id or latest observed), fetches from the server, converts + returns,
and persists each fetched message. Falls back to the local store on
unknown anchor / empty / error.
Persisting fetched history also fixes replying to pre-injection messages:
the reply path resolves message_id→sequence via the store, which now contains
the fetched messages.
Tests: SsoGetGroupMsg request/response round-trip + decode reuse (protocol);
walk-back chunking/dedup/cap/floor (core).
- Grid row height 80→108px so a 1-row stat tile's icon + 3 text lines clear
the padding with real breathing room (the previous fix still clipped ~8px);
stat tile gets an airier layout (size-10 icon, text-lg value, py-3.5).
- Carry the app's Apple language across the homepage: inner surfaces and
empty-state containers round to xl, empty-state icons calm to size-8/opacity,
edit hint + first-run nudge rounded consistently.
- OLED: toggle switch off-track is now a neutral foreground tint with a white
knob (was invisible bg-muted + bg-background on pure black); the OLED sidebar
override is scoped to sidebar-style=follow so panel/accent work in OLED too.
- Accent: the chosen accent now also tints the selected-state surfaces
(--accent / --sidebar-accent) so the sidebar active item and settings nav
rail follow it — default palette only, respecting accent scope.
- Palettes: add 6 more soft schemes (Rosé Pine / Moon / Dawn, Nord, Everforest
dark + light) via a pal() builder; schema unions + normalizer + bootstrap
light/dark classification updated to match.
- Light/dark switch now animates: a circular View Transitions reveal from the
click (mode/palette/intensity changes only; gated off under reduced/disabled
motion; instant fallback where the API is absent).
The settings sub-tab now lives in the URL (typed `?tab=` search param on the
/settings route, validated against the tab whitelist; default 外观 omits it).
The active tab is derived from the URL and tab clicks navigate (replace), so
tabs are shareable/back-button friendly. The sidebar "有新版本可用" banner now
links to /settings?tab=about, landing directly on 关于 (where 软件更新 lives)
instead of the default settings tab.
- Frosted "vibrancy" card (bg-card/65 + backdrop-blur-2xl) floating over the
wavy background, with a centered brand focal point (logo tile, name,
version, subtitle).
- Larger rounded fields/button (h-12, rounded-xl) for a calmer iOS feel.
- The "我觉得这个界面很卡怎么办?" help line becomes a 灵动岛-style frosted
capsule (rounded-full, blur, float, spring on hover/tap).
- New entrance/hover motion gated on reduceMotion/disableMotion.
The login password field showed two eye icons — our custom show/hide
toggle plus Edge/IE's native ::-ms-reveal. Hide the native reveal/clear
pseudo-elements so only our toggle remains.
Login page now has a mouse-reactive wavy-line background (ported a-waves +
embedded Perlin noise, themed to the accent, rendered behind the card which
sits on a 95%-opaque blurred surface so content stays readable). A "我觉得这个
界面很卡怎么办?" link opens a dialog that toggles it off — a device-local
preference (localStorage, works pre-auth, defaults off when the OS prefers
reduced motion), independent of the system motion settings.
Motion settings are now three:
- 关闭登录页的动效 — the login background toggle above (device-local).
- 减弱动效 (reduceMotion) — soft: transforms/springs off, durations
collapsed, page/sidebar instant; gentle entrance fades remain.
- 关闭全部动效 (disableMotion, NEW appearance field) — everything reduce does
plus data-no-motion CSS (kills pulse dots; spinners excluded so "busy"
stays legible) and the entrance opacity fades gated off across
overview/processes/logs/config lists.
- Scrollbars: theme native scrollbars (thin, token-colored thumb) for plain
overflow containers; Radix ScrollArea keeps its own custom bar (it hides
the native one, so no double bar).
- Sliders: new <Slider> component + `.slider` CSS replaces the native range
inputs (UI scale / mask opacity / blur). Filled track + ring-bordered
thumb, themed; input height matches the track so the thumb centers in every
engine; Firefox fills via ::-moz-range-progress.
- Stat tiles: tighten the h=1 tile (py-2 + mt-0.5) so the three text lines +
padding fit the 80px cell without hugging the border.
- Reduce motion: gate the always-present sidebar-width + page-transition
animations to instant when enabled (Framer's reducedMotion only suppresses
transforms, so these needed an explicit opt-out); kill the slider thumb
transition too.
Split the single centered card into a clean three-section stack consistent
with the rest of settings:
- Brand hero card: logo glow + name + version pill + tagline/description.
- "软件更新" grouped list (reuses Group/SettingRow): current-version row +
check-update row (status dot/spinner + 重新检查); when an update exists, a
tinted detail block with version, notes, platform asset hint and 前往下载.
- Support card: refined Star CTA + repo/releases/issues links.
UpdateStatus → UpdateGroup; all update states (checking/disabled/error/
latest) preserved.
Add a "配色方案" picker offering the four Catppuccin flavors as full color
schemes alongside the default sky theme.
- Schema: new `palette` field on UiAppearance (core + client mirror),
normalized to a 5-value whitelist defaulting to 'default'.
- Apply: each flavor's base/surface tokens are injected into the shared
theme stylesheet over :root (Latte) or .dark (the three dark flavors);
accent (--primary/--ring) stays separate so it composes. A flavor fixes
light/dark (Latte→light, rest→dark), overriding `mode`, and forces
dark-intensity to soft so the OLED "black" rule can't fight the palette.
sidebar-style + high-contrast still compose (higher specificity).
- Pre-paint bootstrap is palette-aware (no light↔dark flash on reload).
- Settings: swatch-preview picker; mode + dark-intensity controls disable
while a flavor is active. The top-bar ThemeToggle hides then (it would be
inert under a fixed-scheme palette).
Carry the logs/settings visual language across the remaining pages:
- Processes: sort control becomes a segmented radiogroup; action banner
gains a spinner; cards round to xl with a hover tint; calmer empty state.
- Config: save status collapses from bordered pills into one inline
dot/spinner indicator; node summary rows round to xl with hover and
fold the three flag badges into a single muted "数组 · 不上报自身 · 已设
Token" line; empty states round to xl; account rail rounds to xl.
Presentational only — all state wiring and handlers unchanged.
Rework the settings page around the macOS System Settings layout, reusing
existing UI tokens (no new deps):
- Two-pane: a sticky left nav rail (w-52) on lg, a wrapping tab strip on
mobile. Active section mirrors the app sidebar's selected style.
- Grouped inset lists: each Card is a header above a hairline-divided list
of label-left / control-right rows (SettingRow: auto / inline / stack),
replacing the stacked Field + Separator spam.
- Refined segmented control: recessed track with a raised pill on the
selected segment (font-semibold + full ring so it survives high-contrast
/ OLED); exposed as a proper radiogroup (role=radio + aria-checked) and
named via the row label through a small context + useId.
- Per-section fade transition; section heading anchors the detail pane.
Wide controls (时制, 深色强度, fonts, radius scale, poll interval, default
route) stack so they don't crowd their labels; password-saved confirmation
keeps its success color. All settings preserved 1:1.
Rework the run-log viewer around progressive disclosure and a calmer,
content-first hierarchy, reusing existing UI tokens (no new deps):
- Toolbar: live status dot (red when disconnected) replaces the badge;
real search field with inline icon + clear; icon-only actions with
tooltips, grouped (toggles in a track · divider · data actions).
- One always-visible segmented level filter (iOS-style raised pill +
colored dot) instead of strikethrough pills.
- Collapse the server-level switch + highlight-rule editor into a single
animated "选项" panel; server level also surfaced in the subtitle. The
control area drops from three rows to one.
- Log stream: subtle sticky column header (时间/级别/模块/消息) aligned to
rows via a sm:contents trick, dot-style level indicators, hairline row
separators, and distinct empty / no-match states.
a11y: radiogroup on color swatches, aria-pressed on level toggles,
unique remove-button labels, focus rings on all controls.
Rework the layout edit mode from a separate panel to in-place affordances:
- Each dashboard widget gets a semi-transparent floating overlay in edit mode:
a "拖动" grip bar (the whole card is the drag target — forgiving +
discoverable), an eye to show/hide, and a gear that opens the per-widget
settings in a dialog. Hidden widgets show as re-enableable "已隐藏" ghosts;
the bottom-right resize grip is now clearly visible.
- The sidebar nav becomes a drag-to-reorder list directly in the sidebar
(show/hide + locked pinned items + a 完成 button); editing force-expands the
sidebar (and disables the collapse toggle). Edit mode is a shared
LayoutContext flag so one "编辑布局" toggle drives both the grid and the nav.
- The separate editor panel (overview-layout-editor) is removed; the gear
opens config in a dialog instead.
Also: stat-tile typography tightened (leading-tight, smaller value, overflow
clip) so text no longer hugs the edge on small/short tiles. Verified by an
adversarial review round (no MUST-FIX).
A `pages` section in config/ui.json (3rd section-merge alongside appearance +
layout):
- defaultRoute: the page to open after login (chosen in Settings → 数据与格式;
DefaultRouteRedirect navigates there once, only from '/', respecting
deep-links + later navigation).
- logs: persisted visible levels, max lines, auto-scroll, line-wrap, and
keyword→colour highlight rules — edited live on the logs page; maxLines is
read via a ref so it never re-subscribes the SSE.
- processes: pid / name / status sort (persisted, default from prefs).
- config: remembers the last-used tab.
"重置全部" now also clears pages. The server normalizer is total and caps the
logs fields (levels subset, maxLines 100-5000, <=20 highlight rules, trimmed
keywords). Verified by an adversarial review round (normalizePages total
against 30 adversarial inputs; zero MUST/SHOULD-FIX).
A "高级" Settings tab adds:
- Custom CSS (appearance.customCss, <=50KB) applied post-auth only, as the
last <style> so it overrides everything. It never reaches the login page:
the server strips it from /api/ui/public, applyAppearance gates injection on
a bearer token (so the localStorage cache can't leak it pre-auth), and the
login page clears any lingering <style> on mount. `?safe-mode=1` (sticky for
the tab session) disables custom CSS so a broken rule can't lock the
operator out.
- Export / import the whole UI config as JSON (import is normalized
server-side, then reload); reset-all behind a confirm dialog.
POST /api/ui and the import file are size-capped (256KB) on top of the
per-field caps. Verified across 2 adversarial review rounds — the pre-auth
cached-CSS leak/lockout vector found in round 1 is closed across the boot /
logout / 401-expiry / re-login paths.
Overview widgets can carry an opaque `config` (stored per grid item). v1:
- alerts: how many entries + which log levels (was fixed at 5 warn/error);
the alert row colour now uses the full 6-level palette instead of a
warn/error binary.
- sessions: sort (recent / uin / nickname) + a nickname/uin filter.
A gear in the layout editor opens an inline config form per configurable
widget; edits apply live (no grid re-init — config isn't in the grid key) and
persist via the section-merge save. Config is parsed client-side with total,
clamping parsers (count 1-50, known levels only, filter <=100 chars); the
server stores it opaquely within a 4KB bound. Widgets with no config render
exactly as before. Verified across 2 adversarial review rounds.
The overview becomes a gridstack-backed free grid — drag, resize, and show/
hide 9 widgets (the 5 top stat tiles split out of the old single block, plus
connections/alerts/host/sessions), persisted to config/ui.json's
layout.overviewBlocks as {id,x,y,w,h,visible}.
- gridstack (zero-dep, vanilla; chosen over react-grid-layout whose
react-draggable uses the React-19-removed findDOMNode). React renders each
widget into a gridstack item via a portal (context flows through); gridstack
owns layout/drag/resize. Per-widget minW/minH prevent content collapse.
- Desktop = editable in place; narrow screens = auto single-column, read-only.
Fixed cellHeight => no ResizeObserver => the narrow column-clamp never
persists over the desktop coords.
- Edit mode shows a panel: widget show/hide + nav reorder (the phase-3 motion
list); the grid below is live-editable. Reset + empty-state preserved.
- Server UiLayoutItem gains optional x/y/w/h (clamped); the stats->5-tiles
migration + default placement live client-side (lib/dashboard-layout.ts).
Verified across 2 adversarial review rounds (gridstack+React lifecycle traced
against the library source: StrictMode-safe, no mount-persist loop, no
re-init on coord change, no narrow-visit data loss).
The verify text a user types when requesting to join a group is never in the
push — it lives on the OIDB pending-request queue. The fetch
(resolveGroupJoinRequest) was gated behind the UID→UIN stranger-resolve, so
when the requester's uin was already cached the whole async path (and thus the
comment fetch) was skipped, leaving the OneBot request.group `comment` empty.
Decouple the two: every uid-bearing group_invite now goes through the async
enrich path (needsGroupInviteEnrich → dispatchGroupInvite), which ALWAYS
fetches the comment and resolves the UIN profile only when needed (the
fromUin === groupId cache-pollution self-heal is preserved). Matches
Lagrange's unconditional FetchGroupRequests and NapCat's notify.postscript.
Rewrites the cache-hit test (which had pinned the buggy behavior) into a
regression asserting the comment is fetched + applied on a cache hit.
346/346 protocol tests pass. Verified by adversarial review.
The "C" half — reorder + show/hide the overview cards and the sidebar nav,
persisted to config/ui.json's layout section.
- LayoutContext loads the layout (authed GET /api/ui) and persists edits via
the section-merge POST /api/ui {layout} (debounced). reconcileLayoutItems
keeps stored order/visibility, appends new catalogue entries (forward-compat),
and forces pinned items visible.
- Overview is refactored into a block registry (stats/connections/alerts/host/
sessions). A 「编辑布局」 mode (motion Reorder — zero new deps) edits both the
cards and the nav in place, with per-item show/hide and a reset; an all-hidden
dashboard shows an empty state with 恢复默认.
- Anti-self-lock: PINNED_NAV = ['/', '/settings'] can be reordered but never
hidden — the overview hosts the layout editor and settings hosts
account/appearance, so the user can never strand themselves.
- api.ui.save now takes Partial<UiConfig> (the section-merge contract).
A late initial GET can't clobber an in-flight edit (dirtyRef); a pending edit
flushes on unmount. Verified across 2 adversarial review rounds.
Wires the Phase 1 store into a full appearance layer.
ThemeContext is rewritten to load/save UiAppearance from the server (with a
localStorage anti-flash cache + a one-time migration of the legacy per-browser
theme keys), apply ~12 dimensions to <html>, and theme the login page pre-auth.
New Settings controls (外观 / 数据与格式): free-hex accent (+ presets + apply
scope), dark intensity (soft / OLED black), sidebar style, background
(solid / gradient / uploaded image with opacity + blur), font presets, UI
scale, reduce-motion, high-contrast, default-collapsed sidebar, 12h/24h time
format (wired into logs + overview timestamps).
- index.html gains a pre-paint bootstrap script that applies the cached theme
before the bundle mounts, eliminating the dark/OLED/scale FOUC.
- server saveUiConfig is now a section-level merge so appearance-only and
layout-only saves never clobber each other.
- reduce-motion uses Framer MotionConfig; the background renders on a fixed
layer revealed by transparent app roots; custom-accent hex and ui-scale are
validated/clamped before use (localStorage-tamper hardening).
Verified across 2 adversarial review rounds (FOUC + CSS-injection + impure
state-updater + ref-staleness all fixed).
Backend foundation for the WebUI appearance + layout customization feature.
- core/src/webui/ui-config.ts: config/ui.json store with TOTAL normalization
(any input -> safe config), atomic write, lazy memoized cache, and the
background-image lifecycle (magic-number sniff for PNG/JPEG/WebP).
- server.ts: GET/POST /api/ui (bearer), GET /api/ui/public (unauth cosmetic
subset for pre-login theming), POST/DELETE /api/ui/background (upload with
Content-Length pre-check + 5MB cap + magic-number validation), and unauth
GET /ui-asset/background (serves the uploaded bytes with sniffed mime +
nosniff). SPA fallback excludes /ui-asset/.
- webui: `ui` API-client namespace + mirror types (no UI wiring yet).
Server-managed background fields (hasImage/imageMime/imageVersion) are taken
from the trusted source per path — the file on load, the on-disk config on
save — so a client save can never forge them, and a wallpaper survives a
restart. Verified across 3 adversarial review rounds.
A hot reload that changes the binding signature (url / role / accessToken
/ reconnectIntervalMs) drives `await this.close()` then `await this.open()`
back-to-back. `WsClientAdapter.close()` is synchronous and `WebSocket.close()`
just schedules the close frame, so the OLD socket's `'close'` event fires
AFTER the new socket has already been assigned to `this.socket`.
The close handler did not check the event's origin, so it would null out
`this.socket`, drop `connected`, and schedule an unwanted reconnect against
the freshly-opened connection. Against a single-connection backend that
kicks duplicates this snowballed into a reconnect storm where the bot and
the backend repeatedly kicked each other off; against a permissive backend
it still left the new connection in a degraded state (events couldn't be
sent until the bogus reconnect rebuilt the socket).
Add a one-line identity guard at the top of the close handler — when the
event arrives from a socket that is no longer the current one, do nothing.
The explicitlyClosed / isEnabled checks are kept as the second line of
defence for the normal (current-socket) close path.
Add three regression tests under tests/ws-client-stale-close.test.ts that
mock `@snowluma/websocket` (same approach as adapter-reload-state-machine.
test.ts, which avoids the native addon dependency): the reload race
scenario from the issue, a real-current-socket close that MUST still drive
reconnect, and a close() race against an in-flight close event.
The JSON segment preview was a fixed "[JSON]" placeholder, hiding both
real card titles and QQ's server-side fallback prompts (e.g. cardId
-896954183 "当前QQ版本不支持此应用,请升级", which is what #96 reported
seeing). Best-effort extract a title from the common ARK fields —
meta.{detail_1,detail,news,mannounce,music,video}.title, single-section
meta, then root.prompt / root.desc — and render as "[卡片:...]"; fall
back to "[JSON]" on parse failure or an unrecognised schema. Title is
truncated at 30 chars so a long prompt cannot blow up the log line.
OneBot segment shape is unchanged: downstream consumers still get the
full payload at data.data; this is purely a log-preview readability fix.
The helper is pure best-effort and never throws — every failure path
falls through to the existing "[JSON]" output.
* fix(onebot): support inline file url/path in send_group/private_msg
* test(onebot): update file segment parser test for inline url/path support
* fix(onebot): add canonical fields to file segment and fix log rendering
get_group_member_info serves straight from the in-memory member cache
on its default no_cache=false path — a deliberate choice, since clients
like OlivaDice/MaiBot query it once per inbound message and a per-call
OIDB refetch trips Tencent risk-control. But the group-admin push
decoder only emitted the notice and never updated the cached role, so a
member promoted to admin kept reading back as `member` and permission
gates rejected their commands. Patch the cached role in the decoder on
set/unset, only touching members we already know and never downgrading
the owner.
Also fix promote/demote misdetection: proton materializes an absent
embedded-message field as an empty `{}` rather than undefined, so the
old `extraEnable !== undefined` test was always true — every demote was
reported as a promotion with a fallback userUin. Key off whichever
sub-message actually carries an adminUid instead.
Fixes#93
When the control pipe drops while a send() is awaiting its ack,
rejectControlPending() rejects both the ack and the reply deferreds. The
reply is only awaited after the ack, so an ack-first rejection leaves the
reply promise rejected with NO awaiter — an unhandled rejection that, under
Node's default policy, kills the process. In Docker supervisor restarts it
(every session drops for ~1s); on a bare Windows run it just dies, surfacing
as "connection pipe closed" with the bot offline after hours of uptime.
Attach a no-op .catch() to each deferred at creation so a rejection can never
go unhandled (the real await still observes the value/rejection). Add a
regression test that drops the pipe mid-send and asserts no unhandled
rejection escapes — it fails without the guard.
Also install last-resort process guards (unhandledRejection / uncaughtException)
in the app entry: log and keep serving instead of exiting, so any other stray
rejection cannot take the daemon down and its stack is captured. The bridge's
watcher reconnects the dropped pipe on the next tick.
Advisory, read-only update check: the core polls GitHub's latest stable release, compares it against the running build, and surfaces a notice in the sidebar footer plus a detailed card (changelog + download link + platform-matched asset hint) in Settings → About.
SnowLuma never downloads or applies anything itself — cross-platform self-update is a minefield (Windows locks loaded .node/.dll, Docker bakes the app into image layers, bare installs have no supervisor), so the user downloads and applies manually.
The backend caches results for 6h to stay under GitHub's unauthenticated rate limit, degrades silently on failure, and can be disabled via SNOWLUMA_UPDATE_CHECK=0.
group_decrease collapsed GroupChange.decreaseType into a single isKick boolean, so a group disband (decreaseType 129) was indistinguishable from the bot being kicked — both surfaced as kick_me.
Preserve the protocol reason as a leaveType discriminator (leave | kick | disband) on GroupMemberLeave: decreaseType 129 -> disband, 0/130 -> leave, otherwise kick. The OneBot converter emits disband, or splits kick into kick/kick_me by selfId as before. leave/kick/kick_me are unchanged.
invokeOidb only threw on success=false when an OIDB errorCode was set (the
block added in #69); a send failure with no errorCode silently resolved with an
undecoded result. Surface it — throw errorMessage || 'packet send failed',
matching runOidb and the oidb-service tests that were already red on dev.
npm publish --provenance rejected @snowluma/mcp@1.9.5 with E422 because package.json had no repository.url to match the GitHub repo asserted in the provenance statement. Add repository/homepage/bugs (mirroring @snowluma/sdk and @snowluma/ui, directory packages/mcp) and refresh the now-inaccurate 'read-only' description. 1.9.5 was never successfully published, so re-running the publish workflow will ship it.