The develop rebase merged the #120 title-only-change branch and the
#248/#251 store-side empty-guard into one onStoreDocument. The existing 14
tests exercise each only in isolation (empty-guard tests send no title
fragment; title tests send a non-empty body), so none reached the
empty-guard's blocking branch with titleChanged===true. Add two paired
regression tests on that exact junction:
- empty body + a changed non-empty title over non-empty persisted content,
no intentional-clear → the empty-guard blocks the WHOLE store, dropping
the simultaneous rename too (updatePage not called); the rich content and
old title survive.
- the same doc with a deliberate clear armed via the real stateless
transport → the empty body is allowed and the rename rides along on the
same body-path updatePage (title + empty content persisted).
The pair makes Test 1 non-vacuous: same doc, only the clear differs, and
Test 2 proves updatePage IS reachable — so Test 1's "not called" is the
guard blocking, not an unreached path. Test-only; no production change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add an inline spoiler (Telegram/Discord-style hidden text): a TipTap mark
`spoiler` rendered as <span data-spoiler="true" class="spoiler">, blurred via
CSS and revealed on click (UI-only is-revealed class, never persisted).
- packages/editor-ext: the Spoiler mark (inclusive:false, set/toggle/unset
commands, ||text|| input rule), exported; a lossless turndown rule emitting
raw inline HTML; round-trip test.
- apps/client: SpoilerView mark-view (ReactMarkViewRenderer, Link pattern),
registration in extensions, bubble-menu toggle button (editable only), CSS
(blur + @media print reveal), en/ru i18n.
- apps/server: register Spoiler in collaboration.util tiptapExtensions so the
mark survives HTML<->JSON export/index/import/Yjs; a test proving the public
share keeps the spoiler (it isn't stripped with comments).
No keyboard shortcut: the proposed Mod-Shift-s collides with Strike (and
Mod-Shift-h with Highlight); the ||text|| input rule + the bubble-menu button
cover ergonomics.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F1: add a test that empties a non-empty doc via a change-origin transaction
(ySyncPluginKey meta, the shape y-tiptap sets for remote/merge updates) and
asserts the intentional-clear signal is NOT emitted — pinning the
isChangeOrigin early-return that keeps remote emptiness from punching through
the #248 server guard. The 4 existing tests use local transactions and never
exercised that true-path (verified: removing the guard fails only this test).
F2: record the #248 empty-overwrite guard and the #251 intentional-clear in the
CHANGELOG [Unreleased] Fixed section.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F17: the header's content-clause literal omitted the [[:space:]]* tolerance;
copy page.repo.ts's exact '"type"[[:space:]]*:[[:space:]]*"text"' (jsonb::text
renders a space after the colon, which is why the tolerance exists).
F16: remove expect(ttl).toBeGreaterThan(0) — the twin of the F15 removal;
expect(ttl).toBe(120) strictly subsumes it.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The success path calls notifications.show without a color (Mantine default,
not green), and the test asserts color is undefined — rename the case from
'GREEN success toast' to 'success toast with no error color'.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F14: the lockstep int-spec header still described the pre-F6 two-clause set with
'iff' — add the content-JSON text-node clause so it matches embeddablePredicate.
F15: remove the redundant expect(ttl).toBeLessThanOrEqual(120) that followed
expect(ttl).toBe(120).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F11: PRE_SEED_TTL_SECONDS 45->120 (= client REINDEX_POLL_CAP_MS). At concurrency
1 a queued reindex can wait past the old 45s; if the pre-seed expired while
pending, getMasked fell back to the COUNT and reported done, so the client
stopped polling and missed the climb. Tie the pre-seed TTL to the client cap.
F12: extend the lockstep integration spec — insertPage takes content; a
text_content=null + text-node-content page is IN and a math-only page is OUT,
pinning the structural "type":"text" clause (and the jsonb space-after-colon).
F13: list all three embeddable clauses in the reindex JSDoc/inline comments.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add space-tree-node-menu.test.tsx driving handleMakeAvailableOffline through
the menu: full success → green 'available offline' toast; ydoc not synced
(result.ok but !didSync) → red toast naming 'editor' (the F1 guarantee a
non-warmed page is not reported available); read-query failures → red toast
listing labels; thrown error → extracted reason. Mutation-checked (inverted
gate flips two cases).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F6: extend embeddablePredicate to pages with body content but null text_content,
keyed on the text-node marker "type":"text" (not a bare "text": key, which
also matched math nodes' attrs.text and would leave math-only pages stuck
below 100%). Numerator and denominator share the predicate; tests assert the
compiled WHERE is byte-identical and a math-only doc is excluded.
F7: correct the start() JSDoc (both totals are the real page count).
F8: nextReindexPollInterval reuses isReindexComplete.
F9: getMasked reads progress first and skips the two COUNTs while a reindex is active.
F10: pre-seed the progress entry with a short 45s TTL so a deduped enqueue's
phantom "0 of N" expires quickly instead of sticking for the 1h TTL.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F1: warmPageYdoc now returns didSync (true only on the real 'synced' event,
false on the 8s timeout / error, which are now logged). The tree menu gates
the 'available offline' success toast on result.ok AND didSync, and pushes
'editor' into the failed set otherwise — so a page whose Yjs body never
warmed is no longer reported as fully offline-available.
F2: tests for the page/space/currentUser failure labels and the didSync
true/false paths.
F3: correct the mobile-app-plan / mobile-bootstrap present-state sections to
match shipped code (Capacitor, CORS allowlist, Swagger, offline reading,
/l in the SW denylist).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
guardedFetch blocks loopback/private/link-local/metadata IPs and never calls
fetch; decryptHeaders fails open (returns undefined, warns once, no blob leak).
yjs.util setYjsMark/removeYjsMarkByAttribute/updateYjsMarkAttribute on real
Y.Docs. SHARED_TOOL_SPECS<->in-app parity (name/desc/input-schema; a dropped or
renamed wiring fails). Replace the tautological storage.service spec with
driver-delegation checks across every public method.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
findBreadcrumbPath set node.name='Untitled' in place, mutating the shared
sidebar tree (treeData passed from resolveBreadcrumbNodes). Surface 'Untitled'
via a shallow copy on the returned chain only; input nodes stay untouched.
Add tests for the non-mutation invariant plus applyUpdateOne reducer,
formatRelativeTime buckets, and the pure tree mappers (sortPositionKeys,
pageToTreeNode).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The pub/sub error handlers were `(err) => () => {}` — a noop returning an
inner arrow that never runs, so socket.io redis client errors were silently
swallowed. Log them via Nest Logger. Adjacent pre-existing bug surfaced in
review of #255.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The WsRedisIoAdapter creates two ioredis clients (pubClient/subClient) for
@socket.io/redis-adapter but never closed them, leaking their TCP handles on
application shutdown (#255). The redis-adapter does not own these clients'
lifecycle, and the adapter is instantiated from main.ts (not a DI provider),
so no Nest lifecycle hook applied to it.
Keep references to both clients and override dispose(), which Nest's
SocketModule.close() invokes exactly once during shutdown after all socket.io
servers are closed. Use disconnect(false) to mirror the sibling pub/sub pair
in collaboration/extensions/redis-sync (onDestroy): immediate close, no QUIT
round-trip, no auto-reconnect. Refs are nulled to guard against double-close.
Runtime behavior is unchanged; only the shutdown path is added.
Verified with a script that boots connectToRedis() against a real Redis:
2 sockets to :6379 open after connect, 0 remain after dispose().
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The store-side empty-guard consumed the per-document intentional-clear flag
INSIDE the bounded retry loop. consumeIntentionalClear always deletes the
in-memory Map entry, but a tx rollback cannot un-delete it: attempt 1
consumed the flag then updatePage threw a transient error and rolled back;
attempt 2 re-read the page non-empty, saw the flag gone, and the empty-guard
silently BLOCKED the write — dropping the user's deliberate clear and
defeating the retry guarantee for clears.
Hoist the decision out of the loop (like consumeContributors /
consumeAgentTouched): consume once into `allowIntentionalClear` before the
`for`, and only read that boolean on the empty-over-non-empty branch. The
single hoisted consume still drops a pending flag for a non-empty store
(the "cleared then retyped" case), since every store consumes regardless of
incoming emptiness.
Add a regression test: arm via the real onStateless transport, updatePage
throws once then succeeds, assert it is called twice and the retry writes the
empty doc (the clear survives). It fails on the old consume-in-loop ordering
(updatePage called once) and passes after the hoist.
Document the known fail-safe limitation near the TTL constant: if document
ownership transfers / a node crashes between the stateless signal and the
debounced store, the in-memory flag is lost and the clear is silently not
applied (the doc reloads non-empty) — fail-safe, content is never destroyed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The full-AppModule e2e (apps/server/test/app.e2e-spec.ts) passed but jest
never exited, burning CI to its timeout. Diagnosis (process._getActiveHandles
after app.close()) showed exactly two ioredis sockets to :6379 still open after
shutdown; everything else (BullMQ queues/workers, @nestjs/schedule intervals,
nestjs-ioredis, nestjs-kysely pg pool, @nestjs/cache-manager Keyv store,
hocuspocus pub/sub) already closes on app.close().
The two leaks were owned-but-never-closed clients:
1. ThrottleModule passed a pre-built `new Redis(...)` instance to
ThrottlerStorageRedisService. With an instance, the lib sets
disconnectRequired=false, so its onModuleDestroy never disconnects.
Pass ioredis options instead so the service owns + disconnects the client.
2. CollaborationGateway created a source `new RedisClient(...)` that
RedisSyncExtension only duplicates into pub/sub; the extension's onDestroy
disconnects those duplicates but not the source. Keep a reference and
disconnect it after the hocuspocus onDestroy hook in destroy().
Both are real lifecycle fixes (production shutdown is now clean too), so no
--forceExit is needed. Verified against real Postgres+Redis:
- test:e2e (no forceExit, --runInBand) exits 0 in ~18s (was: hung forever)
- --detectOpenHandles exits 0 with no open-handle report
- active handles after app.close(): none
CI timeout-minutes safety nets left untouched.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The #248 store-side empty-guard (onStoreDocument) unconditionally refuses to
overwrite non-empty persisted content with an empty document, because a
momentarily-empty live Y.Doc is indistinguishable from a real clear at the
store layer. That correctly blocks glitches/bad-merges, but also blocks a user
who genuinely wants to empty a page. This re-introduces a WORKING, narrow,
non-spoofable exception (the dead context.intentionalClear hatch #248 removed
never had a real channel).
Definition of an intentional clear (client, IntentionalClear editor extension):
a LOCAL user transaction (docChanged, NOT a remote y-sync change — filtered via
isChangeOrigin) that reduces a non-empty doc to the empty single-paragraph
shape. This is exactly the select-all + Delete/Backspace keystroke path.
Transport (option b — hocuspocus stateless message): on that transition the
client sends a `{type:'intentional-clear'}` stateless message. The server
(PersistenceExtension.onStateless) records a short-lived (TTL 60s > 45s
maxDebounce), single-use "pending clear" flag keyed by the connection's
document. The next debounced onStoreDocument consumes it on the empty-guard
branch to let that one empty write through.
Why this is the right channel and non-spoofable:
- Yjs transaction origin/metadata does not survive to the server store; awareness
is per-connection and racy. A stateless message ties the signal to a specific
clear, survives the debounce, and rides the authenticated connection.
- The document is taken from the connection, never the payload, so a client
cannot target another page.
- The flag is read ONLY on the empty-over-non-empty branch, so the worst a forged
signal can do is clear a page the connection may already edit; it can never
force or alter a non-empty write. Read-only connections cannot arm it. Every
non-empty store drops a pending flag, so "cleared then retyped" leaves nothing
usable; the flag is single-use and TTL-bounded.
NOTE: #248 is not yet on develop, so the empty-guard block is included here as
the foundation this exception extends. If #248 lands first this rebases cleanly
(the guard logic is identical; the #251-unique additions are the exception,
onStateless, the pending-flag state, and the client extension).
Tests:
- Server (real transport path, not a hand-poke): onStateless sets the flag with
the exact client payload, then the debounced onStoreDocument persists the empty
doc; plus single-use consumption, read-only rejection, non-empty-store drops
the flag, and the unchanged #248 guard tests (empty-over-non-empty blocked,
empty-over-empty allowed).
- Client: a real Editor + the actual selectAll+deleteSelection command emits the
signal; typing / non-emptying edits / already-empty docs do not.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F4: extract the reindex button `loading` predicate into a pure, unit-tested
`isReindexButtonLoading({ mutationPending, deadline, status })` next to the
other reindex helpers, replacing the inline JSX expression. Covers the
load-bearing post-cap case (deadline nulled, reindexing stale-true -> not
loading) plus mutationPending, active-run, and finished cases.
F5: rewrite the `useAiSettingsQuery` poll comment to match the actual
`nextReindexPollInterval` stop condition (continues while reindexing===true OR
within deadline and not fully indexed; stops only when reindexing===false &&
indexed>=total, or the deadline cap) instead of the stale "until indexed===total".
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F2: navigateFallbackDenylist was missing the server's `l/:alias` vanity
short-link, so a top-nav to /l/<alias> after SW registration got the
index.html app shell (which has no /l route) and dead-ended on Error404
instead of the server's 302 redirect. Add /^\/l(\/|$)/ mirroring main.ts.
F3: the partial-failure branch of "make page available offline" showed a
bare generic toast; include result.failed step labels in the message per
AGENTS.md (errors must be specific), matching the catch-branch below it.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A REST/MCP/agent rename (no live editor) wrote the new title to the
page.title column, then writePageTitle loaded the ydoc (fragment still
OLD) and set it to NEW only in memory. On disconnect onStoreDocument saw
titleText(NEW) === column(NEW), took the no-op fast-path, and never
persisted the in-memory fragment — so page.ydoc kept the OLD title and a
later body edit silently reverted the column back to OLD.
writePageTitle now persists the 'title' fragment to page.ydoc DIRECTLY
(PersistenceExtension.persistTitleFragmentYdoc) after the transact,
bypassing the no-op onStoreDocument. The write carries no treeUpdate, so
the tree WS/redis listeners do not re-broadcast (no double broadcast),
and it is idempotent/lock-free so it is safe whether or not a live editor
is connected. Adds a persist-then-reload-then-edit-body regression test
that fails on the old no-op behaviour and passes after the fix.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Apply review suggestions to the real-files block in
ai-agent-roles-catalog.provider.spec.ts (test-only):
1. Fix inaccurate comment: there are 5 content YAML files (index +
four per-bundle/lang files), not 6.
2. Improve isolation: read/parse the real index lazily inside tests
(via loadRealIndex) instead of in the describe body, so a broken
real file fails only these catalog tests, not collection of the
whole spec (incl. the unrelated mocked-remote provider tests).
3. Add the symmetric slug check: each language file's slug set must
equal the declared slug set (no undeclared/extra roles), matching
scripts/check.mjs's exact two-way correspondence.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The blob-sandbox feature (#243/#250) made AiChatToolsService.forUser()
eagerly call this.sandboxStore.asSink() while wiring the stash tool, but
the spec still passed an empty {} as the sandboxStore constructor arg.
That object has no asSink method, so all 19 tests in the suite failed in
CI with 'TypeError: this.sandboxStore.asSink is not a function'.
Replace the stale {} mock at all 4 constructor sites with a no-op sink
exposing asSink() -> { put, has, evict } (jest.fn()). These tests never
execute the stash tool, so a no-op sink is sufficient for forUser() to
wire successfully. Test-only change; production code is unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Provider tests only exercised synthetic stringifyYaml fixtures, so a
hand-conversion error in one of the 6 real catalog files (index.yaml,
bundles/{editorial,research}/{en,ru}.yaml) — a stray quote/colon in a
description, a broken emoji/arrow, a block-scalar indent slip that
silently changes or drops instructions — was caught by no automated
test. scripts/check.mjs is the only other guard and is wired into no
CI/turbo/husky step.
Add a real-files test block that reads each shipped file off disk,
parses it with the SAME options the provider uses
(strict: true, maxAliasCount: 100), and validates it through the
provider's own exported type guards (isCatalogIndex / isCatalogBundleFile
/ isCatalogRole). It is driven from the real index so new bundles/langs
are auto-covered, asserts the editorial bundle still ships fact-checker,
and requires every declared role to be present with non-empty
instructions/name in each language file.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
F1: clear the "Reindex now" spinner once the poll cap fires. Gate the
reindexing part of the button's loading state on the active poll window
(reindexDeadline !== null) so a run that outlives the 120s cap no longer
leaves the button stuck-disabled with a stale `reindexing: true`; the
admin can restart.
F2: rewrite reindexWorkspace JSDoc to describe the EMBEDDABLE page set
(text OR existing embeddings), matching getEmbeddablePageIds /
countEmbeddablePages instead of the old "every non-deleted page".
F3: extract the shared embeddable-content predicate into a private
PageRepo.embeddablePredicate helper, called by both countEmbeddablePages
and getEmbeddablePageIds, removing the verbatim duplication. Behavior is
identical (lockstep int-spec stays green).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Extract the ~110 duplicated lines into one parameterized
useImageTextFieldControl and make useAltTextControl/useCaptionControl
thin wrappers. Behavior identical; t("...") literals stay in the
wrappers so i18n extraction keeps working. sanitizeCaption still
exported for its unit test.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The comment claimed 250 groups -> 499 chars -> slice past 500; the
input is 120 "a b " groups collapsing to 479 chars, under the cap
with no slice. Correct the comment and assert the 479 length.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Mandatory (test-coverage):
- internal-file-urls.test: pin the SSRF/traversal ACCEPT path of
resolveInternalFilePath (the sole guard for content-controlled `src`): an
absolute/protocol-relative URL has its foreign host dropped and only an
/api/files/ pathname survives (http://evil.com/api/files/x/y.png -> /files/x/y.png),
while a host-dropped path that escapes /api/files/ (https://evil.com/api/auth/whoami)
or a backslash-traversal (/api/files\..\auth\whoami) is rejected. Locks the
behavior so a future prefix-only refactor cannot silently open a bypass.
Suggestions:
- index.ts: the stash_page MCP tool now returns structuredContent
{ uri, sha256, size, images } alongside the resource_link, so the MCP output
matches the documented shape (clients get the blob's sha256/ETag and the
mirror counts, not just the link). No outputSchema registered. Rebuilt build/.
- new stash-page-mcp-result.test: server round-trip via InMemoryTransport asserts
both the resource_link and the structuredContent mirror.
- internal-file-urls.test: cover the new URL parse-failure catch branch
(http://[ -> "Invalid internal file src").
- environment.service.spec: assert getPositiveIntEnv warns once per key and
independently across keys (the invalidPositiveIntWarned dedup).
Tests: packages/mcp 383 pass; apps/server sandbox/environment/mcp 235 pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Must-fix:
- mcp.module: drop the now-dead EnvironmentModule import (and its stale
comment). McpService no longer injects EnvironmentService; EnvironmentModule
is @Global and imported at the app root, so DI still resolves.
Stability:
- environment.service: route getSandboxTtlMs + the three SANDBOX_MAX_*_BYTES
caps through a shared getPositiveIntEnv() helper that warns once per key and
falls back to the default on a non-integer or <= 0 value (previously the byte
caps did a bare parseInt, so SANDBOX_MAX_TOTAL_BYTES=0 made every stash_page
fail against a 0-byte cap). TTL behavior is unchanged.
Simplification:
- sandbox.controller: replace the homemade UUID_RE with the project's shared
`uuid` validator (import { validate as isValidUUID } from 'uuid'), matching
the attachment routes; update the spec fixtures to valid v4 UUIDs.
- mcp.service: inline the single-caller one-liner buildSandboxConfig() to
this.sandboxStore.asSink() at the wiring site.
Docs:
- CHANGELOG: add an [Unreleased] > Added entry for #243 (stash_page tool,
anonymous GET /api/sb/:id, five SANDBOX_* env vars).
- AGENTS.md: note that GET /api/sb/:id is in the workspace-gate preHandler's
excludedPaths and is fully tokenless, unlike /api/files/public/... which
still resolves a workspace and needs an attachment JWT.
Tests: cap-getter validation (0/-5/abc -> default, valid -> parsed), updated
UUID fixtures. apps/server jest sandbox/environment/mcp: 233 pass.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Security (must-fix):
- sandbox.controller: the anonymous GET /api/sb/:id response now sets
X-Content-Type-Options: nosniff, a restrictive CSP, and Content-Disposition=
attachment for any mime outside a raster-image allowlist (png/jpeg/gif/webp/
avif). entry.mime is attacker-controlled, so an evil.svg/evil.html could
otherwise execute script inline on the Docmost origin (stored XSS). Mirrors
the public attachment route's hardening.
Stability:
- client.stashPage: reconcile mirrors AFTER the final document put, not only
before it. The doc blob is the newest entry and FIFO eviction drops the
oldest = this stash's own images, so the stored doc could reference an
evicted blob (consumer 404) and over-report images.mirrored. A bounded loop
now reverts doc-put-evicted mirrors, drops the stale doc blob, and re-puts
until stable. Regenerated packages/mcp/build/.
- sandbox.controller: emit Cache-Control on the 304 branch too (ttlSeconds is
computed before the conditional check).
Docs:
- Bump the MCP tool count 39 -> 40 across all READMEs and AGENTS.md (the
registry now exposes exactly 40 tools).
Refactor:
- SandboxStore.asSink() centralizes the {put,has,evict} sink + uri<->id
mapping; the embedded-MCP and in-app agent-tools wiring sites share it.
Tests:
- security headers (inline vs attachment, nosniff, CSP), 304 Cache-Control,
putAndLink URL form, has()/remove(), asSink() round-trip, getSandboxPublicUrl
(trailing-slash trim + APP_URL fallback), and a stash test where the doc put
itself evicts a mirrored image.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Security:
- stash_page: reject path-traversal / percent-encoded srcs before the authed
loopback fetch (resolveInternalFilePath), closing an SSRF/exfiltration hole
where a crafted node.attrs.src could read an arbitrary internal GET endpoint
into the anonymous sandbox.
Stability:
- stash_page: revert + recount mirrors FIFO-evicted by a later put in the same
stash (no dangling sandbox refs, honest images.mirrored/failed); free image
blobs if the final document put throws.
- Reject/clamp non-positive SANDBOX_TTL_MS to the 1h default (warn once).
- Log mirror failures unconditionally (console.warn, no blob bodies).
Cleanup / architecture:
- Remove dead expiresAt from SandboxPutResult.
- Centralize the /api/sb route in SANDBOX_ROUTE_SEGMENT/SANDBOX_API_PATH and
move URL composition into SandboxStore.putAndLink; drop the duplicated sink
closures and the now-unused EnvironmentService injection from McpService and
AiChatToolsService.
- Un-export isInternalFileUrl; document the process-local (instance-bound)
sandbox limitation in the tool description and .env.example.
Docs/tests:
- README/README.ru: 38 -> 39 tools + stash_page entry.
- Add traversal/normalize/recursion unit tests, stash self-eviction +
doc-put-throw + empty/octet-stream mock tests, controller If-None-Match
(wildcard/weak/list) + Cache-Control tests, and SANDBOX_TTL_MS validation
tests. Regenerate packages/mcp/build.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Address six QA findings on the offline-sync feature:
1. HIGH — silent data loss: a paused mutation persisted to IndexedDB and
reloaded while still offline never resumed on reconnect. Seed TanStack
onlineManager from navigator.onLine at boot (it defaults to online:true and
only flips on events, so a cold-boot-offline tab wrongly believed it was
online and never got a true online transition), and call
resumePausedMutations() in PersistQueryClientProvider onSuccess after the
persister rehydrates (defaults are registered before, so the restored
mutation has a mutationFn). New offline-resume.test.ts reproduces the full
persist -> reload -> reconnect path.
2. MEDIUM (security) — logout did not durably clear gitmost-rq-cache: the
throttled persister re-wrote the key ~1s after del() with the still-in-memory
snapshot, resurrecting the previous user's data. Freeze the persister
(persistClient becomes a no-op) before clearing/deleting so neither the
clear()-triggered nor any in-flight write can repopulate the key; re-enable
afterwards for the next sign-in session.
3. MEDIUM (UX) — offline create spun forever: the create-note button awaited a
mutateAsync that stays pending while paused. Detect offline, fire-and-forget
the (queued) mutation, show a "saved offline" notice, and gate the spinner on
!isPaused so it no longer hangs.
4. LOW — an uncached page opened offline showed the generic "Error fetching page
data." instead of the offline fallback (offline fetch yields no HTTP status).
Render OfflineFallback when navigator is offline or the error has no status.
5. LOW — logout teardown threw "Cannot read properties of null (reading
'settings')" in full-editor.tsx: optional-chain the (transiently null) user.
6. Tab title "Untitled": investigated — the tab-title derivation in page.tsx is
byte-identical to develop and already reads page.title from REST/cache (the
recommended source); live edits keep it in sync via updatePageData. Not a
tab-title-derivation regression introduced by this PR; no change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
CHANGELOG: stop presenting the service-worker API cache as an active offline
store (/api is NetworkOnly) — describe it as a defensive purge of the legacy
api-get-cache from older clients; add an explicit upgrade note that the new CORS
allowlist rejects previously-allowed cross-domain REST clients until their origin
is added to CORS_ALLOWED_ORIGINS.
test(offline): cover make-offline ancestor-walk + dedup — a real-ancestor case
exercising the ancestorId===pageId guard (page warmed once), the dedup of
repeated tree failures into a single "tree" label, and the "breadcrumbs" label
when the breadcrumbs lookup rejects.
test(auth): cover clearOfflineCache in handleLogout — purged exactly once before
window.location.replace, and a thrown purge error does not block the redirect.
conventions: use pageKeys.detail() instead of raw ["pages", …] literals in
title-editor and use-page-collab-providers.
cleanup: remove the dead emit() in title-editor (the gateway ignores it; the
cross-user tree refresh is server-side via the Yjs title fragment); drop the
trivial Array.isArray(tiptapExtensions) test (schema is exercised transitively).
refactor: extract the shared page.<id> Yjs doc-name convention into
pageYdocName()/PAGE_YDOC_NAME_PREFIX so the editor providers, offline warm, and
offline purge can no longer drift apart.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Security:
- Clear the offline IndexedDB cache on sign-in (not only logout) so a previous
user's persisted query cache and Yjs page bodies cannot leak to the next user
on a shared device when the prior session ended without an explicit logout.
Regressions:
- Remove the double Yjs title write from the AI title-generation path: the title
editor is bound to the Yjs `title` fragment and the server REST update reseeds
it, so the local setContent raced that reseed and doubled/garbled the title.
Conventions / i18n / docs:
- Remove the unused showAiMenuAtom.
- Register the 3 offline-fallback strings in en-US and ru-RU.
- Fix the 5 broken links to the nonexistent docs/offline-sync-plan.md.
Stability / simplification:
- warmInfiniteAll now reports truncation (returns false) when it hits maxPages
with a cursor still pending instead of silently succeeding.
- space-tree make-offline catch logs the raw error and surfaces the real cause.
- Move the Offline/Mobile/CORS CHANGELOG entries from the released 0.93.0 section
into [Unreleased] (CORS is a documented breaking change).
- Drop the pass-through sync-flag forwarders in use-page-collab-providers; set the
atoms directly.
- Collapse the three isSwaggerEnabled true-cases into it.each.
Tests / architecture:
- Extract collabTokenNeedsRefresh (pure) and cover all four token states.
- Extract shouldPropagateTitleChange and cover the collab-origin skip; add a
TitleEditor render test for the static-h1 vs collaborative-editor switch.
- Add a use-auth test asserting the sign-in cache purge runs before login.
- Add an OFFLINE_PERSIST_ROOTS guard test asserting every persisted root maps to
an exported query-key factory; route make-offline's currentUser warm through a
new userKeys factory.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fixes the offline-sync defects QA found on PR #120 (#237/#238/#220).
Blank-shell / white-screen on offline reload (HIGH):
- auth-query.tsx: the useCollabToken retry predicate read
`error.response.status` unguarded. Offline the collab-token POST rejects as
an axios NETWORK error (isAxiosError true, response undefined), so `.status`
threw an uncaught TypeError in the React Query retryer BEFORE React mounted,
white-screening every route. Extracted the predicate as `collabTokenRetry`
and guarded it with optional chaining (`error.response?.status === 404`).
- user-provider.tsx: gated the whole <Layout> on useCurrentUser() and returned
a bare `<></>` on any error, blanking every authenticated route offline even
when cached data existed. Now renders the cached app when a (stale) user is
present and an explicit OfflineFallback when there is no user to fall back on.
- query-persister.ts / make-offline.ts: persist and warm the ['currentUser']
query so the auth gate can hydrate offline (pinned pages now survive relaunch).
Offline structural create/move/comment silently lost on reload (HIGH):
- offline-mutations.ts: register setMutationDefaults (default mutationFns) for
stable mutation keys and tag useCreatePageMutation / useMovePageMutation /
useCreateCommentMutation with those keys. A paused mutation dehydrated to
IndexedDB while offline now has a mutationFn after reload, so
resumePausedMutations() replays it on reconnect instead of no-op'ing.
Tests (client vitest): collabTokenRetry no longer throws on a no-response
network error; UserProvider renders cached children / the offline fallback (not
a blank fragment) on a network error; a rehydrated paused create/move is
replayable via resumePausedMutations; currentUser persist-root coverage.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Service worker (vite-plugin-pwa/Workbox): add /share/, /mcp, and /robots.txt
to navigateFallbackDenylist so the SPA app-shell never shadows those
server-rendered routes (they mirror the server static-serve exclude list — the
share SEO/OG HTML, the MCP endpoint, and robots.txt must come from the server).
- Remove the dead /api GET NetworkFirst Workbox rule (api-get-cache): offline
reads are served by the persisted TanStack Query cache (IndexedDB) + y-indexeddb,
never by an SW HTTP cache, so caching GET /api only risked stale responses. All
/api is now NetworkOnly. clearOfflineCache still deletes any legacy api-get-cache
defensively (comment updated to note it is no longer created).
- CORS: drop the cleartext 'http://localhost' native-WebView origin. The Capacitor
shell uses the secure scheme (capacitor.config cleartext:false, default Android
scheme https, iOS hosted via CAP_SERVER_URL), so no native client uses it;
allowing it only widened the credentialed-CORS surface. Keeps capacitor://,
ionic://, and https://localhost.
- docs/mobile-bootstrap.md: replace the inaccurate 'hand-rolled service worker'
description with the real Workbox generateSW setup (prompt registration via
virtual:pwa-register, production-only, denylist, NetworkOnly, RQ/y-indexeddb
offline reads) and drop http://localhost from the CORS origins list.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The 'Make available offline' warm path re-typed React Query key literals and
re-declared queryKey+queryFn pairs that the feature hooks already owned, so the
two could silently drift (a hook key change would leave the warm cache under a
stale key). Centralize them so there is one source:
- Add pageKeys (page-query.ts) and spaceKeys (space-query.ts) key factories and
route the inline key literals through them. Partial-match keys and 2-element
spaceMembers invalidations are deliberately left inline so their effective key
VALUE (and invalidation breadth) is unchanged.
- Add queryOptions factories sidebarPagesQueryOptions and spaceByIdQueryOptions,
consumed by both the hooks (fetchAllAncestorChildren, useGetSpaceBySlugQuery)
and the warm path. Comments reuse the existing RQ_KEY factory.
The warm path also stops silently succeeding: warmInfiniteAll returns a boolean
and logs failures; makePageAvailableOffline is best-effort (never throws) and
returns { ok, failed[] }, recording each failed step by label; the tree menu
caller now shows a success or error toast from result.ok. Removed the unused
slugId/parentPageId params from the offline params type.
This is a behavior-preserving centralization: effective query keys, queryFns,
staleTime and enabled are unchanged for every hook.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The onAuthenticationFailed handler is created once per page (effect keyed on
pageId), so it closed over the initial collab token and decoded a STALE value
after a refetch. Worse, jwtDecode(undefined) throws, so when the token had not
loaded (or the request failed) the handler crashed before it could refetch and
reconnect — leaving the editor stuck disconnected.
Mirror the latest token into a ref the handler reads live, and guard the decode:
a missing or malformed token is treated as 'needs refresh' so it refetches and
reconnects instead of throwing. A valid, unexpired token still early-returns.
Also remove two local useState sync flags (isLocalSynced/isRemoteSynced) that
were set but never read — the header indicator consumes the Jotai atoms, and the
hook's return values were never destructured by any caller. The setter wrappers
now drive only the atoms.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
In 2-process deployments (COLLAB_URL set) the standalone collab process runs
Hocuspocus onStoreDocument, which emits PAGE_UPDATED with a treeUpdate snapshot
on a collaborative rename. But CollabAppModule has no WsModule, so PageWsListener
(the broadcaster) only exists in the API process — the collab-originated tree
update never reached clients, and other users' sidebars/breadcrumbs went stale.
Bridge it over Redis pub/sub with the API process as the single broadcast
authority:
- PageTreeBridgePublisher (registered ONLY in CollabAppModule) listens for
PAGE_UPDATED and, when a treeUpdate snapshot is present, publishes it to the
collab:tree-update channel. Gated exactly like PageWsListener so content-only
saves never publish noise.
- PageTreeBridgeSubscriber (registered in WsModule, API process) subscribes on a
dedicated duplicated connection and re-broadcasts each snapshot through
WsTreeService.broadcastPageUpdated — the same restriction-aware emitTreeEvent
path, so authorization is preserved.
Double-broadcast is prevented by module placement: the publisher lives only in
the standalone collab process's root module, so in single-process mode it is
never loaded and the local PageWsListener stays the sole broadcaster.
The bridge is optional and fail-safe: publish errors, malformed payloads,
broadcast rejections, an unlistened 'error' on the subscriber connection, and a
subscribe() failure at boot are all caught and logged, never crashing or blocking
the process. NOTE: assumes a single API broadcaster; horizontal API scaling would
need a consumer-group/leader-election instead of fan-out pub/sub.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
onLoadDocument rebuilds a legacy page (page.content, no page.ydoc) into a Yjs
doc and seeds its 'title' fragment from the page.title column. Both
TiptapTransformer.toYdoc and buildTitleSeedYdoc mint fresh Yjs client-ids on
every call, so the heal must run exactly once per page. Three holes let it run
twice (or lose a write):
- Duplication trap: the initial page read took no row lock, so two processes
(the API process via openDirectConnection and the standalone collab process)
could both observe ydoc IS NULL and each rebuild with different client-ids; a
long-offline client merging an earlier rebuild then duplicates all content.
- Lost-update: persistYdoc wrote updatePage({ydoc}) outside any transaction, so
it could clobber a concurrent onStoreDocument write (which does take a lock).
- Swallowed write errors: a failed heal-persist was logged but the unpersisted
fresh-client-id doc was returned anyway, silently re-arming the trap.
Fix: the heal now runs in healUnderLock, which re-reads the row FOR UPDATE inside
one transaction and re-validates under the lock — if ydoc is now present it
adopts it (no rebuild, no write), otherwise it rebuilds, seeds, and persists the
ydoc in the SAME transaction. The healthy hot path still loads with no lock and
no write. Failure handling surfaces instead of hiding: a rebuild-persist failure
refuses the load (re-throw + error log) so an unpersisted rebuild is never handed
out, while a seed-only persist failure serves the existing healthy ydoc without
the unpersisted seed (non-fatal). Removed the non-transactional persistYdoc.
Deliberately does NOT use a fixed clientID: identical client-ids across docs
built from differing content violate Yjs per-actor uniqueness and corrupt worse
than the trap; serialization under the row lock is the correct fix.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Title now lives in the page's Yjs 'title' fragment, but two paths corrupted it:
- Rename-revert: a REST/MCP title change wrote only the page.title column,
never the Yjs fragment, so the next editor open replayed the stale Yjs title
and reverted the rename. PageService.update now mirrors the new title into the
Yjs 'title' fragment via CollaborationGateway.writePageTitle, which goes
through openDirectConnection directly (Redis-independent: works with
COLLAB_DISABLE_REDIS and in single-process deployments, unlike the
Redis-routed handleYjsEvent path). The write is best-effort: a Yjs failure is
logged and never rolls back the committed column write. Agent provenance
(actor/aiChatId) is threaded into the store context.
- Untitled-on-open: an empty/just-initialized 'title' fragment clobbered a
non-empty page.title to '' on open. onStoreDocument now treats the title as
changed only when the extracted text is non-empty, covering both the
title-only and body+title save branches. Empty-retitling via collab is
intentionally impossible; the REST DTO is the place to enforce non-empty.
writeTitleFragment does a full clear+seed of the 'title' fragment (no
duplication/concatenation) and leaves the body fragment intact. Removed the dead
useTreeMutation.handleRename path. Adds unit tests for writeTitleFragment, the
gateway write, the anti-empty-clobber guard, and agent provenance.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
claude code agent 227
vvzvlad
a
