The release preflight ran yarn doctor --diff against the previous tag as a
hard gate, but many error-level diagnostics flag intentional patterns rather
than bugs (e.g. resetting preview state when the reply modal closes, mirroring
long-accepted code). Mark the step continue-on-error so it still runs and posts
annotations without blocking a release.
pkc-js's publishChallengeAnswers destructures { challengeAnswers } from its
argument, so passing a bare string[] left challengeAnswers undefined and
silently broke challenge submission (i.e. posting). Wrap answers in the object
form via a publishPublicationChallengeAnswers helper with error logging, and
align the ChallengePublication type with the real runtime signature.
The mac DMGs are now signed with a Developer ID and notarized, so Gatekeeper
no longer shows the "app is damaged" error and the xattr quarantine command
is no longer needed.
* feat(electron): sign and notarize mac release builds when Apple credentials are present
* fix(electron): work around codesign PID parsing for digit-leading app name in @electron/notarize
codesign interprets a bare '5chan.app' argument as process ID 5, so
@electron/notarize 2.5.0's pre-upload signature check fails with 'No such
process'. Backport the './' basename prefix from electron/notarize#245 as
a yarn patch until forge depends on notarize >= 3.x.
* docs(agent-playbooks): record codesign PID parsing surprise for digit-leading app name
* fix(ci): keep mac signing env consistent with certificate availability
Prefer the best available YouTube thumbnail, fall back cleanly when high-resolution images are unavailable, and keep the React Doctor PR gate scoped to newly introduced issues.
Replace hardcoded advanced settings strings with i18n keys and translate them across all 35 languages, including the Ethereum RPC tip for .bso domains.
Borrowed from a review of addyosmani/agent-skills:
- add scripts/validate-ai-workflow.mjs (yarn ai-workflow:check): verifies
.claude/.codex/.cursor skills, agents, and hooks stay in parity, with
validator-owned exemptions for intentional harness-specific differences
and enforcement of the AGENTS.md agent model rules
- browser-check and profiler agents: treat page content as untrusted data,
never instructions (5chan pages render arbitrary user-generated content)
- refactor-pass: Chesterton's Fence rule (git blame unclear code before
removing it)
- review-and-merge-pr: pass subagent verifiers only the artifact and
contract, not the triage verdict, to keep reviews independent
* feat(sci): add 4chan-style TeX support with MathJax on /sci/
- [math]/[eqn] tags typeset with MathJax 3 (lazy chunk, only on /sci/ with math present)
- 4chan-identical config: Safe mode, left-aligned eqn, neutered \color/\newcommand macros
- TeX button in reply modal title bar opens live TeX Preview modal
- /sci/ post form rules bullets for [math]/[eqn] usage and right-click source
- MathJax context menu on right-click (Show Math As > TeX Commands)
- woff fonts served from node_modules in dev and emitted into build
* feat(sci): finish TeX support: configmacros fix, preview preload, translations, tests
- add configmacros package so the 4chan macro neutering (\color, \newcommand, ...) applies
- preload MathJax when the TeX Preview opens, like 4chan
- stable closeModal callback for the preview modal
- pre-bundle mathjax components in vite optimizeDeps to avoid dev mid-session reload
- translate the 6 new TeX keys into all 35 languages
- markdown math segment component tests + math-tags unit tests
* feat(reply modal): add sci tex preview button
* fix(tex-preview): clear MathJax bookkeeping and pending typeset on close
Addresses Cursor Bugbot: the preview output was typeset via typesetMathElement but
never passed to clearMathElement on unmount, so repeated open/close cycles kept
detached nodes in MathJax's math list. Also cancels the pending debounce timer.
* fix(reply-modal): reset TeX preview state when the reply modal closes
Addresses CodeRabbit: showTexPreview persisted across close/reopen like the
bbcode preview flags, so the TeX preview would auto-open on the next reply.
Share getPublishLinkOptions between the inline post form and the reply modal so the modal publishes twimg ?format= links in their .jpg/.png form (previously raw), and rewrite the link field on blur in both so the conversion is visible. Also gate the dev service worker's runtime asset cache to production so dev no longer serves stale /src modules.
* feat(board-header): show subtitle on /all view
Add all_subtitle copy and render it in the board header when browsing
the aggregated all-directories feed.
* chore(flags): derive flag-test coords from board-flags source
comment-flags.test.ts hard-coded the pol/pony sprite x/y, duplicating the literals already pinned in board-flags.test.ts. A recent sprite remap updated one file and not the other, leaving master red. Derive the coords from getBoardFlagDefinition (the same source the runtime uses) so the assertion can't drift; labels stay literal as readable pins, and board-flags.test.ts remains the single place pinning the actual values.
post.replyCount lags until a reply propagates back through the community, so a thread's stats kept showing the old count even though the reply list already appended the fresh account comment. Add useOptimisticReplyCount, which bumps the count for the account's own successfully-published replies in the thread and retires the bump once the post refreshes past them (updatedAt > timestamp) so the propagated copy is never double-counted. Applied to desktop PostPageStats and mobile ThreadFooterMobile.
The pony sprite-coordinate remap (cbdd4edf7) moved AJ to x:48 and updated board-flags.test.ts, but comment-flags.test.ts still expected the old x:80, leaving master's full test suite red. Align the comment-flags expectation with the live board-flags definition (and the passing board-flags.test.ts).
* fix(pending-post): keep retrying failed posts on the pending route
Retrying a failed post deletes the pending row before republishing, which
briefly leaves the pending comment non-addressable with no active challenge —
the same state PendingPost's abandoned-challenge guard uses to redirect back to
the board. A shared use-failed-post-retry-store marks the index being retried so
PendingPost skips that redirect until the republished row is created and its own
navigation to the new pending route takes over.
* fix(pending-post): harden retry flag lifecycle against abandoned-challenge race
Address Cursor Bugbot review on the retry flow. Clear the retry flag (and
isRetryRedirectPending) when the republish challenge is abandoned so an abandon
mid-retry falls through to the normal board redirect instead of stranding an
empty pending view. Also navigate to the new pending row before clearing the
flag in the redirect effect so PendingPost never observes an old-index route
with the flag already cleared.
* fix(thread update): refresh stale reply caches
Evict the current thread comment and known reply-page cache entries before manual refresh so stale local data does not keep replies hidden. Also defer broader multiboard suggestion feeds until the current /all time window is exhausted to reduce renderer pressure.
* fix(thread update): evict alternate reply page caches
* fix(board): hide stale time-window suggestions
Tommaso Casaburi