* Harden interrupted recovery control filtering
* Redesign live-to-final assistant replies
* Fix live activity anchor test fixture
* Fix CI lint issues for live reply tests
* Strengthen live progress prompt contract
* Recover PR #3401 refresh on origin/master
* Repair live-to-final refresh regressions
* Fix live worklog refresh regressions
* Show live footer timer on initial stream start
* Restore live stream shell after reload
* Preserve per-frame live SSE replay cursors
* Preserve reasoning as Worklog Thinking cards
* Quiet Worklog Thinking card styling
* Align Worklog Thinking card styling
* Scope live Worklog Thinking cards by segment
* Suppress exact duplicate settled Thinking
* Close#3401 merge review test gaps
* fix(#3401): resolve 4 deep-review regressions (inline-think, reconnect-dup, neon skin, busy-gate worklog)
Deep review (Codex diff-vs-master + live-browser drive) of the live-to-final refactor
surfaced 4 regressions vs master that the rewritten suite no longer guarded:
1. Inline <think>…</think>answer reasoning vanished — _assistantReasoningPayloadText
used $-anchored regexes so a leading think block + visible answer extracted nothing
and the Thinking card never rendered. Removed the 3 $ anchors to match the
(non-anchored) display stripper. Live: inline-think thinking-only turn now renders.
2. (CORE) reconnect/reload duplicated the live reply — _rememberRunJournalCursor advanced
a closure-local seq but never wrote INFLIGHT[activeSid].lastRunJournalSeq, so a reload
replayed the journal from after_seq=0 over restored lastAssistantText. Now mirrors the
cursor onto INFLIGHT + schedules a throttled persist.
3. Neon skin silently broke — PR deleted the :root[data-skin="neon"] CSS but left Neon in
the picker. Restored the neon CSS block from master.
4. Settled tool-worklog rebuild gated purely on !S.busy — dropped every prior settled
turn's worklog when renderMessages re-ran during an active stream (switch-back to an
in-progress session). Restored master's !S.busy || (S.toolCalls && S.toolCalls.length).
Live: busy re-render now preserves tool cards (4→4, was 4→0).
Live-verified all 4 + confirmed #3709/#3592 invariants still hold (1 thinking card, none
below footer; distinct siblings preserved). + tests/test_issue3401_deep_review_fixes.py (7).
* test(#3401): realign 3 stale source-shape assertions to the deep-review fixes
Fix commit changed two source literals that existing stage tests scanned for:
- test_live_activity_timeline.py (x2): split anchor 'if(!S.busy){' → the restored
'if(!S.busy || (S.toolCalls&&S.toolCalls.length)){' guard (fix 4).
- test_run_journal_frontend_static.py: 'after_seq=0' not in source — fix 2's comment
contained that literal; rephrased the comment to 'the zero floor (after_seq of 0)'.
Intent of all three assertions unchanged; only the matched string updated. No code
behavior change.
* docs(changelog): v0.51.294 — Release JJ (stage-3401, #3401 live-to-final redesign)
---------
Co-authored-by: Frank Song <franksong2702@gmail.com>
Co-authored-by: Nathan-Hermes <nesquena-hermes@users.noreply.github.com>
Co-authored-by: nesquena-hermes <[email protected]>
* fix(#3709): thinking card no longer renders twice (in Activity + below answer)
The #3592 inline-render branch (v0.51.258) emitted a thinking card for a
thinking-only message even when a sibling tool-message in the same turn already
built an Activity group carrying that turn's thinking — so the card showed twice,
the second one stranded below the answer + 'Done in …' footer (insertAdjacentHTML
'beforeend' on a segment that already had body+footer).
Fix (keeps #3592, does NOT revert it):
- A1: precompute turnsWithActivityGroup (turns whose segments have tool cards);
the inline branch only renders when the anchor turn is NOT in that set.
- A2: when it does render inline, insert 'beforebegin' the .msg-body/.msg-foot so
the card sits above the answer, not orphaned below the footer.
- B: strip thinking against the TURN's combined visible answer
(_turnVisibleTextByRawIdx), so a trailing thinking-only message that echoes the
answer gets de-duped even though its own body is empty.
Live-verified in browser: #3709 repro (tool+trailing-thinking) → exactly 1 card in
Activity, above footer; #3592 repro (thinking-only) → exactly 1 inline card, not
buried in a collapsed group. + regression test tests/test_issue3709_*.
Supersedes #3708 (which deleted the inline branch outright, re-breaking #3592).
* fix(#3709): merge suppressed sibling thinking into the Activity group (Codex re-gate)
Codex caught a content-loss edge in the first cut: when A1 suppresses a
thinking-only sibling's inline card (its turn has an Activity group), the group
only rendered assistantThinking.get(aIdx) for the TOOL message — so a sibling
with DISTINCT reasoning was neither inline nor in the group → dropped.
Fix: aggregate all of a turn's thinking (turnThinkingParts, de-duped, index
order) and render that merged text once per turn in the Activity group
(_renderedTurnThinking guard). Live-verified: tool-thinking A + distinct
sibling-thinking B → 1 merged node carrying both, no loss. + regression test.
* fix(#3709): shared anchor resolver so inline-suppression & group placement agree (Codex re-gate #2)
Codex caught a fallback-anchor mismatch: turnsWithActivityGroup was populated only
from assistantSegments.get(tcIdx) (direct segment), but the group-render path falls
back to a nearby earlier segment when a tool's assistant_msg_idx has no directly
rendered segment (legacy/rebased). So a fallback-anchored group's turn wasn't in
turnsWithActivityGroup → the sibling rendered inline AND the group rendered → dup
again. Fix: one shared _anchorRowForActivityIdx(aIdx) helper (direct-or-fallback)
used by the precompute, the inline branch, and the group render — they now agree.
Live-verified all three repros still pass.
* test(#3709): update test_compact_activity assertion to mergedThinking var
The brittle source-scan asserted _thinkingActivityNode(thinkingText, false) — the
#3709 fix renders the turn's MERGED thinking via _thinkingActivityNode(mergedThinking,
false) into the same Activity body. Intent (settled thinking renders inside the
Activity disclosure alongside tools) unchanged; only the source variable. Updated to
assert the new variable, kept all intent assertions.
---------
Co-authored-by: nesquena-hermes <[email protected]>
* fix(#3315): surface compression-exhausted/no-final-answer turns as errors (#3316)
When Hermes Agent exhausts context compression in a long tool-heavy turn, the
streamed result can end on a tool result / assistant(tool_calls) turn with no
final assistant answer. WebUI was finalizing that as a completed response.
Now _session_lacks_final_assistant_answer() + _agent_result_terminal_failure()
classify these as terminal failures and surface an apperror instead. The
compression session-id migration + pre-compression snapshot now run BEFORE the
terminal-failure return (ordering bug from the prior hold) so state stays
consistent when exhaustion fires after the agent rotated session_id.
Co-authored-by: Frank Song <franksong2702@gmail.com>
* docs(changelog): v0.51.292 — Release JH (stage-s4, #3316fixes#3315)
---------
Co-authored-by: Frank Song <franksong2702@gmail.com>
Co-authored-by: nesquena-hermes <[email protected]>
* Fix update reload readiness race — poll /health server identity before reload (#3654)
Replaces the raw-uptime comparison (couldn't distinguish a fresh old process
from the restarted one) with a stable server_started_at identity read before
the update POST; reloads only when the identity changes. Both the force-update
and regular apply paths read + pass the baseline. (#874, #3654)
Co-authored-by: Frank Song <franksong2702@gmail.com>
* docs(changelog): v0.51.285 — Release JA (stage-r19)
---------
Co-authored-by: Frank Song <franksong2702@gmail.com>
Co-authored-by: nesquena-hermes <[email protected]>
* feat(composer): surface that messages queue during auto-compaction (#3512, #3079)
Co-authored-by: Rod Boev <rod.boev@gmail.com>
* docs(changelog): v0.51.283 — Release IY (stage-w2)
* fix(composer): restore placeholder on ALL compaction-exit paths, not just clearCompressionUi (#3512)
Codex+Opus both caught: setCompressionUi(done) and the live-anchored SSE
window._compressionUi=null paths bypassed clearCompressionUi, leaving the
'will queue' placeholder stuck after compaction. Factor restore into
_restoreCompressionPlaceholder() + call from every compaction-exit path.
---------
Co-authored-by: nesquena-hermes <[email protected]>
Co-authored-by: Rod Boev <rod.boev@gmail.com>
Absorbs contributor PR #3544 (@rodboev, closes#3340) with two fixes:
1. DETECTION VOCAB (would never fire): the original gated on action names
{save,create,update,upsert}, which don't match the real agent tool enums —
memory.action is add|replace|remove, skill_manage.action is
create|patch|edit|delete|write_file|remove_file. Split into per-tool
predicates with the correct vocabularies: _isMemorySave gates memory on
{add,replace}; _isSkillUpdate gates skill_manage on {create,patch,edit,
write_file}. Deletions excluded so the saved/updated verbs stay accurate;
running/errored excluded.
2. SNAPSHOT/RESTORE PERSISTENCE (Codex catch): classification lived only on the
row._tcData JS property, which does NOT survive the outerHTML/innerHTML
snapshot+restore the live tool-call group uses on session switch/restore —
a restored memory/skill row would be re-counted as a generic tool and the
suffix would silently vanish. buildToolCard now also stamps durable
data-memory-save / data-skill-update attributes, and _syncToolCallGroupSummary
counts them as a fallback when _tcData is absent. Verified live across a real
outerHTML round-trip: label identical before/after.
Replaces the PR's static source assertions with a node-driven behavioral test
(11 cases) covering the real action vocabularies, exclusions, case-insensitivity,
null-arg safety, and the durable-attribute persistence guard.
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
## Release v0.51.266 — Release IH (stage-r16)
One agent-authored APPROVED fix + two un-held streaming/SSE fixes.
### Fixed
| Issue/PR | Author | Fix |
|----------|--------|-----|
| #3635 (#3637) | @nesquena-hermes (nesquena APPROVED) | Composer profile chip reads `S.activeProfile` again — a #3331 regression keyed it on the loaded session's profile, so opening a cross-profile session made the chip disagree with the dropdown checkmark and misrepresent where the next message routes. #3331's project/session-op scoping is unaffected. |
| #3587 (#3605) | @rodboev | Reasoning persists to the correct intermediate assistant message in multi-turn tool flows. The index only advanced in `on_interim_assistant` (suppressed for contentless tool-call messages) → post-tool reasoning was mis-attributed; it now also advances at the `on_tool` boundary, guarded against over-increment. **(un-held — finding resolved)** |
| #2660 (#3558) | @franksong2702 | Session-event SSE no longer wakes every tab across profiles and never drops a relevant refresh — profile attached when known, root/`default` aliases stay unscoped, and the `maxsize=1` queue falls back to unscoped refresh-all on a profile-mismatch coalesce. **(un-held — both findings resolved)** |
### Gate
- Full pytest suite: **7770 passed, 0 failed**
- ESLint: CLEAN · ruff: CLEAN · browser-smoke: CLEAN
- Codex (regression): **SAFE TO SHIP** — chip matches dropdown/routing (no #3331 scoping regression), reasoning-index advance composes with the agent's tool/interim callback ordering, session-events coalesce safely with no dropped refresh and no profile data leak (`/api/sessions` still server-side filtered).
Co-authored-by: nesquena <nesquena@users.noreply.github.com>
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
Co-authored-by: franksong2702 <franksong2702@users.noreply.github.com>
## Release v0.51.261 — Release IC (stage-r11)
Live Todos panel via an explicit `todo_state` SSE contract.
### Fixed
| Issue/PR | Author | Fix |
|----------|--------|-----|
| #3373 follow-up (#3454) | @v2psv | The Todos side panel now tracks `todo` tool state **live during an active run** instead of staying stale until settle / rolling back on a mid-stream reload. A dedicated `todo_state` SSE event sends a full, redacted, idempotent snapshot on todo-tool completion (no more truncated `tool_complete.preview`); the same `api.todo_state` parser feeds live + cold-load; live snapshots persist into INFLIGHT so reload/reattach restores the panel; cold-load vs INFLIGHT reconciled by timestamp (incl. the `coldTs===0` compressed-session edge); legacy reverse-scan kept as fallback for old servers. |
### Gate
- Full pytest suite: **7692 passed, 0 failed**
- ESLint: CLEAN · ruff: CLEAN · browser-smoke: CLEAN
- Codex (regression): **SAFE TO SHIP** — verified the new `todo_state` SSE handler composes with existing dispatch (no double-subscribe), INFLIGHT persistence is cleared on terminal/cancel (composes with discard_session + turn-journal), timestamp reconciliation can't let a stale local snapshot win, redaction holds, the legacy reverse-scan fallback still works with no double-render, and the `models.py` change is todo-scoped (no CLI-classification interaction).
Co-authored-by: v2psv <v2psv@users.noreply.github.com>
## Release v0.51.258 — Release HZ (stage-r6)
Fresh-arrival low-risk pair (both @rodboev, landed in the last sweep window).
### Fixed
| Issue | Fix |
|-------|-----|
| #3597 | The "update available" banner now shows from **any panel** (Settings → System "Check now", etc.), not just the Chat view — it was positioned inside the chat surface so it only rendered there. |
| #3592 | Under Simplified Tool Calling, an assistant turn with **thinking but no tool calls** now renders that thinking inline on settlement instead of burying it in an empty collapsed activity group. |
### Review fix absorbed (Codex)
#3592's inline-render `continue` skipped the activity-group creation that carried the turn's `data-turn-duration`, but the footer still suppressed the "Done in …" duration for any `assistantThinking` turn → thinking-only turns silently lost their duration display. Fixed: footer duration is now suppressed **only** for turns that actually build an activity group (`toolCallAssistantIdxs.has(mi)`), so thinking-only inline turns keep "Done in …". + regression test.
### Gate
- Full pytest suite: **7631 passed, 0 failed**
- ESLint: CLEAN · browser-smoke: CLEAN
- Codex (regression): SHIP-ONLY-WITH-FIXES (duration-drop) → fixed → **SAFE TO SHIP**
### Sweep note
#3603 + #3604 (sidebar CLI-session classification, same author/area) were **not** included — they assert contradictory models for a sidecar-less `source='cli'` recovery row; flagged on both PRs for the author to reconcile.
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
## Release v0.51.254 — Release HV (stage-r2)
Phase-2 medium wave 1 — 4 PRs (UI/mobile/cancel fixes + an un-held model dedup).
### Fixed
| Issue/PR | Author | Fix |
|----------|--------|-----|
| #3528 | @franksong2702 | Render partial tool calls after cancel — interrupted turns keep their `_partial_tool_calls` rows in the transcript + fallback tool-cards. (Codex confirmed it stays render-only, not forwarded to the provider API.) |
| #3550 | @lurebat | Android offline recovery soft-reattaches the live stream instead of hard-reloading the page on a transient background/disconnect. |
| #3479 | @mvanhorn | iOS Safari no longer snaps the conversation to the top when a handoff/compression card is inserted mid-stream or on `refreshSession()`. |
| #3478 | @JayC-L | **Un-held:** named custom providers (`@custom:name:model`) dedup against bare model IDs without regressing Ollama multi-colon tags (`qwen2.5:7b-instruct-q4`). Only `@custom:` IDs strip the two-segment prefix. |
### Hold-sweep note
#3478/#3489 was held yesterday for an Ollama multi-colon-tag regression risk (a blanket `lastIndexOf` would lose the model). The author pushed a scoped fix (only `@custom:` IDs use `lastIndexOf`); I verified `_normId` in node against the regression cases — Ollama bare tags are preserved. Un-held + shipped.
### Gate
- Full pytest suite: **7588 passed, 0 failed**
- ESLint: CLEAN · ruff: CLEAN · browser-smoke: CLEAN
- Codex (regression): **SAFE TO SHIP** — #3552 partial-tool-calls verified render-only (no `_API_SAFE_MSG_KEYS` leak / no 400-on-strict-provider, the v0.50.251 #1375 trap); #3551 no EventSource double-subscribe; #3541 no regression vs the #3525 scroll-follow shipped in v0.51.253; #3489 no over-dedup.
Co-authored-by: franksong2702 <franksong2702@users.noreply.github.com>
Co-authored-by: lurebat <lurebat@users.noreply.github.com>
Co-authored-by: mvanhorn <mvanhorn@users.noreply.github.com>
Co-authored-by: JayC-L <JayC-L@users.noreply.github.com>
## Release v0.51.245 — Release HM (stage-q17)
Small UX bug-fix.
### Fixed
| Issue | Author | Fix |
|-------|--------|-----|
| #3338 | @rodboev (#3502) | **Messaging sessions (Telegram, Discord, WeChat, …) now show their platform source badge in the chat-pane topbar**, not just the sidebar. The topbar badge was gated on `is_cli_session` (intentionally `false` for messaging sources), so it vanished once the session opened. Gate removed; a recovered native session stamped `source_label:"WebUI"` stays un-badged. Reuses the existing `.topbar-source-badge` styling — no new chrome. |
Picked #3502 over the duplicate **#3499** (same issue/files) — #3502 adds the `WebUI` self-source suppression and a stronger regression test. #3499 closed as superseded with credit.
### Gate
- Full pytest suite: **7543 passed, 0 failed** (first run hit the known boot-cascade flake — 376 connection-refused across 29 files; clean on re-run, as expected for a JS-only change)
- ESLint: CLEAN · browser-smoke: CLEAN · 8 source-contract tests pin the fix
- Codex (regression): **SAFE TO SHIP** — both renderers (panels.js + ui.js) fixed consistently, WebUI-suppression correct, read-only suffix intact, `textContent` XSS-safe
Co-authored-by: rodboev <rodboev@users.noreply.github.com>
## Release v0.51.244 — Release HL (stage-q16)
UX-approved direction via Telegram (workspace drag-drop polish you requested). All 4 drag-drop flows verified live in-browser.
### Added
| PR | Author | Feature |
|----|--------|---------|
| #3402 / #3424 | @pamnard | **Drop OS files/folders onto a specific workspace folder row or breadcrumb** to upload into that directory (not just the current dir). OS folder drops are traversed (`webkitGetAsEntry`/`readEntries`) preserving nested structure. Uploads via the existing `/api/workspace/upload` (no new backend). |
### Fixed
- **Composer drop-zone jank**: dragging a workspace file (or OS file) over the composer footer rendered a translucent overlay that let the textarea/chips/icons bleed through and collide with the hint text. Now a clean, fully-opaque box with a single centered **context-aware** label — *"Drop to insert workspace reference"* (workspace file → `@path` insert) vs *"Drop files to attach"* (OS file → message attach).
- **Drag-drop handler coexistence (CORE, caught in review)**: #3424's OS-upload binding assigned `el.ondrop` on folder rows, which **overwrote** the drag-to-move handler from #3422 (also `el.ondrop`) — silently breaking move-to-folder (the ws-path drop fell through to the composer as an `@path` insert). Fixed by binding the OS-upload handlers via `addEventListener` so they compose; each handler gates on its own drag type.
### Drag-drop matrix — all verified LIVE in-browser (real drag→drop, asserted on disk)
| Flow | Result |
|------|--------|
| OS image → composer footer | ✓ attaches |
| workspace file → composer footer | ✓ inserts `@path` |
| workspace file → workspace folder | ✓ moves on disk (report.md → docs/) |
| OS file → workspace folder | ✓ uploads into target folder |
### Scope note
#3424's PR branch carried the OLD pre-hardening `_handle_file_move`. Applied **frontend-only** — master's hardened move backend (v0.51.243, TOCTOU/symlink fixes) is untouched (Codex confirmed no `api/routes.py` diff).
### Gate
- Full pytest suite: **7542 passed, 9 skipped, 3 xpassed, 0 failed**
- ESLint: CLEAN · ruff: CLEAN · browser-smoke: CLEAN
- Codex (regression): CORE handler-clobber → fixed → **SAFE TO SHIP**
Co-authored-by: pamnard <pamnard@users.noreply.github.com>
## Release v0.51.243 — Release HK (stage-q15)
UX-approved direction via Telegram (workspace feature you named). Backend security hardened through 3 Codex rounds.
### Added
| PR | Author | Feature |
|----|--------|---------|
| #3402 / #3422 | @pamnard | **Drag a file or folder in the workspace tree onto another folder row (or breadcrumb segment) to move it** within the workspace. New `POST /api/file/move`. Drop handlers use `stopPropagation` so the composer `@path` drag (#1097) and OS-file upload-drop (#3411) are unchanged. |
### Verified
- Live end-to-end: `notes.txt` → `docs/` confirmed on disk; legit moves return the correct `new_path`.
- 13 tests (incl. folder-into-self/descendant guard, existing-target collision, and 3 security regressions).
### Security hardening absorbed (3 Codex rounds — all fixed + regression-tested)
1. **TOCTOU symlink race on destination**: a path-based `source.rename(dest)` could be raced by swapping `dest_dir` to an external symlink between validation and rename. Now opens both parent dirs via the workspace-anchored `open_anchored_fd` (openat + `O_NOFOLLOW`, same helper as the upload hardening) and uses `os.rename(leaf, leaf, src_dir_fd=…, dst_dir_fd=…)` with an fd-based collision check; path-based fallback only where `dir_fd` is unsupported.
2. **Symlinked workspace root** returned a confusing 400 after a successful move — returned `new_path` now computed against `ws_root.resolve()`.
3. **Symlinked source entry**: `safe_resolve` follows the final symlink, so moving `link.txt` would move its *target* and dangle the link — now rejected via no-follow `lstat` on the lexical path.
### Gate
- Full pytest suite: **7530 passed, 9 skipped, 3 xpassed, 0 failed**
- ESLint: CLEAN · ruff: CLEAN · browser-smoke: CLEAN
- Codex (regression): SHIP-ONLY-WITH-FIXES ×3 → all applied + tested → **SAFE TO SHIP**
Co-authored-by: pamnard <pamnard@users.noreply.github.com>
## Release v0.51.240 — Release HH (stage-q12)
Mobile UX bug-fix — UX-approved via Telegram.
### Fixed
| PR | Author | Fix |
|----|--------|-----|
| #3470 | @cnogrin | On mobile/touch, you can now **swipe up to stop streaming auto-scroll**. Previously the view snapped to the bottom on every token with no way to read earlier content while a response streamed — `_recordNonMessageScrollIntent()` only detected upward intent on the wheel path (`e.deltaY`), and touch events have no `deltaY`. Now tracks `_touchStartY` on `touchstart` and treats a `touchmove` that drags the finger down >8px (= scroll up into history, `scrollTop` decreases) as upward intent, setting the same `_messageUserUnpinned` flag the wheel path + scroll listener use. |
### Why this is safe for existing installs
- **Only ADDS a touch-unpin path** — wheel + desktop behavior completely untouched, no existing branch modified.
- New `touchstart`/`touchend`/`touchcancel` listeners are **passive + capture-only** (they only write `_touchStartY`), so they can't interfere with existing touch handling.
- Net effect for users: mobile users *gain* the ability to scroll up during streaming (which was simply broken before). No one's working flow changes.
### Absorbed on review (Codex CORE MUST-FIX)
The contributor's gesture sign was inverted (`dy<-8` = finger up = *follow* the stream), which would have unpinned in the wrong direction. Corrected to `dy>8` to match the existing scroll listener's `movedUp` semantics; fixed the comment; strengthened the regression test to pin the gesture direction.
### Gate
- Full pytest suite: **7503 passed, 9 skipped, 3 xpassed, 0 failed**
- ESLint runtime gate: CLEAN · browser-smoke: CLEAN
- Codex (regression): SHIP-ONLY-WITH-FIXES (inverted sign) → fixed → re-reviewed **SAFE TO SHIP**
- Regression test `tests/test_issue3470_touch_unpin_streaming_scroll.py` (verified to fail against master)
UX-approved via Telegram (mobile touch-gesture behavior, no visual/layout delta to screenshot).
Co-authored-by: cnogrin <cnogrin@users.noreply.github.com>
* Release v0.51.229 (stage-p13): /model never silently snaps versioned name to -tier variant (#3368, @nesquena-hermes)
Agent-authored, nesquena-APPROVED. Rebased onto current master. Both _findModelInDropdown
(ui.js) and _bestModelMatch (commands.js) reject a prefix-snap when the typed target ends
in a version digit and the longer option's extra text is a variant/tier suffix (.pro) rather
than a version continuation (.digit). Adds _nearestModelSuggestion + 'did you mean?' toast.
34 tests pass (14 new #3368 + 20 regression: #1188 fuzzy + #3360 collision).
* fix(commands): /model did-you-mean toast renders suggestion + single quotes (#3368 review)
Live-render review of the approved #3437 caught two toast-assembly bugs in cmdModel:
(1) t('model_did_you_mean') was called WITHOUT the suggestion arg — model_did_you_mean
is a (m)=>... template that t() invokes, so it rendered 'did you mean "undefined"?';
fixed to t('model_did_you_mean', suggestion). (2) no_model_match already ends with an
opening quote, so '"${args}"' doubled it ('No model matching ""deepseek-v4""'); fixed
to '${args}"'. +4 source-assertion regression tests. Verified live: toast now reads
'No model matching "deepseek-v4" — did you mean "deepseek/deepseek-v4-pro"?'.
* fix(commands): slash-qualified versioned no-snap falls through to suggestion (#3368 Codex CORE)
Codex review found a 2nd no-snap layer the version guard missed: for a slash-qualified
versioned query (e.g. 'xiaomi/mimo-v2.5') whose only near catalog entry is a rejected
tier variant ('xiaomi/mimo-v2.5-pro'), cmdModel's cross-provider /api/session/update
fallback would silently persist the invalid model + 'Switched to...'. Now gated on
!versionedNoSnap (_looksLikeVersionedModel(bare) && a near suggestion exists) so it falls
through to the 'did you mean?' toast; genuinely off-catalog providers (no near variant)
still direct-update. Verified live: '/model deepseek/deepseek-v4' no longer switches, shows
suggestion toast. +1 regression test.
---------
Co-authored-by: nesquena-hermes <[email protected]>
* feat(composer): replace mobile config-button sliders icon with context-usage ring (#3062, @NottheGuy007)
Squashed net diff of #3062 (the PR branch's tip commits were deletes of files not
present in our tree). Replaces the composerMobileCtxBadge text badge with an SVG
progress ring (ctx-arc + ctx-num) showing real-time context-window usage: ring
fill via stroke-dashoffset, centered percentage, color-coded green<=50%
orange<=85% red>85%. Ring resets to 0%/green on new session.
* feat(activity): add 'expand activity feed by default' appearance setting (#3080, @AJV20)
Squashed net diff of #3080. Adds a Settings -> Appearance checkbox
(activity_feed_expanded_default, default off) to expand new Activity disclosures
by default; preserves manual per-turn collapse/expand (explicit toggle still
wins); live 'Waiting on model' rows explain what the agent is doing before/after
tool calls. i18n keys for all locales.
* docs(changelog): v0.51.226 — context-usage ring (#3062) + activity-feed default-expand setting (#3080)
* test(mobile): update touch-target test for #3062 ring (badge removed)
#3062 replaced the composerMobileCtxBadge text badge with the SVG context-usage
ring (composerMobileCtxRing) but left 3 stale references in tests/test_mobile_layout.py.
The full suite caught test_mobile_composer_primary_controls_keep_touch_friendly_sizing
asserting the removed .composer-mobile-ctx-badge CSS rule + #composerMobileCtxBadge
element. Updated the assertion to the new ring: confirm composerMobileCtxRing exists,
the old badge is fully gone (not dangling), and the ring SVG is aria-hidden so it
stays a decorative overlay that doesn't steal the config button's 44px touch target
(which is still asserted via .composer-mobile-config-btn above). 56/56 mobile-layout
tests pass. Codex: no production JS dereferences the removed badge.
---------
Co-authored-by: nesquena-hermes <[email protected]>
v0.51.218 fixed getModelLabel() (the visible chip) for URI-scheme model IDs but
left the same first-segment-slash-strip bug in the matching/dedup paths:
api/config.py _norm_model_id + _get_label_for_model, and static/ui.js
_normalizeConfiguredModelKey. For gpt://folder/model/latest those treat the
path slashes as provider delimiters, mis-normalizing the model-identity key
(the #3360 collision class, for URI ids). Adds a URI-scheme guard to all three
so the scheme is not stripped. Kept v0.51.218's getModelLabel (more thorough
than #3436's variant — it extracts the model name vs returning the whole id);
took #3436's backend + _normalizeConfiguredModelKey halves with backend/frontend
parity tests.
Co-authored-by: b3nw <b3nw@duck.com>
Codex re-check: degenerate URIs still leaked — gpt://folder123 returned the
authority, gpt://folder123/${MODEL} returned the placeholder. Removed the _all[0]
authority fallback and guarded the literal-last-path fallback against placeholders;
a URI with no usable model segment now falls back to the raw id. Added regression
cases for gpt://folder123 and gpt://folder123/${MODEL}.
Codex gate found two edges in the first cut: (1) the candidate segment list
included the URI authority, so gpt://folder123/v4 and .../latest returned the
folder; (2) _isVersionTail matched any digit-leading segment, dropping a real
model named 2026-model. Fix: build path segments from AFTER the authority only;
tighten the version-tail regex to pure version tokens (latest/stable/v4/1.2),
not mixed names; fall back to last-usable (non-placeholder) path segment so the
authority is never returned. Added edge-case regression tests.
#3366 changed getModelLabel() to strip only the first /-segment (fixing #3360
multi-slash proxy IDs). That regressed URI-scheme IDs like Yandex
gpt://${FOLDER}/deepseek-v4-flash/latest — indexOf('/') lands inside the ://
and leaves /${FOLDER}/... path junk in the composer model chip. Detect a
scheme:// id, drop scheme+authority, and take the last meaningful path segment
(skipping ${...} env-var placeholders and bare version tails like latest).
Non-URI multi-slash IDs keep the #3360 first-segment-strip behavior unchanged.
Node-driven regression test covers the URI case + the #3360 non-regression.
Capture a same-session force-reload hint (loaded renderable/message counts,
known count, truncation flag) BEFORE clearing the in-memory transcript, so the
authoritative reload requests a width that preserves what was loaded instead of
collapsing a long session to the default 30-message tail window mid-read.
Same-session force reloads render with preserveScroll. Resolves conflict with
the shipped #3306 carry-forward snapshot by keeping both (complementary).
Closes#3239.
Co-authored-by: viraatdas <viraatdas@users.noreply.github.com>
When a custom/proxy provider serves models whose IDs share the same base
name across vendor prefixes (e.g. vendor_a/deepseek/deepseek-v4-pro vs
vendor_b/deepseek/deepseek-v4-pro), several normalization functions use
split('/').pop() (or split('/')[-1]) which discards all segments except
the last. This causes three user-facing symptoms: (1) clicking one
model selects a different colliding model, (2) configured-model badges
attach to the wrong dropdown entry, and (3) the model-chip label in the
composer bar is truncated to just the base model name.
Root cause: all three callers take only the last slash-segment instead
of stripping only the first (provider) segment and preserving the
remaining vendor hierarchy.
Fix 1 — _findModelInDropdown (static/ui.js): Move the exact string match
before the provider-aware normalized match. Previously, when all models
share the same provider ID (common with LLM proxy setups), the normalized
match returned whichever colliding option appeared first in DOM order,
even though an exact match existed.
Fix 2 — _normalizeConfiguredModelKey (static/ui.js) and _norm_model_id
(api/config.py): Replace split('/').pop() / split('/')[-1] with a first-
segment-only strip (regex on frontend, split('/',1) on backend), matching
the strategy already used by _findModelInDropdown's norm lambda. This
prevents multi-slash IDs from colliding in badge assignment and the
configured-entry dedup set. Additionally, strip colon-qualified provider
prefixes (e.g. custom:name/) before the slash strip so badge-key variants
like 'custom:llm-proxy/opencode_go/model' merge correctly with the bare
'opencode_go/model' in the configured section dedup.
Fix 3 — getModelLabel (static/ui.js) and _get_label_for_model
(api/config.py): Same split('/').pop() to first-segment-strip change so
the composer-bar model chip and backend label preserve vendor context
(e.g. shows 'opencode_go/deepseek-v4-pro' instead of 'deepseek-v4-pro').
Verification: 9 new regression tests (test_issue3360) covering exact-
match priority, multi-slash normalization, and backend/frontend parity.
Updated 1 existing test (test_norm_model_id_trailing_empty_guard) that
asserted the old split('/').pop() pattern. All 25 related tests pass.
AI Usage: Gemini (gemini-2.5-pro), via Antigravity IDE, pair-programmed.
(cherry picked from commit a454fecd2b3a83f7da34473c883be069171aeac9)
Both advisors (Opus MUST-FIX + Codex SHIP-ONLY-WITH-FIXES) caught that the Edge
TTS playback was broken despite the settings selector working:
1. ui.js _playEdgeTts used new Audio('/api/tts?text=...') — a GET — but /api/tts
is POST-only (405) and registered only in handle_post. The per-message speaker
button + auto-read silently failed in edge mode, and the GET leaked message
text into the query string/access log. Rewritten to POST JSON + blob object
URL (mirrors the working boot.js path), and now surfaces server errors
(503 not-installed, 429 rate-limit) via toast instead of silent dead air.
2. boot.js hands-free Edge audio was a local var never assigned to the shared
_playingEdgeAudio handle, so stopTTS() (from _deactivate) couldn't stop it.
Now registered + cleared on end/error.
3. Test isolation: the _tts_limiter function-attribute singleton persisted across
the whole suite, flaking 2 of my endpoint tests in the full run. Converted the
reset to an autouse fixture (before+after each) + unique per-test client IPs.
Opus iter2 SHOULD-FIX (both non-blocking, applied for cleanliness):
- esc() the media_download title/aria-label so a future translator's quote
can't break out of the attribute (defense-in-depth).
- Remove the now-dead media_open key from all 11 locales + the ui.js fallback
dict (the Open button was dropped in the clean-image redesign).
#3220: redesign generated-image rendering from a permanent bordered card
(filename + Open/Download buttons) to a clean inline image with click-to-zoom
lightbox + a hover/focus-revealed Download overlay, matching ChatGPT/Claude/
Gemini. Drops redundant Open (lightbox already covers it). Tests updated.
#3337: fix two-tone code background — Prism's prism-tomorrow theme styled the
parent <pre> gray while the <code> was navy var(--code-bg), so dark theme showed
a gray frame around navy code. Override BOTH .preview-code[class*=language-] and
its <code> to var(--code-bg) (mirrors the chat code-block fix at .msg-body pre).
#3223: fix action-menu clipping — the new Regenerate-title row made the 9-item
menu tall enough to overflow the viewport bottom when opened on a top-anchored
row at short viewports (e.g. 1280x720). _positionSessionActionMenu now clamps
the menu within both viewport edges and caps max-height with scroll when the
menu is taller than the viewport.
Supersedes the v0.51.199 proximity-re-pin (#3330) and the #3250 upward-intent
timeout with a sticky-unpin model (ChatGPT/Claude/Codex behavior): scroll up =
stay put until you return to the bottom or click the scroll-to-bottom control.
Reconciled against the shipped #3330 code: removed the now-dead
_recentMessageUpwardIntent reference from the #3319 rAF retry, kept the
load-time -Infinity intent-init fix, kept the pinned-only >500 catch-up.
Co-authored-by: pamnard <pamnard@users.noreply.github.com>
Codex+Opus regression-gate finding: _isRecoveryControlMessage /
_streamRecoveryControlMessage fell back to matching
provider_details_label==='interruption details'. But a GENUINE 'Response
interrupted' card (Stop button, real provider crash) carries that exact label,
so the filter would drop a real user-facing interruption from the transcript on
the next render/restore — the inverse of the #3300 data-loss class. Require the
explicit server-set recovery_control marker; keep only the two fully-anchored
synthetic-text matches for pre-marker backward-compat. Adds a node-driven
regression test (revert-verified) asserting a label-only interruption card and a
real user turn stay visible while marker + strict text are filtered.
Codex follow-up finding: _lastMessageUpwardIntentMs/_lastNonMessageScrollIntentMs
initialized to 0, so _recentMessageUpwardIntent() returned true for the first 2s
after load (performance.now() < MESSAGE_UPWARD_INTENT_MS=2000) even with no user
scroll — which would disable the new #3319 retry guard during initial load. Sentinel
-Infinity makes 'no event recorded yet' read as no-intent. Also tightens the
pre-existing scrollIfPinned/settle callers that read the same helpers.
Codex regression-gate finding: the new requestAnimationFrame retry in
_setMessageScrollToBottom re-asserted scrollTop=scrollHeight + _scrollPinned=true
on the next layout frame unconditionally, so a user who scrolled up in that ~16ms
window during streaming would be snapped back and re-pinned, bypassing the
scrollIfPinned early-return guards. Re-check _messageUserUnpinned / upward-intent /
non-message-scroll-intent / !_scrollPinned inside the retry; on hit, only release
the programmatic-scroll latch and bail.
Increase MESSAGE_UPWARD_INTENT_MS from 450ms to 2000ms to fix a race
condition where the user scrolls up during streaming, pauses to read
for >450ms, and then gets snapped back to the bottom.
The root cause: after the 450ms upward-intent window expires, DOM layout
changes from the streaming markdown parser (smd), tool card insertions,
or code re-highlighting can trigger scroll events that the handler no
longer recognizes as user-initiated. When the resulting position lands
inside the 250px near-bottom zone for two consecutive samples, the
hysteresis counter re-pins (_scrollPinned=true) and the next streaming
token's scrollIfPinned() call forces scrollTop to the bottom.
With a 2-second window, the user's upward intent persists through typical
streaming DOM churn. Downward scrolling and the scroll-to-bottom button
are unaffected — movedUp requires top < _lastScrollTop-2 which is false
for downward movement regardless of the intent timeout.
Refs: #1360 (macOS momentum protection), #1731 (direction-aware unpin)
1. routes.py: hoist a single case-folded path helper (_norm/_within_ci/_equal_ci)
used for ALL deny + carve-out comparisons (consistent macOS/Windows safety).
2. routes.py: split the deny into (a) dir-based denies that ALWAYS fire (even
inside the active workspace — so a workspace overlapping a state dir cannot
expose sessions/memories), and (b) filename denies relaxed only by the
carve-out. Fix the over-block: a workspace that is a proper DESCENDANT of a
Hermes root (e.g. STATE_DIR/workspace) is a legit project workspace and keeps
the carve-out; only a root-itself / ancestor / $HOME / profiles / state-subdir
workspace disables it.
3. ui.js: move the bare file:// media-stash pass after the raw-<pre> stash too,
so file:// inside a raw <pre> block stays literal (not just fenced/inline code).
1. api/routes.py: case-fold /api/media deny filename + dir containment checks
(os.path.normcase + casefold) so STATE.DB / Sessions/ cannot bypass the
state/secret deny on case-insensitive filesystems (macOS/Windows).
2. static/ui.js: move the bare file:// media-stash pass to run AFTER fenced-block
and inline-code stashing, so a file:// inside a code block / backtick span
stays literal text instead of becoming an auto-loaded <img>. The MEDIA: stash
keeps its first-position precedent.
Adds behavioral renderer tests (real renderMd via node) for fenced + inline code
file:// staying literal, bare file:// becoming media, and anchors keeping the
link path. Closes the last Codex review items for #3234.
