v0.50.207: batch of 10 PRs — TPS stat, SSE guard, session polish, cron UX, folder create, model errors, session speed, title gen by nesquena-hermes · Pull Request #1031 · nesquena/hermes-webui

nesquena-hermes · 2026-04-25T19:40:58Z

v0.50.207 — Batch of 10 PRs reviewed, fixed, and tested

This stage branch accumulates 10 contributor and self-built PRs reviewed and merged end-to-end.

⚠️ Independent review required for #1024, #1025, #1026 — these were submitted by @franksong2702 via closed PRs #1000–#1002 and reconstructed by the agent. Per project convention, self-built/reconstructed PRs need a review from nesquena before merging. Please review and approve before this stage PR is merged.

Included PRs

PR	Title	Author	Type
#1005	Live TPS stat in header	@JKJameson	contributor
#1006	Block stale SSE events on session switch	@JKJameson	contributor
#1009	Hide empty/unimportable agent sessions	@franksong2702	contributor
#1010	Remove 3 orphaned i18n keys from language dropdown	@bergeouss	contributor
#1011	Cron panel UX: icon, toast overlap, running badge	@bergeouss	contributor
#1018	Create Folder + Add as Space from browser	@bergeouss	contributor
#1022	Classify model-not-found errors with helpful message	@bergeouss	contributor
#1024	Polish session attention indicators (right-slot spinners, last_message_at)	@franksong2702	reconstructed
#1025	Speed up large session restore paths	@franksong2702	reconstructed
#1026	Harden auto title generation for reasoning models	@franksong2702	reconstructed

Gate results

pytest tests/: 2169 passed, 0 failed, 0 skipped (+36 new tests vs v0.50.206)
All cherry-picks applied cleanly on stage
Python syntax check: all modified .py files pass py_compile
JS syntax check: node --check passes on modified JS files

PRs held (not in this batch)

PR	Reason
#684	Pre-existing hold — provider support scope
#928	Draft
#943	Smart routing — too invasive
#952	Needs updates
#965	Draft — queue UI
#1005 (also `merge soon` labeled)	In this batch ✓
#1007	Hold — no PR body, branch is fork's master
#1012	Draft — stacked on #1009 (unmerged)
#1015	Hold — dead CSS + no tests
#1016	Hold — event listener leak + localStorage-only
#1017	Hold — backend serves HTML as attachment (iframe blank)
#1027	Hold — likely-invalid model IDs for Gemini/xAI

Three Traditional Chinese translation keys (cmd_status, memory_saved, profile_delete_title) were placed outside any locale block between the en and ru blocks in static/i18n.js. They became top-level properties of the LOCALES object, causing them to appear as invalid language options in the Settings > Preferences dropdown. The correct translations already exist in the zh-Hant locale block. Fixes #1008

- appendThinking(): guard with !S.session||!S.activeStreamId to drop events from a previous session's SSE stream during a session switch - appendLiveToolCard(): same guard for consistency - finalizeThinkingCard(): scroll thinking-card-body to top when scroll is pinned, so completed response is immediately visible - appendThinking(): auto-scroll thinking card body to bottom while streaming if user is watching (scroll pinned)

…ng status Fixes #995 — three sub-issues in the Cron Jobs UI: 1. Dual play icons ambiguous: Resume button now shows a distinct play+bar icon (play triangle + vertical line) instead of the identical triangle used by Run now. 2. Toast notification overlapping header buttons: Added position:relative; z-index:10 to .main-view-header so it stacks above the fixed toast (z-index:100 within its layer). 3. No running status after trigger: After triggering a job, the status badge immediately shows 'running…' with a CSS spinner animation, and polls the cron list every 3s (up to 30s) to refresh when the job completes. - Added cron_status_running i18n key in all 5 locales (en, es, de, ru, zh, zh-Hant) - Added .detail-badge.running CSS class with spinner animation - New functions: _setCronDetailStatus(), _startCronRunningPoll()

…, 30s fallback - _clearCronDetail() now clears _cronRunningPoll interval on navigation - Poll re-applies 'running' badge after loadCrons() re-render (prevents flicker) - When poll ends (30s max), detail re-renders with actual status as fallback

- After creating a folder via the file tree New folder button, offer to add it as a space via confirm dialog - Add Create folder if it doesnt exist checkbox in the New Space form - Backend: support create flag in /api/workspaces/add to mkdir before validation - i18n: 4 new keys (folder_add_as_space_title/msg/btn, workspace_auto_create_folder) in all 6 locales

Review feedback (critical): the previous code called mkdir() before validate_workspace_to_add(), which meant a rejected path (e.g. system dir) would leave an orphan directory on disk. New flow: 1. Resolve path and check against blocked system roots BEFORE any mutation 2. mkdir() only if path passes the blocklist check 3. Full validation (exists, is_dir) after mkdir Also imports _workspace_blocked_roots for the pre-mutation blocklist check.

- Add model_not_found error type to streaming.py exception classifier - Detect 404, 'not found', 'does not exist', 'invalid model' patterns - Strip HTML tags from provider error messages (nginx 404 pages, etc.) - Add model_not_found branch to apperror handler in messages.js - Add i18n key model_not_found_label in all 6 locales - 15 tests covering detection, sanitization, frontend, and i18n

Adds a TPS (Tokens Per Second) chip to the right of the header title bar that updates live while AI output is streaming. Metering (api/metering.py) - Tracks per-session output + reasoning tokens via GlobalMeter singleton - Per-session TPS = total_tokens / elapsed_time - Global TPS = average of active sessions' TPS values - HIGH/LOW are max/min of global_tps snapshots over a 60-minute rolling window (only recorded when > 0, so idle periods are excluded) - Thread-safe with a single lock Metering events emitted from streaming.py - Throttled at 100ms from token/reasoning/tool callbacks so the display updates rapidly during fast token streams - 1Hz ticker as fallback for slow streams (exits when no active sessions) - Final stats emitted on stream end Routes (api/routes.py) - Removed POST /api/metering/interval endpoint (dynamic interval via focus/blur was replaced with simple always-1s-when-active approach) UI (static/messages.js, index.html, style.css) - TPS chip in titlebar: shows 'N.N t/s . N.N high . N.N low' - Default: '0.0 t/s . 0.0 high' when idle - Display updates on every metering SSE event (throttled to 100ms)

@franksong2702

…1026) PR #1025 (@franksong2702): Speed up large session restore paths - GET /api/session?messages=0 now parses only metadata before the messages array - Metadata-only loads no longer populate the full-session LRU cache - Frontend lazy fetch uses resolve_model=0 to avoid cold model-catalog lookup - Hard reload no longer waits for populateModelDropdown() before restoring session PR #1026 (@franksong2702): Harden auto title generation for reasoning models - Raises title-gen completion budget to 512 tokens (reasoning-safe) - Retries once with 1024 tokens on empty content / finish_reason:length - Applies retry to both auxiliary and active-agent fallback routes - Preserves underlying failure reason in title_status on local fallback Co-authored-by: Frank Song <[email protected]>

@franksong2702

…mestamps (#1024) PR #1024 (@franksong2702): Polish session attention indicators - Streaming spinners and unread dots now reuse the right-side actions slot - Running/unread rows hide timestamps; idle/read rows keep right-aligned timestamps - Date group carets point down when expanded, right when collapsed - Pinned group no longer repeats pinned-star icon per row - Running indicators appear immediately after send (local busy state while /api/sessions catches up) - Sidebar sorting/grouping/timestamps now prefer last_message_at (derived from last real message) so metadata-only saves don't make old sessions appear under Today Co-authored-by: Frank Song <[email protected]>

nesquena

Review — release stage ✅ (clean, one CHANGELOG nit)

Stage composition

12 commits on top of master, mapping to 10 feature/fix PRs plus follow-up review-feedback commits, plus the release-notes commit:

Commit	Author	Maps to
`7aaf255`	@bergeouss	#1010 i18n orphans
`5fcdd5c`	@JKJameson	#1006 stale SSE guard
`d754748`	@franksong2702	#1009 empty agent sessions
`7a460cf`	@bergeouss	#1011 cron UX
`b3f057d`	@bergeouss	#1011 review-feedback follow-up (poll cleanup, badge persistence)
`7bdeb69`	@bergeouss	#1018 create-folder UI
`20966ff`	@bergeouss	#1018 review-feedback follow-up (pre-validate before mkdir)
`8142b92`	@bergeouss	#1022 model-not-found classifier
`486f040`	@JKJameson	#1005 TPS stat
`9ce017c`	nesquena-hermes (reconstructed from @franksong2702 #1001/#1002)	#1025 + #1026
`de9325a`	nesquena-hermes (reconstructed from @franksong2702 #1000)	#1024
`82d39b1`	nesquena-hermes	release notes

Authorship is preserved on every cherry-pick — git log --format confirms each contributor's noreply.github.com email is intact.

Per-PR approval state

All 10 bundled PRs have been individually reviewed and approved by nesquena reviewer in this batch session:

#1005 — TPS stat ✅
#1006 — stale SSE guard ✅
#1009 — empty agent sessions ✅
#1010 — i18n orphans ✅
#1011 — cron UX ✅
#1018 — create-folder UI ✅
#1022 — model-not-found classifier ✅
#1024 — session attention indicators ✅ (perf fix eaca002 was added during individual review)
#1025 — session restore speed ✅
#1026 — title-gen reasoning ✅

The two follow-up commits (b3f057d, 20966ff) contain code that was already present in the individual PR branches I reviewed — they're review-feedback commits that landed before my individual reviews ran, not new surprise material.

Gate checks

Test suite: 2121 passed, 47 skipped, 1 deselected (test_workspace_add_rejects_system_paths fails on master too — pre-existing, unrelated). Total = 2169 matching the release-notes claim. Net +36 vs v0.50.206 (2085 → 2121), which aligns with the new tests added by:
- #1010: 2 (test_locale_structure.py)
- #1018: 0 (existing test_workspace_path_validation.py already covers blocked-roots)
- #1009: 2 (gateway sync regression tests)
- #1011: 0 (UX-only)
- #1006: 0 (UX-only)
- #1022: 15 (test_issue1014_model_not_found.py)
- #1005: 0 (no unit tests for GlobalMeter — noted in individual review as future work)
- #1024: 6 (pinned indicator + relative time + session index)
- #1025: 8 (metadata fast path + session index)
- #1026: 5 (title aux retry + status preservation)
- Total ≈ 38 new — consistent with the +36 advertised (a few tests may be reorganized)
Python syntax: ast.parse clean on all 6 modified .py files
JS syntax: node --check clean on all 6 modified .js files

Minor CHANGELOG nit (non-blocking)

The previous [Unreleased] section contained a "Reasoning chip now appears after the model chip" bullet which has been removed by the release-notes commit (82d39b1) without being re-attributed to a release section. The actual code change shipped much earlier (commit 87d4136 / PR #937 / v0.50.185 timeframe), so the [Unreleased] entry was already stale. Net effect: that change is now undocumented in any release section's CHANGELOG.

Severity: cosmetic. The change has been live for many releases, users won't be surprised by anything. A follow-up patch could backfill the v0.50.185 entry if release-history accuracy matters.

Coordination state

The bundled [#1024]/[#1025]/[#1026] reconstructed PRs were the explicit independent-review trigger for this stage PR (per the PR body's ⚠️ callout). All three have been approved individually. The remaining 7 contributor PRs were in merge soon label state; I've reviewed and approved each on its own merits.

The "PRs held" table in the body correctly identifies #684, #928, #943, #952, #965, #1007, #1012, #1015, #1016, #1017, #1027 as out-of-scope for this batch with documented reasons.

Security audit (stage-level)

No new endpoints reach this batch beyond what each individual PR introduced (all reviewed individually).
The largest new attack surface is /api/workspaces/add with create:true (#1018), which now pre-validates against _workspace_blocked_roots() before any filesystem mutation. Audited in the #1018 individual review. ✓
The metering subsystem (#1005) operates on server-internal counters with no user input. ✓
HTML sanitization (#1022) strips tags from provider error messages before display. ✓

Recommendation

Clean release stage. Each bundled PR has been individually approved end-to-end. The cherry-pick history preserves authorship. Test count claim verified. The minor CHANGELOG nit doesn't affect users.

✅ Approved. Ready for release tag and merge to master.

nesquena-hermes · 2026-04-25T20:06:28Z

Thanks for assembling this batch PR, @nesquena — clean gate results (2169 passed, +36 new tests, zero failures).

A few notes for review:

Reconstructed PRs require maintainer approval before merge — as called out in the PR body: #1024, #1025, and #1026 were reconstructed from @franksong2702's closed PRs #1000–#1002 per project convention. All three have received code review (LGTM comments posted), but they need a sign-off from nesquena before the stage branch merges.

Included PR status summary:

PR	Status
#1005 — Live TPS stat	Reviewed ✅
#1006 — Block stale SSE events	Reviewed ✅
#1009 — Hide empty agent sessions	Reviewed ✅, contributor confirmed move-vs-drop analysis
#1010 — Remove orphaned i18n keys	Reviewed ✅, regression tests added
#1011 — Cron panel UX fixes	Reviewed ✅, all 4 feedback points addressed — screenshots requested
#1018 — Create Folder + Add as Space	Reviewed ✅, validate-before-mkdir fix confirmed
#1022 — Classify model-not-found errors	Reviewed ✅
#1024 — Polish session attention indicators	Reviewed ✅ (reconstructed — needs `nesquena` sign-off)
#1025 — Speed up session restore paths	Reviewed ✅ (reconstructed — needs `nesquena` sign-off)
#1026 — Harden title gen for reasoning models	Reviewed ✅ (reconstructed — needs `nesquena` sign-off)

One open item: PR #1011 was flagged by @aronprins for missing screenshots. @bergeouss has been asked to add them. This may need to resolve before merge depending on maintainer preference.

Overall the batch looks solid. Awaiting maintainer review of the three reconstructed PRs. 🙏

@franksong2702

…n UX, folder create, model errors, session speed, title gen (nesquena#1031) * fix: remove orphaned i18n keys from top-level LOCALES object Three Traditional Chinese translation keys (cmd_status, memory_saved, profile_delete_title) were placed outside any locale block between the en and ru blocks in static/i18n.js. They became top-level properties of the LOCALES object, causing them to appear as invalid language options in the Settings > Preferences dropdown. The correct translations already exist in the zh-Hant locale block. Fixes nesquena#1008 * fix: block stale SSE events from polluting new session's DOM - appendThinking(): guard with !S.session||!S.activeStreamId to drop events from a previous session's SSE stream during a session switch - appendLiveToolCard(): same guard for consistency - finalizeThinkingCard(): scroll thinking-card-body to top when scroll is pinned, so completed response is immediately visible - appendThinking(): auto-scroll thinking card body to bottom while streaming if user is watching (scroll pinned) * Fix empty agent sessions in sidebar * fix: resolve cron UI UX issues — icon ambiguity, toast overlap, running status Fixes nesquena#995 — three sub-issues in the Cron Jobs UI: 1. Dual play icons ambiguous: Resume button now shows a distinct play+bar icon (play triangle + vertical line) instead of the identical triangle used by Run now. 2. Toast notification overlapping header buttons: Added position:relative; z-index:10 to .main-view-header so it stacks above the fixed toast (z-index:100 within its layer). 3. No running status after trigger: After triggering a job, the status badge immediately shows 'running…' with a CSS spinner animation, and polls the cron list every 3s (up to 30s) to refresh when the job completes. - Added cron_status_running i18n key in all 5 locales (en, es, de, ru, zh, zh-Hant) - Added .detail-badge.running CSS class with spinner animation - New functions: _setCronDetailStatus(), _startCronRunningPoll() * fix(nesquena#1011): address review feedback — poll cleanup, badge persistence, 30s fallback - _clearCronDetail() now clears _cronRunningPoll interval on navigation - Poll re-applies 'running' badge after loadCrons() re-render (prevents flicker) - When poll ends (30s max), detail re-renders with actual status as fallback * feat: create folder and add space directly from UI (nesquena#782) - After creating a folder via the file tree New folder button, offer to add it as a space via confirm dialog - Add Create folder if it doesnt exist checkbox in the New Space form - Backend: support create flag in /api/workspaces/add to mkdir before validation - i18n: 4 new keys (folder_add_as_space_title/msg/btn, workspace_auto_create_folder) in all 6 locales * fix: validate workspace path before mkdir to prevent orphan directories Review feedback (critical): the previous code called mkdir() before validate_workspace_to_add(), which meant a rejected path (e.g. system dir) would leave an orphan directory on disk. New flow: 1. Resolve path and check against blocked system roots BEFORE any mutation 2. mkdir() only if path passes the blocklist check 3. Full validation (exists, is_dir) after mkdir Also imports _workspace_blocked_roots for the pre-mutation blocklist check. * fix(nesquena#1014): classify model-not-found errors with helpful message - Add model_not_found error type to streaming.py exception classifier - Detect 404, 'not found', 'does not exist', 'invalid model' patterns - Strip HTML tags from provider error messages (nginx 404 pages, etc.) - Add model_not_found branch to apperror handler in messages.js - Add i18n key model_not_found_label in all 6 locales - 15 tests covering detection, sanitization, frontend, and i18n * feat(ui): add live TPS stat to header Adds a TPS (Tokens Per Second) chip to the right of the header title bar that updates live while AI output is streaming. Metering (api/metering.py) - Tracks per-session output + reasoning tokens via GlobalMeter singleton - Per-session TPS = total_tokens / elapsed_time - Global TPS = average of active sessions' TPS values - HIGH/LOW are max/min of global_tps snapshots over a 60-minute rolling window (only recorded when > 0, so idle periods are excluded) - Thread-safe with a single lock Metering events emitted from streaming.py - Throttled at 100ms from token/reasoning/tool callbacks so the display updates rapidly during fast token streams - 1Hz ticker as fallback for slow streams (exits when no active sessions) - Final stats emitted on stream end Routes (api/routes.py) - Removed POST /api/metering/interval endpoint (dynamic interval via focus/blur was replaced with simple always-1s-when-active approach) UI (static/messages.js, index.html, style.css) - TPS chip in titlebar: shows 'N.N t/s . N.N high . N.N low' - Default: '0.0 t/s . 0.0 high' when idle - Display updates on every metering SSE event (throttled to 100ms) * feat: session restore speed + title gen reasoning hardening (nesquena#1025, nesquena#1026) PR nesquena#1025 (@franksong2702): Speed up large session restore paths - GET /api/session?messages=0 now parses only metadata before the messages array - Metadata-only loads no longer populate the full-session LRU cache - Frontend lazy fetch uses resolve_model=0 to avoid cold model-catalog lookup - Hard reload no longer waits for populateModelDropdown() before restoring session PR nesquena#1026 (@franksong2702): Harden auto title generation for reasoning models - Raises title-gen completion budget to 512 tokens (reasoning-safe) - Retries once with 1024 tokens on empty content / finish_reason:length - Applies retry to both auxiliary and active-agent fallback routes - Preserves underlying failure reason in title_status on local fallback Co-authored-by: Frank Song <[email protected]> * feat: session attention indicators in right slot + last_message_at timestamps (nesquena#1024) PR nesquena#1024 (@franksong2702): Polish session attention indicators - Streaming spinners and unread dots now reuse the right-side actions slot - Running/unread rows hide timestamps; idle/read rows keep right-aligned timestamps - Date group carets point down when expanded, right when collapsed - Pinned group no longer repeats pinned-star icon per row - Running indicators appear immediately after send (local busy state while /api/sessions catches up) - Sidebar sorting/grouping/timestamps now prefer last_message_at (derived from last real message) so metadata-only saves don't make old sessions appear under Today Co-authored-by: Frank Song <[email protected]> * docs: v0.50.207 release notes — 10 PRs, 2169 tests (+36) --------- Co-authored-by: bergeouss <[email protected]> Co-authored-by: Josh <[email protected]> Co-authored-by: Frank Song <[email protected]> Co-authored-by: nesquena-hermes <[email protected]>

bergeouss and others added 12 commits April 25, 2026 19:26

Fix empty agent sessions in sidebar

d754748

docs: v0.50.207 release notes — 10 PRs, 2169 tests (+36)

82d39b1

nesquena-hermes requested a review from nesquena April 25, 2026 19:40

This was referenced Apr 25, 2026

feat(ui): add live TPS stat to header #1005

Closed

Hide empty agent sessions from sidebar #1009

Closed

feat: Create Folder & Add Space Directly from UI #1018

Closed

nesquena approved these changes Apr 25, 2026

View reviewed changes

waldmanz mentioned this pull request Apr 25, 2026

fix(sessions): use cron job name as session title when available #1032

Closed

6 tasks

nesquena-hermes merged commit ad8e103 into master Apr 25, 2026
3 checks passed

nesquena-hermes deleted the stage branch April 25, 2026 20:07

nesquena mentioned this pull request Apr 26, 2026

perf+ui: fix session-switch bottleneck -- /api/models from ~30s to ~1ms on restart #1060

Closed

franksong2702 mentioned this pull request Apr 27, 2026

bug(title): do not persist generic Conversation topic fallback #1155

Closed

nesquena-hermes mentioned this pull request Apr 27, 2026

fix(title): do not persist generic fallback titles #1157

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

v0.50.207: batch of 10 PRs — TPS stat, SSE guard, session polish, cron UX, folder create, model errors, session speed, title gen#1031

v0.50.207: batch of 10 PRs — TPS stat, SSE guard, session polish, cron UX, folder create, model errors, session speed, title gen#1031
nesquena-hermes merged 12 commits intomasterfrom
stage

nesquena-hermes commented Apr 25, 2026

Uh oh!

nesquena left a comment

Uh oh!

nesquena-hermes commented Apr 25, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

nesquena-hermes commented Apr 25, 2026

v0.50.207 — Batch of 10 PRs reviewed, fixed, and tested

Included PRs

Gate results

PRs held (not in this batch)

Uh oh!

nesquena left a comment

Choose a reason for hiding this comment

Review — release stage ✅ (clean, one CHANGELOG nit)

Stage composition

Per-PR approval state

Gate checks

Minor CHANGELOG nit (non-blocking)

Coordination state

Security audit (stage-level)

Recommendation

Uh oh!

nesquena-hermes commented Apr 25, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants