Commit Graph

1086 Commits

Author SHA1 Message Date
nesquena-hermes d71b8977d6 Stage 401: PR #2742 2026-05-22 15:22:01 +00:00
nesquena-hermes 742c9cbd9b Stage 401: PR #2730 2026-05-22 15:21:32 +00:00
Isla-Liu 5b41f03a92 fix(webui): close sqlite3 connections in handoff-summary path (#2233)
Two functions on the /api/session/handoff-summary hot path were opening
sqlite3.connect(...) inside a bare `with` statement, which commits the
transaction at scope exit but does NOT close the connection. Per-turn
invocations accumulated state.db / state.db-wal file descriptors and
CPython heap pages on long-lived worker threads, surfacing as the
multi-GB VmRSS and 6x duplicated state.db fds observed on the live
instance (D0 pre-restart baseline: VmRSS 1,334,248 kB, 55 fds; cold
baseline after restart: VmRSS 136,668 kB, 10 fds).

Wrap both call sites with contextlib.closing(...) (already imported and
used at seven other sites in the same files) so the connection is
closed deterministically:

  - api/models.py :: count_conversation_rounds
  - api/routes.py :: _persist_handoff_summary_to_state_db

Regression test:
  tests/test_issue2233_sqlite_connection_leak.py loops both functions
  20 times against a tmp state.db and asserts /proc/<pid>/fd count
  does not grow more than 2. Linux-only via sys.platform skip.

D1 live soak against a freshly-built worktree server (port 8799,
isolated HERMES_HOME / HERMES_WEBUI_STATE_DIR) hitting
/api/session/handoff-summary 20 times:

  fd_before      = 5
  fd_after       = 5     (growth 0, threshold < 5)
  vmrss_before   = 52636 kB
  vmrss_after    = 52636 kB  (growth 0 kB, threshold < 30 MB)

The patched fix curve trends below the leak curve.

Rollback: single git revert <this-sha> reverts both file edits.

Refs #2233.
2026-05-22 18:34:06 +08:00
wdzhou a4e6ffccd9 fix(session): deduplicate _write_session_index full rebuild entries by session_id
The full rebuild path scans SESSION_DIR via glob('*.json') and appends every loaded session to a plain list without deduplicating by session_id. When old-format session_*.json files coexist alongside WebUI-format xxx.json files (both sharing session_id), the index gets duplicate entries, causing frontend Vue key crashes.

Fix: use dict[session_id -> compact_entry] to naturally deduplicate.
2026-05-22 18:02:49 +08:00
wdzhou 16f9887846 fix(session): deduplicate _write_session_index full rebuild by session_id
The full rebuild path of _write_session_index scans SESSION_DIR via
glob('*.json') and appends every loaded session to a plain list without
deduplicating by session_id. When old-format session_*.json files coexist
alongside WebUI-format xxx.json files (both sharing the same session_id),
the same session appears multiple times in the index, causing frontend
Vue key collisions and a blank page.

Fix: use dict[session_id -> compact_entry] to naturally deduplicate.
Prefer the entry with the larger message_count when conflicts arise.
2026-05-22 16:13:42 +08:00
Ashish Vaja b63bdae09b fix: redact update-check git diagnostics 2026-05-22 02:28:17 +00:00
Hermes Agent 0718770fe1 Stage 400: PR #2719 — fix: apply release update target (clear sessionStorage check-once stamp + force banner re-evaluation on user click)
Co-authored-by: ai-ag2026 <ai-ag2026@users.noreply.github.com>
2026-05-21 22:59:56 +00:00
Hermes Agent b14aae4ee5 Stage 400: PR #2717 — fix: surface update check fetch errors instead of failing silently
Co-authored-by: ai-ag2026 <ai-ag2026@users.noreply.github.com>
2026-05-21 22:59:54 +00:00
Hermes Agent 654f62e0bd Stage 400: PR #2721 — fix(session): treat active runs as live during repair (skip restart-stale prune for sessions with live streams)
Co-authored-by: ai-ag2026 <ai-ag2026@users.noreply.github.com>
2026-05-21 22:59:43 +00:00
Hermes Agent 3874781a57 Stage 399: PR #2704 — fix(streaming): prevent exponential empty _partial accumulation across reasoning-only cancellations
Co-authored-by: wirtsi <wirtsi@users.noreply.github.com>
2026-05-21 17:56:46 +00:00
Hermes Agent 7152da2756 Stage 399: PR #2705 — fix(api): dedupe replayed context_messages so the agent doesn't see duplicates (follow-up to v0.51.96 #2620)
Co-authored-by: AlexeyDsov <AlexeyDsov@users.noreply.github.com>
2026-05-21 17:56:43 +00:00
Hermes Agent 4db8df5e29 Stage 399: PR #2686 — fix(session): dedupe restamped state.db replay rows in /api/session display merge
Co-authored-by: ai-ag2026 <ai-ag2026@users.noreply.github.com>
2026-05-21 17:56:40 +00:00
Hermes Agent 905b3eba5e Stage 398: PR #2700 — feat: make pinned session limit configurable (builds on shipped #2614 3-cap)
Co-authored-by: ai-ag2026 <ai-ag2026@users.noreply.github.com>
2026-05-21 17:43:56 +00:00
Hermes Agent f563d37244 Stage 398: PR #2696 — feat(runtime): add runner-local adapter selection (RuntimeAdapter slice 4c, feature-flagged)
Co-authored-by: Michaelyklam <Michaelyklam@users.noreply.github.com>
2026-05-21 17:43:54 +00:00
Hermes Agent 6864739e55 Stage 398: PR #2703 — fix: detect agent version from copied source (Docker two-container System panel) (closes #2691)
Closes #2691

Co-authored-by: Michaelyklam <Michaelyklam@users.noreply.github.com>
2026-05-21 17:43:51 +00:00
Hermes Agent 7d3013245a Stage 398: PR #2687 — feat: hide suggestions preference (closes #2679)
Closes #2679

Co-authored-by: Michaelyklam <Michaelyklam@users.noreply.github.com>
2026-05-21 17:43:48 +00:00
Hermes Agent 92f1896754 Stage 397: PR #2684 — fix: repair stale Codex OpenAI slash-qualified model state
Co-authored-by: ai-ag2026 <ai-ag2026@users.noreply.github.com>
2026-05-21 17:13:56 +00:00
nesquena-hermes c6587091a2 Stage 396: PR #2663 2026-05-21 00:26:54 +00:00
nesquena-hermes f867b4520b Stage 395: PR #2662 2026-05-21 00:14:45 +00:00
nesquena-hermes cc5f6e3a78 Stage 394: PR #2636 2026-05-20 23:53:04 +00:00
nesquena-hermes 45c7a693af Stage 394: PR #2625 2026-05-20 23:53:04 +00:00
nesquena-hermes 38933b288d Stage-394 follow-up: profile-switch reconciliation + a11y switch role + server-side chat/settings filter
Per deep-review verdict SHIP-WITH-FIXES on PR #2636:

1. Profile-switch reconciliation: _refreshProfileSwitchBackground now re-fetches
   /api/settings and re-applies hidden_tabs for the new profile. Without this,
   Profile A's hidden-tabs choice stayed in effect under Profile B until the
   user opened Settings → Appearance.

2. A11y: switched chips from role=button + aria-pressed to role=switch +
   aria-checked. The pressed/not-pressed wording confused screen-reader users
   because chip-off looks like the off state. Added role=group +
   aria-labelledby on the container, and a :focus-visible style on the chips.

3. Server-side belt-and-suspenders: api/config.py now strips 'chat' and
   'settings' from hidden_tabs at validation time, matching the client's apply-
   time filter. A tampered POST can no longer persist the forbidden values.

3 new regression tests added (chat/settings rejection, profile-switch wiring,
chip a11y attributes).

Co-authored-by: FrancescoFarinola <francesco.farinola@example.com>
2026-05-20 23:05:19 +00:00
Francesco Farinola 7f1feca3fe feat: sidebar tab visibility toggle in Settings > Appearance
Add chip row in Settings > Appearance that lets users toggle individual
sidebar/rail tabs on or off. Chat and Settings are always visible.

- Backend: hidden_tabs list setting with validation (no bool coerce)
- Frontend: pill chips that scan rail buttons, autosave via appearance
- Boot: _restoreTabVisibility IIFE applies hidden tabs before first paint
- i18n: 11 locales (label + description)
- Tests: 5 regression tests covering backend, frontend contracts,
  boot restore, i18n coverage, and settings session tracking
2026-05-20 22:57:36 +00:00
Michael Lam f17d4e204f fix: keep resumed CLI sessions in sidebar cap 2026-05-20 15:54:44 -07:00
nesquena-hermes feb35893b9 Stage 393: PR #2637
# Conflicts:
#	static/sessions.js
2026-05-20 22:24:40 +00:00
nesquena-hermes 4d8b8d0ffe Stage 393: PR #2633
# Conflicts:
#	CHANGELOG.md
2026-05-20 22:23:53 +00:00
nesquena-hermes e35c94bf55 Stage 393: PR #2615 2026-05-20 22:23:53 +00:00
nesquena-hermes aaf30b7b0a Stage 392: PR #2643 2026-05-20 21:48:04 +00:00
nesquena-hermes fa459aa01e Stage 392: PR #2651 2026-05-20 21:48:04 +00:00
nesquena-hermes b4a00b5aae Stage 392: PR #2650 2026-05-20 21:48:04 +00:00
nesquena-hermes dc0c833744 Stage 392: PR #2647 2026-05-20 21:48:04 +00:00
nesquena-hermes 6ed66daac2 Stage 392: PR #2638 2026-05-20 21:48:04 +00:00
Lumen Yang 71fbc796b2 fix: dedupe replayed context tail after compression 2026-05-20 23:15:54 +02:00
dobby-d-elf 6278222596 tighten session refresh invalidation 2026-05-20 14:40:13 -06:00
starship-s 153e035d12 fix: forward title generation api key 2026-05-20 14:39:38 -06:00
dobby-d-elf 14dd5aa00d address session event review 2026-05-20 14:33:36 -06:00
Colin Chang 9c3e37d2ee fix: custom_providers models allowlist takes priority over live /v1/models fetch
Custom providers that have a curated models: list in config.yaml
(e.g. ZenMux gateways) should show ONLY those configured models in
the picker dropdown, not the full /v1/models catalog.

Before this fix, _named_custom_groups unconditionally called
_read_custom_endpoint_models() which would pull hundreds of models
from aggregator gateways and overwrite the user's curated list.

Now the build checks if the custom_provider entry has a non-empty
models dict/list in config.yaml — if so, it skips the live fetch
and uses only the configured models (same behavior as hermes-agent
model_switch.py Section 4 patch).

Closes: configure-model-list-should-be-authoritative
2026-05-20 20:22:11 +00:00
dobby-d-elf 87527ff4f6 Fix state db legacy dedup repeat preservation 2026-05-20 14:18:47 -06:00
nesquena-hermes 1e3ca07575 Stage 390: PR #2634
# Conflicts:
#	CHANGELOG.md
2026-05-20 20:16:30 +00:00
nesquena-hermes 495991c2db Stage 390: PR #2642 2026-05-20 20:16:30 +00:00
Arsh Kumar Singh 2253cf5a32 chore: address review notes — dedup comment and 409-path clarification 2026-05-20 19:57:20 +00:00
Michael Lam 6e64068f0f fix: cap CLI session sidebar state scans 2026-05-20 12:47:03 -07:00
nesquena-hermes 7c2d56c920 Stage 389 follow-up: close TOCTOU race in pin-cap (Opus advisor #2614) 2026-05-20 18:12:38 +00:00
dobby-d-elf 19ad20afff Fix new chats using profile default model 2026-05-20 10:57:04 -06:00
nesquena-hermes dd36d09f89 Stage 389: PR #2626
# Conflicts:
#	CHANGELOG.md
2026-05-20 16:41:45 +00:00
nesquena-hermes 3d34eef02d Stage 389: PR #2620 2026-05-20 16:41:45 +00:00
nesquena-hermes 4d8e1ccc10 Stage 389: PR #2618 2026-05-20 16:41:44 +00:00
nesquena-hermes eaff4d0b8e Stage 389: PR #2614
# Conflicts:
#	CHANGELOG.md
2026-05-20 16:41:44 +00:00
nesquena-hermes 3bcd81b79f Stage 389: PR #2612
# Conflicts:
#	CHANGELOG.md
2026-05-20 16:41:44 +00:00
Arsh Kumar Singh d385db69d5 fix(clarify): require stable clarify_id and wait for backend ack so stale responses are rejected
The WebUI clarification popup had a response-delivery failure: users
submitted answers in the popup, but the agent still fell through to the
timeout fallback message.  Three bugs conspired:

1. No stable clarify_id — _ClarifyEntry had no unique identifier, so
   the frontend could not reference a specific pending prompt.  The
   backend used FIFO resolution which silently failed for stale/late
   responses.

2. Frontend hid the card before confirmation — respondClarify() called
   hideClarifyCard(true, 'sent') BEFORE the API call completed.  If the
   backend rejected the response, the card was already gone and the
   user's draft was discarded.

3. Backend lied about success — _resolve_clarify_legacy() returned
   bool(resolved) or not bool(clarify_id).  Since the frontend never
   sent clarify_id, the backend always reported ok:true even when
   nothing was resolved.

Changes:

api/clarify.py:
- _ClarifyEntry now auto-generates a stable clarify_id (uuid4.hex[:12])
- submit_pending() injects clarify_id into the data dict visible to the
  frontend via SSE and polling
- New resolve_clarify_by_id() for O(1) lookup by id instead of FIFO pop

api/routes.py:
- _resolve_clarify_legacy() uses resolve_clarify_by_id when clarify_id
  is provided; returns actual bool result (no more unconditional True)
- _handle_clarify_respond() returns HTTP 409 + {ok:false, stale:true}
  when resolution fails

static/messages.js:
- respondClarify() now sends clarify_id in the POST body
- Waits for a positive backend acknowledgement before hiding the card
- Saves a draft copy before POST and restores it on failure
- On 409/network error: re-enables controls, shows error toast
- Guards against parallel-SSE race where clearing the cache after a
  successful response could erase a newly queued next prompt (codex P1)

tests:
- Updated test_sprint30.py for new ack-before-hide behaviour
- Updated test_clarify_unblock.py for 409 on stale responses

Closes #2639.
2026-05-20 16:35:15 +00:00