jpmschweitzer/clide
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

pql: retry transient db-busy so sidebar panes don't stick (T-350)

Browse Source 
The pql-backed sidebar panes fetch once on first build. If that fetch
fired too early — the planning DB still settling at startup, or a db-busy
SQLite lock under concurrent pql writes (pql exits 69) — the pane showed
"pql … failed" and stayed there until a manual refresh re-fired it.

Retry transient failures at the single chokepoint, PqlClient._run: on a
busy/locked signal (exit 69, or stderr mentioning database is locked /
busy) retry a few times with short backoff before throwing. Genuine
errors aren't busy, so they still surface immediately.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
jpmschweitzer
2026-06-10 20:20:36 +02:00
co-authored by Claude Opus 4.8
parent b648523a1a
commit 9a8175903b
5 changed files with 106 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.pql/changelog/ticket_history/2026-06.sql
+1
View File
@@ -3510,3 +3510,4 @@ INSERT INTO ticket_history (ticket_record_id, field, old_value, new_value, chang
INSERT INTO ticket_history (ticket_record_id, field, old_value, new_value, changed_by, changed_at, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FB4XCM5KBXDDSCWJ37GPYG3R', 'status', 'backlog', 'done', NULL, '2026-06-10 16:57:07', '2026-06-10 16:57:07', '2026-06-10 16:57:07', NULL, '1ce291d54831d3ede687384037c379ff', 2) ON CONFLICT(hash) DO NOTHING;
INSERT INTO ticket_history (ticket_record_id, field, old_value, new_value, changed_by, changed_at, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FB50YE6S6YWNP2ZSFWES9B2W', 'status', 'backlog', 'done', NULL, '2026-06-10 17:13:53', '2026-06-10 17:13:53', '2026-06-10 17:13:53', NULL, '87b069d3f25c91622c9720147d359360', 2) ON CONFLICT(hash) DO NOTHING;
INSERT INTO ticket_history (ticket_record_id, field, old_value, new_value, changed_by, changed_at, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FB58X0TFJ02YTMVPD0D9Q838', 'status', 'backlog', 'done', NULL, '2026-06-10 17:47:01', '2026-06-10 17:47:01', '2026-06-10 17:47:01', NULL, 'ba3d17ec4e668ade82b07d2bb848ab91', 2) ON CONFLICT(hash) DO NOTHING;
INSERT INTO ticket_history (ticket_record_id, field, old_value, new_value, changed_by, changed_at, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FB5CW7JPT6BR2RWMNYVCXJ50', 'status', 'backlog', 'done', NULL, '2026-06-10 18:20:28', '2026-06-10 18:20:28', '2026-06-10 18:20:28', NULL, '213109fcd67b4375ebbd3b59c4a1ed04', 2) ON CONFLICT(hash) DO NOTHING;
.pql/changelog/tickets/2026-06.sql
+6
View File
@@ -3213,3 +3213,9 @@ INSERT INTO tickets (record_id, type, parent_record_id, title, description, stat
So a build host/container missing the Wayland client dev headers (very plausible in a Bazzite/immutable distrobox or toolbox) silently compiles out the decoration suppression — the rebuilt app ships with the compositor''s native title bar (double title bar on KDE Plasma Wayland). Reported live on Bazzite KDE.
Fix: make wayland-client a hard build requirement — fail the CMake configure with a clear, actionable message (name the package: Fedora wayland-devel, Debian/Ubuntu libwayland-dev) instead of dropping the feature. Frameless chrome is a core guardrail; never ship without it. Also fix the stale ''xdg-decoration'' comment (the code uses the KDE server-decoration protocol, not xdg-decoration).', 'done', 'high', NULL, NULL, NULL, '2026-06-10 17:45:28', '2026-06-10 17:47:01', NULL, '6eb6120c0ef64a7f5ce78073b398143c', 2) ON CONFLICT(record_id) DO UPDATE SET type=excluded.type, parent_record_id=excluded.parent_record_id, title=excluded.title, description=excluded.description, status=excluded.status, priority=excluded.priority, assigned_to=excluded.assigned_to, team=excluded.team, decision_ref=excluded.decision_ref, updated_at=excluded.updated_at, deleted_at=excluded.deleted_at, hash=excluded.hash, canonical_version=excluded.canonical_version WHERE excluded.updated_at > tickets.updated_at OR (excluded.updated_at = tickets.updated_at AND excluded.hash > tickets.hash);
INSERT INTO tickets (record_id, type, parent_record_id, title, description, status, priority, assigned_to, team, decision_ref, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FB5CW7JPT6BR2RWMNYVCXJ50', 'bug', '06FB0TNQM5TWC00GW0P3X02HZW', 'pql sidebar panes stick on error when the first fetch fires too early', 'The pql-backed sidebar panes (tickets, decisions, pql, search) fetch once on first build via pql.* IPC. If that first fetch hits a TRANSIENT pql failure — the planning DB still settling at startup, or a db-busy SQLite lock under concurrent pql writes (pql 1.10 exits 69 ''db busy'', as seen in the serial test suite) — the pane shows ''pql ticket failed'' and stays there until a manual refresh / tab-switch re-fires the fetch. Reproduced live on 2.3.0: left pane ''pql ticket failed''; works on manual refresh. pql works fine in isolation, so it''s purely a too-early / transient timing issue with no retry.
Fix: make pql invocations resilient to transient failures at the single chokepoint, PqlClient._run (lib/src/pql/client.dart) — on a busy/locked signal (exit 69, or stderr mentioning database is locked / busy), retry a small bounded number of times with short backoff before throwing PqlException. Keep genuine errors fast (don''t blanket-retry every non-zero). Fixes all pql panes at once.', 'backlog', 'high', NULL, NULL, NULL, '2026-06-10 18:02:50', '2026-06-10 18:02:50', NULL, 'd03c864f5e64284f3582e59556a4ab4f', 2) ON CONFLICT(record_id) DO UPDATE SET type=excluded.type, parent_record_id=excluded.parent_record_id, title=excluded.title, description=excluded.description, status=excluded.status, priority=excluded.priority, assigned_to=excluded.assigned_to, team=excluded.team, decision_ref=excluded.decision_ref, updated_at=excluded.updated_at, deleted_at=excluded.deleted_at, hash=excluded.hash, canonical_version=excluded.canonical_version WHERE excluded.updated_at > tickets.updated_at OR (excluded.updated_at = tickets.updated_at AND excluded.hash > tickets.hash);
INSERT INTO tickets (record_id, type, parent_record_id, title, description, status, priority, assigned_to, team, decision_ref, created_at, updated_at, deleted_at, hash, canonical_version) VALUES ('06FB5CW7JPT6BR2RWMNYVCXJ50', 'bug', '06FB0TNQM5TWC00GW0P3X02HZW', 'pql sidebar panes stick on error when the first fetch fires too early', 'The pql-backed sidebar panes (tickets, decisions, pql, search) fetch once on first build via pql.* IPC. If that first fetch hits a TRANSIENT pql failure — the planning DB still settling at startup, or a db-busy SQLite lock under concurrent pql writes (pql 1.10 exits 69 ''db busy'', as seen in the serial test suite) — the pane shows ''pql ticket failed'' and stays there until a manual refresh / tab-switch re-fires the fetch. Reproduced live on 2.3.0: left pane ''pql ticket failed''; works on manual refresh. pql works fine in isolation, so it''s purely a too-early / transient timing issue with no retry.
Fix: make pql invocations resilient to transient failures at the single chokepoint, PqlClient._run (lib/src/pql/client.dart) — on a busy/locked signal (exit 69, or stderr mentioning database is locked / busy), retry a small bounded number of times with short backoff before throwing PqlException. Keep genuine errors fast (don''t blanket-retry every non-zero). Fixes all pql panes at once.', 'done', 'high', NULL, NULL, NULL, '2026-06-10 18:02:50', '2026-06-10 18:20:28', NULL, 'd48c86ad83f3da7cd4f8c9308c852448', 2) ON CONFLICT(record_id) DO UPDATE SET type=excluded.type, parent_record_id=excluded.parent_record_id, title=excluded.title, description=excluded.description, status=excluded.status, priority=excluded.priority, assigned_to=excluded.assigned_to, team=excluded.team, decision_ref=excluded.decision_ref, updated_at=excluded.updated_at, deleted_at=excluded.deleted_at, hash=excluded.hash, canonical_version=excluded.canonical_version WHERE excluded.updated_at > tickets.updated_at OR (excluded.updated_at = tickets.updated_at AND excluded.hash > tickets.hash);