* fix: coalesce ci auto-fix queue rounds
* fix: centralize ci auto-fix round cap
* fix: keep ci auto-fix state fresh
* fix: refine ci auto-fix round status
* fix: handle ci auto-fix exhaustion edge cases
* fix: reduce ci auto-fix dispatch complexity
* fix: refresh pending ci auto-fix before direct prompt
* fix: harden ci auto-fix queue replacement
---------
Co-authored-by: root <root@kandev.ts.net>
* fix(web): suppress kanban spinner for TODO tasks
A task moved back into TODO can still carry an active primary session
state (CREATED/STARTING/RUNNING). The kanban card was painting the
blue running-spinner in that case, making queued tasks look like they
were executing.
Add an explicit TODO guard at the top of shouldShowTaskRunningSpinner
so the queued column never shows the spinner, regardless of the
session-state desync. Existing decoupling for IN_PROGRESS/REVIEW/etc.
is untouched.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
* test: clarify spinner test description excludes TODO
Greptile flagged that "returns true when any task state has an actively
running primary session" overstates the invariant now that TODO
short-circuits to false. Reword to "non-TODO task states" so the
description matches behavior.
* ci: retrigger after artifact expiry
---------
Co-authored-by: Kandev Agent <agent@kandev.dev>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
* perf(backend): cap cumulative diff bytes and file count during large rebases
* perf(web): coalesce cumulative-diff refetches and skip redundant status compares
* feat(web): warn when the changes panel hides files from a capped diff
* fix(backend): enforce cumulative diff total byte budget strictly
* fix(web): keep coalesced refetch timer shared across envKey subscribers
* fix(web): drop unsound git-status SHA fast-path that staled live diffs
* fix(web): cancel coalesced timer on mid-flight drain to avoid duplicate refetch
* refactor(web): move changes-panel banner and helper to a sibling file
* docs(backend): clarify cumulative diff truncated-files-count semantics
* fix(web): tighten useSelectedFileKey mode param to the all|file union
* perf(kanban): memoize derived board data to stop re-render storms
The board rebuilt the `tasks` array (and `stepOptions`/`sharedProps`) on every
store change, producing fresh object identities that defeated downstream memos
and re-rendered every card on each WebSocket message. Wrap them in useMemo so
the board only re-renders when its inputs actually change.
* perf(session-runtime): bound runtime memory and dedupe git-status compare
Three runtime-store fixes for the crash a heavy rebase triggered:
- Cap shell + terminal output buffers at 2MB (matching the existing process
buffer) so a chatty shell can't grow the store without limit.
- Purge per-session runtime state (messages, turns, shell/process buffers, and
the bySessionId maps) when a session is removed; env-scoped maps are dropped
only once no other session shares the environment.
- Run setGitStatus's deep per-file diff comparison once per update and return
whether anything changed, instead of recomputing it 2-3x. Under a massive
rebase (thousands of files, frequent updates) that comparison was the
dominant CPU cost and drove the render storm.
* refactor(session-runtime): drop now-unused gitStatusWouldMutate
The git-status handler no longer pre-checks for changes (setGitStatus now
returns whether it changed), leaving gitStatusWouldMutate with no callers.
* fix(session-runtime): enforce terminal cap on first and oversized chunks
appendTerminalChunk only evicted when more than one chunk was buffered, so a
single chunk larger than the 2MB cap slipped through untrimmed; the new-terminal
branch also pushed the first chunk without going through the cap at all. Clamp
each chunk to the tail of the cap and route the new-terminal path through
appendTerminalChunk. Addresses CodeRabbit review.
* fix: scope multi-repo file editor open/save to the right repo
The editor open/save/sync/delete chain dropped the repository_name in
multi-repo task workspaces, so the backend resolved paths against the
bare task root and failed with "file not found". Thread the repo
subpath through openFile, save, sync, and delete, and store it on
FileEditorState.
* fix: pass repo from diff Edit button to multi-repo file editor
The diff/changes "Edit" button opened files without their repository
scope, so multi-repo files resolved against the bare task root and
failed with "file not found". Thread repository_name through the diff
and review toolbars into onOpenFile.
* fix: preserve repo across tab restore and stale syncs
Address review feedback: persisted file tabs dropped the repo subpath,
so a multi-repo file reopened against the bare task root after a browser
refresh. Persist and restore `repo` on StoredFileTab. Also guard the
workspace sync against a tab being swapped to a different repo mid-fetch,
and widen the remaining onOpenFile type declarations to carry repo.
* fix: thread repo into editor panel fallback fetch
Address review feedback: FileEditorPanel's useFileLoader fallback fetch
(triggered when the open-files map is transiently empty, e.g. after a
session switch) called requestFileContent without the repo subpath,
re-introducing the multi-repo "file not found" failure. Record repo on
the panel params and forward it through useFileLoader. Also assert the
full restored-tab shape in the local-storage round-trip test.
* fix: guard editor fetches against stale session/repo races
Address review feedback: file open and the panel fallback fetch could
apply a late response after the user navigated away. openFile now drops
the result if the active session changed mid-fetch (via fetchFileEditorState),
and the panel loader keys its in-flight guard by session+path+repo so a
changed target re-fetches and stale responses are discarded.
* feat(stats): add avg turn duration and messages per turn to averages card
Adds two new metrics to the workspace stats Averages card: mean turn
duration and mean message count per turn. Both exclude outliers — turns
shorter than 1s (no-op/aborted), longer than 1h (zombie turns where
completed_at was backfilled across an agent restart), and turns with no
messages — so the displayed means reflect typical work rather than tail
skew. On a ~4k-turn sample the filter drops avg duration from ~1062s to
~289s.
* fix(stats): exclude exact 1h turns and use Tooltip for averages help text
Address PR feedback: the BETWEEN bound was inclusive so turns exactly at
the 1h ceiling were retained, contradicting the documented half-open
range. Also swap the title attribute on the new Averages rows for the
shared Tooltip component so the help text is reachable via keyboard and
screen readers.
* fix(stats): align turn-row zero state with em-dash convention
Address PR feedback: avg_messages_per_turn rendered "0.0" while
avg_turn_duration_ms rendered "—" under the same empty-data condition.
Also corrects the GetGlobalStats docstring to acknowledge that
clean_turn_agg adds one correlated per-turn lookup on top of the four
single-scan CTEs.
* test(stats): cover exact 1h turn boundary in clean-turn exclusion test
Add a turn of exactly 3,600,000ms (cleanTurnMaxDurationMs) to
TestGetGlobalStats_CleanTurnExcludesOutliers so the exclusive upper
bound is exercised directly, not just via the 2h zombie turn that
sits well above the boundary.
---------
Co-authored-by: Kandev Agent <agent@kandev.dev>
Carlos Florêncio