Repository navigation
fix(session): delete reverted messages boundary-last and tie-break ids by raw order - #52588
Open
austinborn wants to merge 3 commits into
Open
austinborn wants to merge 3 commits into
austinborn wants to merge 3 commits into
Conversation
- Update .actual/rules/cross-cutting-message-ingestion-synchronization-layers-e283.md (claude) - Update .actual/rules/cross-cutting-session-history-compaction-routines-reve-df90.md (claude) - Update .actual/rules/cross-cutting-session-message-timelines-transcript-exp-f4fc.md (claude) - Update AGENTS.md (agents) - Update docs/adr/e2838170-cdcd-4036-ae98-5e997b53d7cd-standardize-chronological-ordering-for-session-message-ingestion-and-timelines-message-ingestion-synchronization-layers-not-assume.md (docs) - Update docs/adr/df903f58-c272-4a3f-b640-a93922945540-standardize-chronological-ordering-for-session-message-ingestion-and-timelines-session-history-compaction-routines-revert-boundary.md (docs) - Update docs/adr/f4fcb389-bd12-4d7e-8adc-0d32c57f98c4-standardize-chronological-ordering-for-session-message-ingestion-and-timelines-session-message-timelines-transcript-exports-sync.md (docs)
Rule Files Sync
…s by raw order Revert cleanup removed the boundary message first, so an interrupted cleanup left no marker and the next cleanup cleared the revert while resurrecting the remaining reverted messages. Delete messages and parts newest-first so the boundary is removed last and an interrupted cleanup stays resumable. The TUI and share page tie-broke same-millisecond messages with localeCompare, which disagrees with storage's SQLite BINARY collation (and with the TUI's own binary search) and can treat distinct ids as equal. Compare ids by code unit. Refs anomalyco#42816 Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Contributor
|
The following comment was made by an LLM, it may be inaccurate: |
This branch has not been deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Issue for this PR
Closes #42816 (items 1 and 2). Supersedes #42819, which was auto-closed as stale. The completion-gate fallback from the same issue is handled separately in #48472.
Type of change
What does this PR do?
Revert cleanup.
SessionRevert.cleanupdeleted the revert boundary message first, and each removal is its own write. If cleanup was interrupted after that, the next cleanup couldn't find the boundary (findIndex→ -1), deleted nothing, andclearRevert()dropped the marker. The rest of the reverted messages then stayed in the transcript. Messages and parts are now deleted newest-first, so the boundary goes last and an interrupted cleanup can still be finished by the next one.Tie-break collation. Storage pages with
ORDER BY time_created, id(SQLite BINARY). The TUI (compareMessage,listSessions) and the share page tie-broke onlocaleCompare, which can order ids from the same millisecond the other way (e.g.msg_A…vsmsg_a…) and returns 0 for canonically equivalent distinct ids. In the TUI this also disagreed with its ownsearch(), which uses<. All three now compare by code unit.How did you verify your code works?
packages/opencode/test/session/revert-compact.test.ts: boundary-last removal order (messages and parts), and a cleanup that resumes after a partial delete.packages/tui/test/context/sync.test.ts: case-differing ids sort in storage order whatever the input order, and canonically equivalent distinct ids never compare equal.bun testfor the session revert/message-v2/pagination suites (100 pass) and the TUI context tests;bun typecheckacross the repo (pre-push hook).Screenshots / recordings
N/A, no UI change.
Checklist
🤖 Generated with Claude Code