mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-12 06:10:44 +01:00
DottaandClaude Opus 5.5
bd97a3e6f7
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - A follow-up turn on a per-turn session must retire the previous turn's runner state before it starts a new runner. Retirement proves the previous owner finished, and it also proves that every quarantined copy of that session root is settled > - If restart recovery interrupts a run's own start (for example a hard kill during `session.open`), recovery proves the half-started bootstrap held no provider work, quarantines that root, and bootstraps the same run again. The quarantined copy keeps the run's identity > - After that run finished, retirement checked the old bootstrap copy as if it were a finished owner. An unfinished `session.open` (or an opened session without a turn) is never "settled", so the follow-up failed with `runner_state_identity_mismatch: prior_owner_terminal_state_indeterminate` and needed a Retry > - This pull request skips a quarantined copy of the same owner when the same bootstrap proof that let recovery quarantine it still holds > - The benefit is that a follow-up after a restart-interrupted start continues on its own ## Linked Issues or Issue Description Refs #20137 **What happened?** Local restart-chaos K13 regression storm w-20261108T235957-s9401 (master4420a5adcf): 102 of 105 tasks passed. Two of the three failures share this cause. - **t58 (codex C1):** - A K1 at 00:15:42 interrupted run `78dd97f1` during its first `session.open`. Recovery quarantined that root (`run.prepare` completed, `session.open` pending, no events) and bootstrapped again. - The run then completed. A second K1 hit 190 ms after its result was accepted, and recovery finalized the result. - Follow-up `def668a6` failed with `runner_state_identity_mismatch: prior_owner_terminal_state_indeterminate; retirement_refused=quarantine_journal_commands_unsettled(session.open.pending)`. - **t67 (claude C1):** the same pattern. The quarantined copy of run `990936c0` holds a completed `session.open` with only session-open events and no turn, so the refusal was `quarantine_journal_terminal_unsettled`. Both copies are in the kept instance's quarantine directory under the previous run's own identity. Each Retry succeeded. **Expected behavior** A quarantined copy of the finished owner's own earlier bootstrap that provably held no provider work does not hold the follow-up. Retirement proceeds and the follow-up starts without a Retry. **Steps to reproduce** 1. Let a per-turn run's first bootstrap be quarantined after a restart (the run bootstraps again and finishes). 2. Start a follow-up turn. 3. Before this change, retirement refuses with `quarantine_journal_commands_unsettled(session.open.pending)` or `quarantine_journal_terminal_unsettled`. This is not a code regression. The quarantine-copy check dates from #19061. Today's recovery fixes let more interrupted starts bootstrap again within the same run instead of blocking, and each such restart leaves this kind of copy behind. **Paperclip version or commit**4420a5adcf(wave). Branch rebased on current master. ## What Changed - `native-session-executor.ts`: the quarantine loop in `inspectFinishedPerTurnRunnerdOwner` skips a copy when both of these hold: - its durable identity matches the finished owner exactly; - `runnerdStateProvesIncompleteBootstrap` holds for it. That is the proof restart recovery uses before it quarantines an interrupted bootstrap: no provider turn, and for a local runner, its journal and processes proven gone. Every other copy still has to pass the settled-projection proof. - `native-recovered-owner-followup.test.ts`: a red/green test covering both cases: - an interrupted bootstrap copy is retired past; - a copy that had already issued a turn still holds the follow-up. ## Verification - Red/green: on master, the new admitted case fails with exactly the t58 error. With the change it is admitted. The turn-issued copy is refused on both. - `vitest run` of `native-recovered-owner-followup.test.ts`, `native-session-executor.test.ts` and `native-runner-restart-recovery.integration.test.ts`: 3 files, 1504 tests passed (14 skipped). - `PAPERCLIP_TYPECHECK_LOW_MEMORY=1 pnpm typecheck` (`server`): passed. ## Risks - Low risk. The skip reuses the existing bootstrap-incomplete proof and requires an exact identity match, so a foreign or work-holding copy still blocks. Nothing is deleted; the copy stays in quarantine. - Not covered: t22 in the same storm, a follow-up whose own unconnected `run.attach` root was hit by a K1 (`authority_indeterminate`), has a different shape and is tracked separately. ## Model Used - Claude Opus 5.5 (`claude-opus-5-5`), via Claude Code with tool use (shell, file edits, test execution). ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [ ] All Paperclip CI gates are green - [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>