mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-11 13:24:55 +01:00
eaacadeba3e27c5e3ec0e8b8d453c229e047eba9
4
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 DottaandPaperclip
|
1678aae862 |
fix: expire unused bases with generated repository ignore metadata (#16695)
## Thinking Path
> - Paperclip is the open source app people use to manage AI agents for
work.
> - Finished tasks can release Git caches that Paperclip can restore
later.
> - Nested repository preparation adds an ignore entry to the base
repository.
> - Base cleanup currently treats this generated entry as custom
metadata.
> - This pull request accepts the exact generated entry and keeps other
metadata protected.
> - Unused private bases can then expire with their workspaces and
return on demand.
## Linked Issues or Issue Description
Refs #16691. Related #15404 and #14443 change repository preparation or
sync, not base expiration.
**What happened?**
A finished workspace with a nested project repository expires, but its
unused private base remains with `custom_git_metadata`. Normal
preparation added `/.paperclip-repositories/` to the stock Git exclude
file.
**Expected behavior**
Cleanup removes a verified, unused base with this generated entry.
Custom metadata still prevents base deletion.
**Steps to reproduce**
Create an isolated workspace with a nested repository through normal
preparation. Finish the task, wait for the retention period, and run
expiration. The workspace disappears but its private base remains.
**Paperclip version or commit**
|
||
|
DottaandPaperclip
|
790b729086 |
fix: restore nested project repositories and settings after expiration (#16691)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Agents use isolated Git workspaces with managed project repository copies. > - Finished workspaces can expire when Git can reproduce their contents. > - The saved recipe records the main commit but omits nested commits and local settings. > - Reopening can therefore use newer nested code and lose merged user settings. > - This pull request saves and restores those contents before workspace provisioning. > - It also permits verified FETCH_HEAD metadata so ordinary fetches do not block safe expiration. ## Linked Issues or Issue Description Refs #16457, #16586, #16621. Related performance work: #16579. **What happened?** After a finished workspace expires, its nested project checkout can reopen at the current upstream commit instead of its previous commit. Cleanup also treats local Claude settings as disposable, although that file can contain user settings. Ordinary linked-worktree FETCH_HEAD metadata prevents otherwise safe cleanup. **Expected behavior** Restore the main and nested repositories at their saved commits. Preserve local settings bytes and permissions. Skip files and Git state that cannot be reproduced. Permit standard fetched metadata only when its commits remain recoverable. **Steps to reproduce** 1. Prepare a managed nested project repository at commit A and add ignored local settings. 2. Advance upstream to B, finish the task tree, and expire its workspace and base. 3. Reopen the workspace through the normal service path. Before this fix, nested preparation selects B. ## What Changed - Extend the existing restoration recipe with exact nested commits, branches, source identity, and encrypted local settings. Keep the schema compatible with older recipes. - Restore nested repositories and settings before provisioning. Recheck company/project source access and preserve conflicting files on retry. - Recheck recorded contents under cleanup locks before deletion. Persist lock ownership before atomic publication so process exits recover safely and foreign locks stay intact. Preserve changed or unsupported contents, including tracked settings. - Keep saved project repository selections through normal preparation. Checkpoint restored contents before provisioning so failed setup can retry without overwriting later edits. - Accept bounded linked-worktree FETCH_HEAD records only after local ancestry or fresh upstream proof. Preserve local-only fetched commits. - Add real Git/Postgres lifecycle cases and focused content, access, publication-retry, and fetched-state controls. Document the behavior. ## Verification - `pnpm -r typecheck` passed. - `pnpm build` passed. - The original real Git/Postgres experiment failed with nested B instead of A. The same path passes with this change. - The integrated lifecycle suite passed 42 cases in 91.66s. Related preparation, credential, content, and Git metadata controls passed 46 cases in 24.41s. The final heartbeat checkpoint persistence case passed separately in 8.17s. - Coverage includes exact main/nested commits, encrypted settings and modes, upstream failure and retry, provisioning order, concurrent reopening, changed unique files, tracked settings, malformed metadata, and company boundaries. - Exact-head CI passed the implementation, typecheck, build, security, and test matrices. One browser shard passed on a single rerun after an unrelated fixture teardown raced an activity-log write; its feature assertions passed in the initial attempt. - The broad local test run was stopped after about 12 minutes of unrelated serial tests. Full parallel CI is required before merge; no local full-suite pass is claimed. - Single-workspace deployment, real-backlog reclamation, and ordinary product reopening remain required acceptance work after merge. ## Risks - Expired Git repositories still depend on upstream access. A failed restoration returns the existing retryable error and never substitutes the default branch. - Saved settings use the existing local encryption provider. Recovery requires that encryption key to remain available. - Unsupported files, tracked settings, unknown sources, and unverified Git state remain on disk. This can reduce reclaimed space. - No database migration, recovery-retention change, public API, or fleet-default change is included. ## Model Used OpenAI Codex, GPT-6 family, with tool use, code execution, and an independent agent review. The runtime does not expose a more specific model identifier. ## 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 - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
2aca41464b |
fix(workspaces): preserve compaction and expire standard Git bases (#16476)
## Thinking Path > - Paperclip manages AI agents and their task workspaces. > - Finished tasks can release reproducible Git files and restore them later. > - Unique work must stay on disk, while regenerable dependencies can be removed. > - Linux Git templates include an empty legacy directory that blocked base cleanup. > - The new expiration branch also skipped the existing guarded dependency cleanup. > - This change accepts the standard empty directory and retains that cleanup order. > - Private bases can expire, and protected workspaces still shed regenerable files. ## Linked Issues or Issue Description Refs: #16457 Live acceptance of #16457 removed a worktree but retained its private base because Linux Git created an empty `.git/branches` directory. A pending-base dry run then tried to resolve the already removed worktree and hid the preservation reason. A local integration comparison also showed that enabling Git expiration skipped existing dependency compaction. It deferred pending startup compaction by an hour. Both failures were reproduced before the fix. ## What Changed - Accept `.git/branches` only when it is an empty real directory. Preserve populated entries, files, and symlinks. - Inspect pending base cleanup without resolving an absent worktree. - Run existing guarded compaction before Git expiration. Exclude already closed rows from compaction and retain its pending deadline. - Keep failed expiration checks from reaching legacy whole-workspace deletion. Existing compaction remains independent of Git expiration dry-run mode. - Extend real Git/PostgreSQL tests and retain the acceptance failures in the delivery state note. ## Verification - Linux-template baseline: private-base expiration failed before the fix. The first sandboxed attempt could not bind PostgreSQL; the authorized retry reproduced the behavior. - Compaction baseline: two regression cases failed in 7.67 seconds. They reproduced skipped dependency cleanup and the lost startup deadline. - After the fixes, `pnpm exec vitest run server/src/__tests__/execution-workspace-compaction.test.ts server/src/__tests__/git-workspace-expiration.test.ts server/src/__tests__/terminal-workspace-work.test.ts` passed all 49 cases in 43.79 seconds. - Controls cover unique work, active runs, unfinished descendants, previews, concurrent reopening, interrupted removal, exact-commit restoration, and retryable upstream failures. - Full `pnpm -r typecheck` and `pnpm build` passed for the final commit. All exact-head CI and security checks passed. Greptile is 5/5 with all findings resolved. - Final staging acceptance will let the restarted worker resume the retained base cleanup, then reopen normally at the recorded commit and persist subsequent work. The base has not been manually removed. The superseded preview build was cancelled before dispatching a replacement. ## Risks - Legacy branch definitions can hold unique data. Only an empty real directory qualifies for deletion. - Compaction preserves its existing lifecycle guard and regenerable-directory allowlist. It cannot fall through to legacy deletion after an expiration failure. - Already closed workspaces must only resume their recorded expiration; they do not enter compaction. ## Model Used OpenAI Codex, GPT-6.1-sol, with reasoning, code execution, Git, database, and browser tools. The exact context-window size is not exposed by this session. ## 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 - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
|
DottaandPaperclip
|
59a70dc9e2 |
fix(workspaces): expire reproducible Git files and restore exact commits (#16457)
## Thinking Path > - Paperclip manages AI agents and their task workspaces. > - Isolated Git worktrees and private repository copies remain on the host after tasks finish. > - These files can fill the disk even when every commit is available upstream. > - Cleanup must retain unique files, local Git state, live work, and recovery data. > - This pull request records and verifies an exact-commit restoration recipe before expiration. > - Normal reopening restores that commit even after upstream advances. > - The benefit is lower host disk use while finished tasks remain available for later work. ## Linked Issues or Issue Description Refs #10285 and #2022. Follows #16433 for skill-cache expiration. Related #11936 provides separate host-local cleanup tooling; this change extends the existing company-scoped lifecycle worker and normal workspace reopening. ROADMAP.md does not specify a competing implementation. ## What Changed - Add opt-in expiration for verified Paperclip-owned Git worktrees after their whole task tree has been done or cancelled for 24 hours. - Preserve active and scheduled runs, previews, pending reopening, unresolved recovery, dirty files, unknown ignored files, and unsupported or unique Git state. - Verify an authenticated fetch in a fresh object store. Save repository identity, exact commit, branch, and ownership before any deletion. - Rename eligible files into recorded managed trash. Resume interrupted cleanup under source and lifecycle locks. - Expire private base clones only after consumer, file, reference, reflog, and object checks pass. Preserve shared project base clones. - Rebuild missing bases and worktrees through normal reopening at the recorded commit. Return an explicit retryable error when repository access fails. - Add independent workspace/base switches, dry-run reports, reclaimed-byte counts, and operational documentation. No table, public API, UI, or recovery-retention change. ## Verification - Real local bare remote plus Postgres: create at A, advance upstream to B, expire the worktree and private base, reopen at A, then make and retain a subsequent commit. - Final lifecycle suite: all 20 tests passed in 46.41 seconds at `d72947a2e45be5d0c912debb03c46b15863b4798`. Review fixes cover second-cycle expiration, newer-branch recovery, normal preparation after restoration, branch adoption, and Git administration path controls. Eleven existing branch reconciliation controls also passed. Covers unique files and Git state, live descendants, scheduled runs, recovery, previews, unauthorized access, upstream failure, concurrent reopening, interrupted deletion, shared bases, and nested private worktrees. - Final guard regressions: 3 affected tests passed. Existing workspace/runtime/preparation/source controls: 328 tests passed in 238.69 seconds. - Full repository typecheck and build passed locally. The full local test attempt did not pass: initial native-session cases hit macOS Postgres shared-memory exhaustion, source changes invalidated two discovery cases, and the checkout disappeared near the end. The final-code retry of the remaining skill and unrelated controls passed all 103 tests. All checks passed on this exact head, including typecheck, general and serialized tests, Runner checks, build, canary dry-run, browser E2E, and security scans. Greptile reviewed the head at 5/5 with all findings resolved. - Fixed a blocking CI policy false positive for `docker-compose-secrets.test.mjs`. The test still rejects GitHub secret expressions, including dot, bracket, and multiline forms. Its full 15-test suite passes. - Map validation and `git diff --check` passed. - Staging dry-run inspection, measured reclamation, and ordinary product acceptance remain pending. No live acceptance is claimed. ## Risks - Git restoration depends on continued upstream availability and credentials. Cleanup skips unverified commits; restoration never substitutes the default branch. - File deletion requires multiple conservative proofs. Unknown or unsupported state is retained and can limit reclamation. - Source and lifecycle locks serialize preparation, reopening, and expiration. Remote verification has a deadline but can delay work on the same source. - Start with both expiration switches disabled. When enabled, dry-run remains the default. Rollback disables further deletion while retaining restoration support. ## Model Used OpenAI GPT-6 through Codex, with reasoning, repository inspection, tool use, and code execution. The exact backend model ID and context-window size are not exposed in this session. ## 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 (focused tests; full local limitation documented above) - [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 - [x] All Paperclip CI gates are green - [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing> |