mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-22 11:54:35 +01:00
28ed2f1db0eec8750be6ecb20c74694cb584d8e0
2
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
|
|
28ed2f1db0 |
fix(runner): support native OpenCode on macOS SSH hosts (#22534)
## Thinking Path > - Paperclip coordinates agents across local, SSH, and sandbox environments. > - Native Runner must launch artifacts for the execution host. > - A Linux controller currently selects Linux OpenCode assets for a macOS SSH host. > - Remote launch-file validation also accepts only ELF executables. > - The standard Core image must carry a verified Mac runtime for this host pair. > - This pull request adds that runtime and selects it for Apple Silicon SSH hosts. > - Self-hosted and hosted controllers can use the same generic implementation. ## Linked Issues or Issue Description **What happened?** A native OpenCode run on an Apple Silicon SSH host receives Linux controller artifacts. Its private launch-file verifier also rejects Mach-O files. A working legacy CLI agent does not exercise this path. **Expected behavior** The controller must select and verify the target's native daemon, Node, OpenCode executable, and proxy. Existing Linux targets and legacy agents must keep their current behavior. **Steps to reproduce** 1. Run a standard Linux Core image. 2. Select an SSH environment on an Apple Silicon Mac. 3. Start a native OpenCode agent with a compatible model connection. 4. Observe the incompatible runtime selection before useful agent work. **Paperclip version or commit** Base: `324ec1f537a9ba1e84239faa81bcef326a5dd1b7`. **Deployment mode** Docker or another Linux controller, with macOS ARM64 execution over SSH. Related work: #21662 covers local OpenCode confinement. #22236 covers SSH process cleanup. This change does not depend on either PR. No duplicate macOS OpenCode runtime PR was found. ## What Changed - Select a verified `darwin-arm64` runtime for OpenCode SSH targets. Preserve the Linux image default and explicit operator overrides. - Verify Mach-O launch files against the admitted architecture, file digests, and remote account. Keep existing Linux receipt and recovery behavior. - Build the Mac runtime on an Apple Silicon CI host with standalone Node 24.21.0. Keep OpenCode pinned to 1.18.34. - Package the sealed OpenCode runtime artifacts and required package metadata. Omit other harness distributions and their optional qualification inventories. - Verify archive provenance and safe extraction on Linux. Include the runtime in standard and merge-queue Core images. Avoid storing the transport archive in an image layer. - Reuse artifacts only from successful protected master image builds with identical runtime inputs. Test cache trust boundaries and register the script tests in CI. - Document runtime delivery, manual image inputs, and the initial Apple Silicon scope. No Cloud source, schema, or agent configuration changes. ## Verification - `pnpm -r typecheck` — passed. - `pnpm build` — passed. - Focused OpenCode artifact selection and private staging tests — 21 passed. - Focused native executor tests for OpenCode, SSH, and provider packs — 60 passed. - Real Rust daemon transport tests with the OpenCode protocol fixture — 4 passed. - Installed OpenCode 1.18.34 server qualification with a local provider fixture — 3 passed. - Runtime packaging, cache trust, workflow, image-contract, and merge-queue tests — 26 passed. - The CI release-workflow test group, including canary reuse and provenance — 218 passed. - Docker workflow and build-stamp guards — 29 passed after adding the Mac digest expectation, updating the excluded-source COPY expectation, and removing the redundant production commit ARG. - The source SDK registry, public-type, and evidence checks — 59 passed after updating only the root package metadata hash. Lock, compiler, parser, and model pins did not change. - Fresh-host provider-pack upload harness — 7 passed. The new regression first failed with `package_version:acpx`, then passed after the required metadata was preserved. - Node version policy — passed against a clean tracked-file snapshot. Generated local provider assets make the direct workspace scan fail; they are absent from the CI checkout. - Built, packaged, installed, and verified the actual Mac runtime. Both the controller verifier and the remote startup verifier accept its pins, target, sealed files, manifest, and package metadata. The pre-change private launch verifier rejects the real Mach-O files; the candidate accepts those same files. The dedicated archive is about 101 MB. - Live native smoke: real Mac daemon, private staged Mach-O files, OpenCode 1.18.34, and a local GLM-5.3 backend. The model read two files, edited the implementation, ran three tests, completed the semantic finish call, and summarized the three passes. An independent test run passed. The test file stayed unchanged. This used the existing `allow` mode and a scratch account-home fixture on a Mac. It does not prove hosted SSH dispatch. - The full CI workflow passed on `ae74808631f1933edec89aa1e80db7f2b756aca3`: server test shards, Runner Rust and native server tests, type checking, builds, browser suites, image checks, the native Mac producer, and Linux runtime installation. Greptile is 5/5 with no unresolved threads. - Local `pnpm test:run` used a disposable PostgreSQL 18 fixture after embedded Postgres failed to start because its macOS package lacks dynamic libraries. All six chat shards passed (1,971 tests). The redundant unsharded local run was interrupted during the general-server group after all current-head CI test jobs passed. The later local workspace and serialized groups did not run. The owned PostgreSQL fixture is stopped. This is not a local full-suite pass. Retained review failures: the first PR revision omitted package metadata and did not ignore downloaded image inputs. The fresh-host upload regression reproduced the metadata failure before its fix. Generated inputs are now ignored. CI also exposed a missing Node policy exception for the pinned Mac producer; the exception is limited to that producer job. A later policy attempt exposed an old exact-dependency assertion in canary reuse tests. The assertion now covers the added runtime and evaluates the publication gate for success, failure, cancellation, and skip outcomes. The Runner support suite then found that the new test registration changed a pinned root package input; the standard source registry generator updated that one hash, and all 59 affected tests passed. The server matrix then found four outdated image guard expectations and a duplicate production commit ARG. The focused image guards now pass after updating those expectations and keeping one commit ARG before Mac verification. Retained failures: the first producer attempt used Homebrew Node, which failed the relocation check. The successful producer used standalone Node 24.21.0. Live smoke attempts also encountered a stopped tunnel, an upstream model failure, and the existing Rust restriction on OpenCode permission requests in `ask` mode. The successful tool trace used the existing `allow` mode. Its initial reporter looked for tool names in generic item events; verification of the retained canonical tool events and an independent rerun confirmed the result. The upstream model failure is not explained by a later pass. Not run locally: a full Docker image build, because the existing Docker daemon does not respond. The PR adds native producer qualification and Linux installation verification to CI. Hosted deployment, remote service-account dispatch, and hosted continuation remain acceptance steps after deployment. ## Risks - A cold Mac runtime build adds CI time when Runner or producer inputs change. Reuse is bounded and accepts only a successful protected master producer. The artifact digest and runtime input identity remain verified. - Standard images gain a dedicated Mac runtime payload. Manual image builds can omit it; Mac OpenCode SSH runs then fail with a named missing-runtime error. Linux behavior remains available. - This release covers Apple Silicon SSH distribution. It does not qualify Intel Mac distribution or expand other harness support. - The OpenCode permission bridge keeps its existing behavior. This PR does not add `ask` approval support. - No saved run/session schema or provider pin changes. Existing Linux receipts retain their interpretation. A surviving admitted Runner bypasses new artifact discovery. Legacy agents are not converted. Existing account-change and old-receipt restoration tests remain in coverage. - No migration, cluster change, new sandbox image requirement, or Cloud control-plane change. A staging application rebuild and deployment are separate from this PR. ## Model Used OpenAI GPT-6, through Codex, with repository tools, code execution, tests, and SSH inspection. The session does not expose a more specific model ID or context-window size. ## 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> |
||
|
|
374a918d67 |
fix(runner): retain qualified OpenCode artifacts for cold recovery (#19622)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Native recovery must keep admitted provider state and useful workspace files. > - OpenCode binds its saved session to exact executable paths and hashes. > - A replacement image can remove those files even when session data is saved. > - Retaining a manifest alone cannot restore the missing launch files. > - This pull request saves the qualified files before provider work and restores the same owned paths. > - The benefit is cold continuity without relaxing the Runner guard or replaying a prior turn. ## Linked Issues or Issue Description Refs #19558. That fix retains the selected pack for ordinary continuation. This change adds durable file retention for a cold allocation. Related #19487 adjusts the admission budget; it does not retain launch files. A search of open and closed OpenCode cold/durable artifact PRs found no duplicate. **What happened?** A saved OpenCode session could outlive the image that supplied its Node executable, proxy, and native executable. A new image supplied different bytes. The Runner then correctly refused the changed launch profile. The saved harness did not contain the old files. **Expected behavior** Core must save qualified launch files before the first turn. It must restore those exact files under the same owned account and absolute paths. Unknown ownership, missing files, and changed hashes must still stop recovery. **Steps to reproduce** Start an admitted remote OpenCode session. Save its workspace, encrypted checkpoint, and harness data. Destroy the original allocation and controller files. Restore into a new allocation with an image whose qualified proxy has different bytes. Resume the same session. The new opt-in local qualification command automates this path with actual OpenCode and a local synthetic model. ## What Changed - Store company-bound, encrypted launch objects and canonical run/backup references. Generate migration 0438 through the normal database workflow. - Stage qualified Node, proxy, and native OpenCode files under a private content-addressed OS-account home. Record the account UID and exact launch paths. Preserve the Rust profile guard. - Gate first provider work on durable capture under the current native claim. Bound Stop/deadline waits while the actual late transfer keeps its worker and scratch ownership until settlement. Restore cold caches with verified staging and atomic rename. Keep live reattachment and historical image-rooted sessions on their existing paths. - Reuse bounded canonical expiration owners. Atomically fence old checkpoint authority with a strict company/run-scoped retirement marker before external deletion. Keep its indexed cleanup anchor across failed deletes and restart. Preserve earlier resume handles and reject later header or source-consumer writes. Add no idle scan or network call inside a transaction. - Add storage, race, executor, filesystem, and actual Rust resume regressions. Add the opt-in local Linux qualification command and operating limits in `doc/opencode-cold-recovery.md`. ## Verification - RED before repair: real PostgreSQL/AES cold restore lacked the launch files; normal executor preparation chose the new image pack. Separate regressions exposed lost upload receipts, missing cache repair, last-owner retirement, an account-UID mismatch, and a resume-handle race during external deletion. Each expected failure was captured before its repair. - Local qualification before the final master integration: 1,169 executor tests, 233 storage/archive/admission/expiration tests, all seven server typecheck layers, migration safety, and the actual Rust cold-resume test passed. - The maintained Linux command passed with two cached immutable Paperclip images whose proxy hashes differ. It removed the original account/controller trees, restored the production published cwd under a new lease, resumed the same actual OpenCode session, kept a useful file and prior model context, and made one new synthetic model request. No external model call or network was available in either container. - The object-store fixture uses real encrypted StorageService streams with an independent owned local directory. It is not hosted S3 acceptance or a complete task HTTP/actor workflow. Historical tasks still require normal effects reconciliation and recovery after deployment. - Review regressions reproduced partial-cache publication, resume-versus-deletion, fresh-writer and source-consumer resurrection, and unbounded Stop/deadline waits before each repair. They now pass through real filesystem, encrypted storage, PostgreSQL, and SDK boundaries. File validation is split from database ownership (384/831 source lines). - Clean runtime head `aadab7acc69ff959678030db46ffe2dcfbefe28f`: the maintained actual OpenCode cold qualification passed with `dirty:false`. It preserved the provider session and recorded launch paths, restored useful workspace content and prior model context, and made exactly one resumed synthetic model request. The 314,560,000-byte launch capture took 23.118 seconds; the later harness was 833,536 bytes. Subsequent repairs change backup-reference retirement and governed continuity-cache classification; capture, restore, launch, and published-workspace inputs are unchanged, so that runtime receipt remains scoped to those paths. The continuity-break consumer is verified separately. - Retention repair `c14d386cecf7123257918ba6a8d47307df2da397`: 68/68 affected real PostgreSQL/AES storage tests and 15/15 focused executor authority/admission cases pass. All seven server typecheck layers pass (source layers and unchanged test shards, with the final 590-file test layer repeated after fixture edits). All review threads have evidence-backed resolutions. Its subsequent CI caught the separate lifecycle-classification gap described below. The refreshed Paperclip review reported paused work, not a completed new review. - Final head `4d1cd3e3309e328b11abc85adc6d78c6d0c7f55d`: all 19 focused root-entry/executor cases and all three source declaration layers pass. [Greptile reviewed this exact head at 5/5](https://github.com/paperclipai/paperclip/pull/19622#issuecomment-6695984874). Final-head [CI verification](https://github.com/paperclipai/paperclip/actions/runs/42308908167/job/128214100339) passes: 60 successful checks, five conditional skips, and the explicitly neutral paused Paperclip review. No check is pending or failing. - CI also caught the launch cache missing from the session-root lifecycle contract. Both the writer scan and a real governed-replacement consumer failed locally before repair: old launch bytes remained in the active root. The cache now archives with the prior session state; the consumer verifies exact bytes are preserved in the archive and absent from the active root. - The latest review found that backup retirement could delete an independent run pin while a selected session retained the checkpoint. A real PostgreSQL/AES regression failed before the repair. Backup-only ownership now uses a NULL-safe exact backup predicate; the run pin belongs to fenced checkpoint expiration. The regression verifies checkpoint decryption and launch-file restoration after pending-backup cleanup, followed by deletion only after removing the final resume handle. - An executor first-session case checks new-session durable capture through the real capture helper and receipt propagation into the actual harness manifest at the backup boundary. Real storage cases verify the run-profile receipt and canonical backup/run pins. - Sequential closure reuse and simultaneous first admission are different scopes. A deterministic two-uploader test measures both independent uploads, one verified ready winner, loser-only deletion, and usable restoration for both runs. Each capture is bounded to 512 MiB and five minutes; up to three pending receipts per run are retained. There is no claim of global single-flight capture or a company-wide egress bound. - Earlier optional warm-ownership verification encountered missing local binaries. After exact-source setup, all nine ownership assertions passed; two other suites encountered local PostgreSQL setup/teardown contention, so those interrupted attempts are not passes. Required current-head CI supplies the broad final coverage. No deployment or hosted historical recovery is claimed. ## Risks - This adds two tables and a forward migration. The file object is bounded to 512 MiB. Capture adds one initial transfer; later harness archives exclude those large files. Company-scoped ready objects are shared and retained until the last canonical owner retires. Concurrent initial admissions can duplicate bounded transfers; each keeps independent ownership and cancellation rather than waiting on another controller. A continuity break retains the controller cache with its prior state until normal session-root cleanup, adding up to one bounded launch closure per retired state. - Cold continuity requires the same qualified Linux platform, supported runtime dependencies, owned account UID/home, and published absolute cwd. A mismatch fails closed. No alias or Rust guard change is included. - Existing image-rooted profiles have no retained file object. This change cannot backfill missing historical binaries. A fresh governed session must reconcile prior effects first. - Live reattachment does not read artifact objects. Storage or provenance failures before a new turn hold admission. They do not grant new execution, model, tool, or replay authority. ## Model Used OpenAI Codex, GPT-6 family, with reasoning, repository tools, code execution, and local browser tools. The exact served model ID and context window are not exposed in this session. No claim is made for unavailable model metadata. ## 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> |