Commit graph

11 commits

Author SHA1 Message Date
Scott Werner
3ab584691d Drop the branch-head compatibility wrapper
`branch_head_sha` existed to keep callers compiling while the branch-head
lookup moved onto the repository reader. Its last caller now opens a
reader directly, so the wrapper only made the typed API worse: it joined
an already-validated owner and repo into a slug so the reader could split
them apart again, invented an "invalid repository coordinate" error for a
value validated upstream, opened a fresh credential session per call, and
flattened `RepositoryReadError` into `anyhow` while keeping one variant —
leaving callers unable to tell a rate limit from a rejected token.

Its integration test pinned the wrapper rather than the behavior. Replace
it with one that asserts the same 404-means-not-observable semantics
through `resolve_commit`, which is where that contract actually lives.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-19 15:43:08 -04:00
Scott Werner
8dfda7ede0 Reduce duplication in GitHub repository reads
Share the twin's handler scaffolding, collapse the reader's parallel URL
and error machinery, and resolve branch-head credentials once per verify.

Twin GitHub server:
- Add handlers/support.rs holding the response envelope, installation-token
  authorization, Accept matching, and commit-SHA checks. The commits and
  contents handlers carried byte-identical copies of all six items, and
  pulls.rs had its own copy of the two response mappers.
- Add AppState::find_repository and repository_mut, replacing four
  open-coded repository lookups.
- Add head_refs and heads_selector so the heads/{branch} mapping is
  spelled once instead of in add_repository, the fixture conversion, and
  the branch handler.
- Key repository files by commit SHA then path rather than by a
  (String, String) tuple, which drops two allocations and two full-map
  scans per content request.

Repository reader:
- Use DisplaySafeUrl, which removes the file-scope disallowed_types
  suppression and the direct url dependency. The suppression covered the
  whole module and everything later added to it.
- Build {api_base}/repos/{owner}/{repo} once when the session opens, so
  the URL builders become infallible methods and three unreachable
  cannot-be-a-base error paths disappear.
- Collapse the per-operation NotFound and Unavailable variants into ones
  carrying the operation, derive its rendering with strum, and mark the
  error non_exhaustive.
- Return the status classification as one Err(match), size the body
  buffer from Content-Length, and lowercase the resolved SHA in place.

Pull request pipeline:
- Open one reader before the branch-head retry loop instead of once per
  attempt. With App credentials each attempt previously minted a fresh
  installation token, costing two extra round trips per retry. Only the
  ref lookup is retried now; credential failures surface immediately.

Tests keep their coverage: one helper opens readers across eight call
sites, and the repository file fixtures become a table.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-19 15:15:43 -04:00
Scott Werner
13bc1f35de Add bounded GitHub repository reads 2026-08-17 12:56:09 -04:00
Bryan Helmkamp
5c6289df80
Simplify async pull request creation
Structural cleanup of the durable pull request creation feature, from a
three-agent review (reuse, quality, efficiency) of the branch:

- Move the supervisor out of handler/ into server/pull_request_supervisor.rs,
  collapse its double bookkeeping into one task-id map, and fold the five
  copy-pasted failure arms into attempt_pull_request_creation.
- Tag pull_request.failed events with the creation id they resolve, so a
  publish-stage failure can never fail an unrelated explicit creation. The
  reducer gains PullRequestCreation::succeed/fail transition methods.
- Scan pending creations through a narrow projection-cache accessor instead
  of materializing every run summary, raise the scan interval to 30s (notify
  covers the live path), and cap retries for runs whose worker cannot even
  record a failure.
- Answer "creation already pending" POSTs before taking the per-run create
  lock, which a worker can hold for the whole creation.
- Replace the hand-rolled per-run lock map with fabro_store::KeyedMutex.
- Reuse cheap Arc'd projections (cached_run_projection) on the poll endpoint
  and in the worker instead of deep-cloning run summaries and diffs.
- Merge ExistingPullRequest into fabro_github::CreatedPullRequest and
  extract one reconcile_existing_pull_request helper for both call sites.
- Give the client poll loop a 15-minute deadline; document that Retry-After
  and the poll interval are the same constant.
- Resolve a wedged pending creation (run already has a pull request) as a
  durable failure instead of skipping it forever.
- Tests: shared wait_for_pull_request_creation helper, a pinned generation-
  failure assertion, and a new pipeline test proving reconciliation adopts
  an existing PR without an LLM call or create request.

Verified: cargo build --workspace, cargo nextest run --workspace (7,767
passed), nightly clippy -D warnings, fmt --check, insta (no pending), bun
typecheck in fabro-api-client.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-04 14:51:29 -04:00
Bryan Helmkamp
30ce02f2a8
Make pull request creation durable and asynchronous 2026-08-01 19:28:52 -04:00
Scott Werner
1575cc174f Stop requesting workflow write permission 2026-07-29 13:19:55 -04:00
Bryan Helmkamp
1aa7a153b0
Merge pull request #652 from fabro-sh/fix/publish-failures
Make publish failures terminal
2026-07-28 17:03:18 -04:00
Bryan Helmkamp
4727ee8e75
Harden sandbox clone retries 2026-07-28 16:26:41 -04:00
Bryan Helmkamp
73f48eeddb
refactor: simplify publish pipeline and collapse duplicated stage errors
Follow-up cleanup on the publish-failures change.

Error model:
- Collapse `Error::{Engine, Publish, Handler}` into one `Error::Stage` with an
  `ErrorStage` discriminator. The three shared a field shape and had to be
  edited together in four match groups; nine near-identical constructors
  become two private helpers.
- Add `Error::failure_reason()`, replacing the same error -> FailureReason
  mapping written out in four places.
- Publish errors are now terminal. Publish runs once, after execution, so no
  caller could ever act on the retryable classification.

Publish phase:
- Fix: a branch that was pushed is now still reported when pull request
  creation fails afterwards. `PublishOutcome` records what happened and
  carries the error separately, instead of hiding both behind a `Result`.
- Drop `PublishOutcome::NoChanges`, which no consumer distinguished from
  `Published { pr_url: None }`.
- Move publish onto `Concluded` as methods and replace three near-identical
  precondition guards with one `publish_target()`.

Pull requests:
- `maybe_open_pull_request` -> `open_pull_request` returning the record
  directly. Both callers already reject empty diffs, so the `Ok(None)` path
  was unreachable.
- Drop `CreatedPullRequest.head_sha`, which echoed back its own input.

GitHub client:
- Delete `branch_exists`, which had no callers and duplicated
  `branch_head_sha`. Give `branch_head_sha` the `_with_client` split every
  sibling has and port the tests to `MockHttpClient`.
- Collapse the copy-pasted credential match in `resolve_clone_credentials`.

Events:
- `PullRequestCreated.head_sha` is `Option<String>` instead of using an empty
  string to mean absent.
- Centralize the run-branch refspec in `lifecycle::git::push_run_branch`, so
  `git.push` reports a branch name from both emitters as documented.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-28 15:09:10 -04:00
Bryan Helmkamp
1c82bd9008
fix(workflow): make publish failures terminal 2026-07-27 11:25:18 -04:00
Scott Werner
47bc772f7b refactor: organize crates into three layers 2026-07-23 17:59:34 -04:00