Conversation
…ist and stay pinned closure_records.py reads every spec/compatibility/closure/*.json: each cited repository path must exist, a #fragment must name an entry (a heading anchor in Markdown), a <name>Path / <name>Sha256 pair must match, and cited records under spec/ and conformance/ are pinned in record-digests.json so they cannot change without an explicit --write in the same commit. A closure may not cite a suite that CI no longer runs (RETIRED_SUITES in the compatibility-inventory workflow).
…on doc The checked-in matrix cites file:line locations that moved with later code changes (rest/mod.rs and others), so test_fs_config_surface_matrix and test_fs_config_docs failed. Regenerated with surface_matrix.py --write and doc_render.py --write; only the cited line numbers change.
…ee directory test_historical_worktree_parent_accepts_the_real_repository_parent required the developer checkout's .worktree directory and failed on every CI runner. Build a real worktree parent under tmp_path instead; the refusal test for a symlinked parent is unchanged.
…shard by duration Owner decision 2026-09-25. The retired test files are listed once in the workflow's RETIRED_SUITES with the reasons: each fails because 866bb06 moved the shared request_bytes_preflight.py to the sandbox project (passes at 866bb06^, fails at 866bb06), because an O7/O8 shadow or runtime binding predates later source changes, because it fails only on the CI runner, or because it alone exceeds the 45-minute job limit. They run on demand through workflow_dispatch (run_retired) with the O7-era history.py frozen-bytes check; no test code or expectation changed. compat-broad shards are planned from measured per-file durations (7 shards, about 27 minutes for the fullest). A new required job checks the records every closure cites and runs the closure contracts.
Artifact-backed tests refuse a dirty working tree, and the list written into the checkout made it dirty.
fs-rules-publication/test_o5_user_token_comparator_v2.py builds a fully bound collection per test (about 2 minutes each, 178 tests), so no file-level plan fits the 45-minute limit. broad_shards.py now runs a file longer than 0.8 of the budget as contiguous chunks of whole tests, rebases the collected IDs on the repository path, and plans 22 shards (about 31 minutes for the fullest). Also retire fs-write-limits/test_limits_03_baseline_prep.py: limits-03 takes its project from the shared request-byte preflight, which now names the sandbox, while baseline preparation requires fireemu-35fe6.
Owner
Author
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.
Temporary draft PR to run the broad-timing workflow on the CI runner (per-file durations for duration-based compat-broad shards). It will be closed without merging; the retirement changes land on feat separately after review.