fix(tests): key the corpus-column scratch dir per call, not per process - #2969
Closed
gHashTag wants to merge 2 commits into
Closed
fix(tests): key the corpus-column scratch dir per call, not per process#2969gHashTag wants to merge 2 commits into
gHashTag wants to merge 2 commits into
Conversation
Closes #2968 `tri harness scratch --gate` has been red on master since `994f3c8a3`. `bootstrap/tests/corpus_zig_bodies.rs:86` keys its scratch directory by `std::process::id()` -- one value for the whole test binary -- and the test that uses it ends with `remove_dir_all(&dir)`. Measured, and stated as measured: only one of that file's three tests touches the directory today, so nothing collides on this tree. The gate is structural and right to be: the second test to use that key would collide, and it would pass the first time it ran. Fixed as the gate's own message prescribes -- an `AtomicUsize` counter, unique per call. `tri harness scratch --gate` prints `none` and exits 0; the three tests in that file still pass. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Contributor
PR DashboardGenerated at: 2026-08-30 18:09:30 UTC
Summary
Seal Status
|
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Refs #2968 `git add -A` in this branch's first commit swept in 44 lines of `bootstrap/src/compiler.rs` and its `FROZEN_HASH` that I did not write: a C and Rust emitter fix for string literals (`extra_kind == "string"`), belonging to another session. How it got into a clean worktree: the stash stack is **shared across every worktree of a repository**. An earlier command of mine ran `git stash -q -u` on a tree with nothing to stash -- which pushes nothing -- and then `git stash pop -q`, which popped the top of the shared stack: somebody else's work in progress. Both files are restored to master's version, so this branch is again the one line it claims to be. Their work is not lost: it is the diff of this branch's first commit `1c00afecb`, and a copy is attached to #2968. Measured before and after: with a release build of this branch, `tri seals drift` read 152 drifted seals; on master it reads 0. After this revert it reads 0 -- the drift was the swept-in emitter change, not my test file. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Contributor
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
Owner
Author
|
Superseded by #2973, which landed the same repair (an
The one thing here worth keeping is on #2968 and not in this diff: the 44 lines of |
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.
Closes #2968
tri harness scratch --gatehas been red on master since994f3c8a3.bootstrap/tests/corpus_zig_bodies.rs:86keyed its scratch directory bystd::process::id()— one value for the whole test binary — and the test thatuses it ends with
remove_dir_all(&dir).Measured, and stated as measured: only one of that file's three tests creates
or deletes the directory today, so nothing collides on this tree. The gate is
structural and right to be — the second test to use that key would collide, and
it would pass the first time it ran.
Fixed as the gate's own message prescribes: an
AtomicUsizecounter, unique percall.
Verify
none, exit 0.cargo test --manifest-path bootstrap/Cargo.toml --test corpus_zig_bodies→ 3 passed.Not mine to have written, and one line to repair — the gate caught it on new code
within an hour of being wired, which is the reading worth keeping.
🤖 Generated with Claude Code