Repository navigation
Fix macOS test failures on main - #1281
Merged
Merged
Conversation
eval_arms compared recorded binary paths against an uncanonicalized temp dir, which macOS reaches through /var -> /private/var. The Console GitHub tests snapshotted fake gh calls straight after killing the Console, so a gh child it had already spawned could record afterwards. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01M5VdjUGU154umHfxS3ZTBN
JacobStephens2
added
the
test:macos
On macOS, spawning a file that has execute permission but no valid executable format falls back to /bin/sh, which exits 127. Linux refuses the spawn instead. A Harness that exits 126 or 127 before any turn never ran, so its Trial is void (session_unavailable) on both platforms, not failed against its Arm. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LSAzXiA6BD9neiPgVQRMKD
tool_bin_dirs_on matches tools by canonical path, and macOS temp dirs sit behind /var -> /private/var. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LSAzXiA6BD9neiPgVQRMKD
This was referenced Oct 25, 2026
Merged
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes the macOS failures on main since #1122 (CI runs 40131162566, 40141759916, 40145328626). Linux was green throughout.
macOS shards stop at their first failure, so each fix uncovered the next. The full macOS suite (
test:macos) now passes on all four shards.tests/eval_arms.rs: the fixture now canonicalizes its temp root, sinceprepare-armsrecords canonical binary paths and macOS temp dirs sit behind/var -> /private/var.tests/console/github.rs: after killing the Console, wait for fakeghcalls to settle before snapshotting. Agh api graphqlchild spawned just before the kill was recording one extra call.src/harness/interpretation.rs,src/trial.rs). macOS runs a file it can't execute through/bin/sh, which exits 127. Linux refuses the spawn instead. A Harness that exits 126 or 127 before any turn never ran, so its Trial is nowvoid(session_unavailable) on both platforms, rather thanfailedagainst its Arm. The new testa_harness_the_shell_cannot_run_is_a_void_trialreproduces this on Linux.src/sandbox.rs:tool_bin_dirs_names_only_bound_rootscanonicalizes its temp root, since tools are matched by canonical path.🤖 Generated with Claude Code
https://claude.ai/code/session_01M5VdjUGU154umHfxS3ZTBN
https://claude.ai/code/session_01LSAzXiA6BD9neiPgVQRMKD