🤖 fix: build dist/ before running terminal-bench #512
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| archive = build_app_archive(self._repo_root, self._INCLUDE_PATHS) | ||
| self._archive_bytes = archive | ||
| return archive | ||
| self._staged_container_id = container_id |
There was a problem hiding this comment.
The new _ensure_payload_staged unconditionally caches self._staged_container_id = container_id and returns early when container_id == self._staged_container_id. When session.container has no id attribute (the code already handles this case with getattr(..., None)), container_id becomes None. After the first call the cached ID is also None, so every subsequent call returns immediately without re-running stage_payload. For sessions whose container cannot be identified, this skips installing the cmux bundle for new benchmark containers and tasks will fail because /installed-agent was never repopulated. The previous version only cached when a real container ID was available specifically to avoid this problem.
Useful? React with 👍 / 👎.
PR #507 added `dist/` to the terminal-bench archive include paths to fix worker crashes. However, the workflow wasn't building `dist/` before running the benchmark, causing all tasks to fail immediately with: ``` Error running agent for task <name>: Required file /home/runner/work/cmux/cmux/dist missing ``` Now runs `make build` before `make benchmark-terminal` to ensure dist/ exists and contains the compiled worker files. Verified with workflow run #19140594821 which successfully completed the modernize-fortran-build task.
ammar-agent
force-pushed
the
tb-timeout
branch
from
853a074 to
8542a87
Compare
|
Closing to create fresh PR without stale review comments |
1 participant
Problem
PR #507 added
dist/to the terminal-bench archive include paths to fix worker crashes. However, the workflow wasn't buildingdist/before running the benchmark, causing all tasks to fail immediately with:Solution
Add
make buildstep beforemake benchmark-terminalin the workflow. This ensures:dist/directory existsTesting
Manually triggered workflow run #19108392805 which failed with this exact error. With this fix, the workflow should build dist/ successfully and allow terminal-bench tasks to run.
Generated with
cmux