Feature/bencher accuracy tracking - #223
Closed
jenskeiner wants to merge 35 commits into
Closed
Conversation
…he Bencher upload - emit per-case NDJSON during build-linux's make check (serial -> *.ndjson, OpenMP -> *.ndjson.threads with openmp:1); no separate test re-run - add runtime (serial/omp) axis to the converter metric name - new environment-gated bencher-upload job (PRs gated, develop auto); project nfft - remove redundant bench-accuracy-linux.yml; update ADR-0004, docs, CONTEXT, plan
Route the upload job to a branch-restricted 'bencher-baseline' environment on push (auto baseline) and the reviewer-gated 'benchmarks' on PR/dispatch, so the API key lives only as an environment secret in those two environments rather than as a repo-wide secret.
Route only fork pull_requests through the reviewer-gated 'benchmarks' environment; push, same-repo PRs, and workflow_dispatch go unattended through 'bencher-baseline'. CodSpeed (bench-linux.yml) is untouched.
jenskeiner
temporarily deployed
to
bencher-baseline
June 29, 2026 19:08 — with
GitHub Actions
Inactive
bencher run's deprecated --token reads the BENCHER_API_TOKEN env var and validates it as a JWT. Setting that var made --token reject the (non-JWT) project-scoped bencher_run_* key at arg-parse time, failing the upload even though we pass --key. Expose the secret as BENCHER_API_KEY instead; the GitHub secret keeps the name BENCHER_API_TOKEN (no rename needed). Verified with a real upload to project nfft.
jenskeiner
temporarily deployed
to
bencher-baseline
June 30, 2026 11:43 — with
GitHub Actions
Inactive
Accuracy reportNo |
Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>
Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>
Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>
… in a list' Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>
The shopt -s nullglob added earlier made `ls base-bmf/*.bmf.json` expand to no args -> `ls` listed the cwd and falsely succeeded, so the script always took the with-baseline branch and diffed the PR against an EMPTY baseline -> a misleading 'Accuracy: 0 cases unchanged' with no heatmap link. Use an array-length check (empty array under nullglob) instead, in both the PR and fork scripts.
|
|
||
| import pytest | ||
|
|
||
| from diff import DiffResult, diff, digits, load_bmf_tree |
jenskeiner
temporarily deployed
to
bencher-baseline
June 30, 2026 12:08 — with
GitHub Actions
Inactive
Contributor
Author
|
Closing in favor of #224. |
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.
No description provided.