Testing ci with feature#3
Conversation
re create functionality of criterion compare action
|
/bench --tags listing-optimization |
|
PR title does not match the required pattern. Please ensure you follow the conventional commits spec. Your title should start with Title: Your title should start with Title: Your title should start with Title: Your title should start with Title: |
add list with backward checkpoint scan added tests, some restructuring fixes fix restored list() tests for reviewer readability quick fix quick fix fix final tests test fixes fix
git checkout -- <name> treats the argument as a pathspec, not a branch name, causing "pathspec did not match" errors. Using FETCH_HEAD avoids passing any user-controlled input to the checkout command entirely.
20f0d5c to
b55bee6
Compare
|
/bench --tags listing-optimization |
The post-comment job runs on a fresh runner with no checkout, so gh cannot auto-detect the repository from git. Pass --repo explicitly.
|
/bench --tags listing-optimization |
Benchmark for c651b7dClick to view benchmark
|
|
/bench |
1 similar comment
|
/bench |
|
/bench --filter snapshotConstruction |
|
/bench --tags base --filter 101kAdds.*snapshotConstruction |
|
/bench |
|
/bench --filter snapshotConstruction |
|
/bench --tags base --filter 101kAdds.*snapshotConstruction |
|
/bench --filter 101kAdds|10kAdds |
|
/bench --tags listing-optimization,metadata-only |
|
/bench --tags metadata-only, listing-optimization |
Benchmark for 7a7215atags: `base`
|
Benchmark for 7a7215atags: `listing-optimization,metadata-only`
|
|
/bench --filter 101kAdds|10kAdds |
|
/bench --tags metadata-only, listing-optimization |
|
/bench --filter snapshotConstruction |
|
/bench --filter snapshot |
|
/bench --filter 101kAdds|10kAdds |
|
/bench --tags listing-optimization, base |
|
/bench --tags base --filter 101kAdds.*snapshotConstruction |
Benchmark for 7a7215atags: `listing-optimization,`
|
Benchmark for 7a7215atags: `base` | filter: `101kAdds.*snapshotConstruction`
|
Benchmark for 7a7215afilter: `101kAdds|10kAdds`
|
Benchmark for 7a7215afilter: `snapshot`
|
|
/bench --tags crc-optimization --filter SinceCrc.*snapshotConstruction |
Benchmark for 47d1fb2tags: `crc-optimization` | filter: `SinceCrc.*snapshotConstruction`
|
|
/bench --tags listing-optimization, base |
|
/bench --tags listing-optimization |
Benchmark for 47d1fb2tags: `listing-optimization`
|
…action (#2264) ## What changes are proposed in this pull request? This PR introduces the following functionality: - Replaces the boa-dev/criterion-compare-action GitHub Action with a custom bash script (.github/scripts/run-benchmarks.sh) that replicates its output format, including a markdown table with base and PR durations, a percentage change column, and bolding of the faster result when the difference is statistically significant based on Criterion's error bounds. - Splits the benchmark workflow into two jobs — run-benchmark (which checks out and runs untrusted PR code with only contents: read permission) and post-comment (which holds pull-requests: write and posts results) — so that GH_TOKEN is never in the environment when untrusted code executes. - Extends the /bench comment syntax to support both a --tags flag (comma-separated list passed as BENCH_TAGS) and a --filter flag (Criterion benchmark name regex), where bare /bench defaults to --tags base. - Pins all GitHub Actions to full commit SHAs and passes all user-controlled values through environment variables rather than inline expressions to prevent expression injection. ## How was this change tested? Made a pull request on my fork and checked that it commented correctly and that --tags and --filter work: lorenarosati#3
What changes are proposed in this pull request?
How was this change tested?