mirror/pnpm
1
0
Fork 0
mirror of https://github.com/pnpm/pnpm.git synced 2026-07-19 20:22:21 -04:00
Code Issues Packages Projects Releases Wiki Activity

ci(benchmark): drop --show-output (#325)

Browse Source 
Hyperfine's --show-output piped each install's stdout/stderr (with
TRACE=pacquet=info enabled) into the workflow log, drowning the PR
summary in noise. Suppress per-run output and remove the now-pointless
TRACE env var; the markdown report still captures the timing summary.

Co-authored-by: Claude <noreply@anthropic.com>
This commit is contained in:
Khải
2026-04-27 20:29:19 +07:00
committed by GitHub
parent 51571b78c5
commit 840bd8685e
1 changed files with 3 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
pacquet/.github/workflows/integrated-benchmark.yml vendored
View File

@@ -128,35 +128,22 @@ jobs:
# timeout (6 h). A normal run is ~2 min — 10 min is generous
# headroom, and a failure surfaces fast enough to iterate on.
timeout-minutes: 10
# Enable pacquet's tracing at `info` so the next hang shows
# which phase the stalled install was in (download,
# extraction, import, index write …) before it went silent.
# Without this, `--show-output` only captures pacquet's
# empty stream since the default filter is off.
env:
TRACE: pacquet=info
# `--show-output` pipes each binary's stderr through hyperfine
# so the step log captures traces / profile output / panics
# inline with each run. Essential for diagnosing a hang on CI
# where we can't attach a debugger.
run: |
just integrated-benchmark --show-output --scenario=frozen-lockfile --verdaccio --with-pnpm HEAD main
just integrated-benchmark --scenario=frozen-lockfile --verdaccio --with-pnpm HEAD main
cp bench-work-env/BENCHMARK_REPORT.md bench-work-env/BENCHMARK_REPORT_FROZEN_LOCKFILE.md
cp bench-work-env/BENCHMARK_REPORT.json bench-work-env/BENCHMARK_REPORT_FROZEN_LOCKFILE.json
- name: 'Benchmark: Frozen Lockfile (Hot Cache)'
shell: bash
# Same timeout / tracing rationale as the cold-cache step above.
# Same timeout rationale as the cold-cache step above.
# Mirrors pnpm's "Headless install (frozen lockfile, warm
# store+cache)" benchmark: each timed run wipes only
# `node_modules`, the per-revision store survives, and
# hyperfine's warmup run is what populates it on the first
# iteration so timed runs all see a hot store.
timeout-minutes: 10
env:
TRACE: pacquet=info
run: |
just integrated-benchmark --show-output --scenario=frozen-lockfile-hot-cache --verdaccio --with-pnpm HEAD main
just integrated-benchmark --scenario=frozen-lockfile-hot-cache --verdaccio --with-pnpm HEAD main
cp bench-work-env/BENCHMARK_REPORT.md bench-work-env/BENCHMARK_REPORT_FROZEN_LOCKFILE_HOT_CACHE.md
cp bench-work-env/BENCHMARK_REPORT.json bench-work-env/BENCHMARK_REPORT_FROZEN_LOCKFILE_HOT_CACHE.json