[jdx]

🤖 Refreshed benchmarks

benchmarks/results.json was pinned to aube 1.21.0; the workspace is now 1.22.0. Re-ran mise run bench:bump on the hermetic Verdaccio registry (500mbit / 50ms per the mise task) and regenerated benchmarks/results.json plus the README BENCH_RATIOS block. The benchmark matrix pins aube's GVS mode via npm_config_enable_global_virtual_store=true|false (the auto-synthesized env alias for the enableGlobalVirtualStore setting), so GitHub Actions' inherited CI=true environment does not change whether aube runs with GVS enabled or disabled.

Benchmark changes

Versions:

• aube: 1.21.0 -> 1.22.0
• pnpm: 11.6.0 -> 11.7.0
• yarn: 4.16.0 -> 4.17.0

Public ratios: warm installs vs Bun 2x -> 1.8x; warm installs vs pnpm 3x -> 3x; repeat test vs Bun 5x -> 8x; repeat test vs pnpm 30x -> 42x.

Benchmark	aube	bun	deno	pnpm	npm	yarn
Fresh install (warm cache)	886ms -> 787ms (-11%)	1813ms -> 1412ms (-22%)	1616ms -> 1488ms (-8%)	2512ms -> 2074ms (-17%)	7136ms -> 7089ms (-1%)	7708ms -> 10121ms (+31%)
Fresh install (cold cache)	5860ms -> 6821ms (+16%)	4856ms -> 5429ms (+12%)	6748ms -> 7964ms (+18%)	12925ms -> 11206ms (-13%)	8758ms -> 9054ms (+3%)	11079ms -> 12186ms (+10%)
npm install && npm run test	11ms -> 8ms (-27%)	60ms -> 64ms (+7%)	59ms -> 49ms (-17%)	331ms -> 333ms (+1%)	681ms -> 679ms (0%)	1158ms -> 1496ms (+29%)

Review the numbers before merging — if anything looks wildly off vs. the previous release, investigate before landing. Hermetic proxy jitter or an npmjs uplink hiccup can occasionally skew results.

Once merged to main, the updated bench results flow into the next release-plz-pr run automatically.

---

Generated by the bench-refresh workflow.

---

Note

Low Risk
Documentation and benchmark artifact updates only; no runtime or package-manager logic changes.

Overview
Re-runs the hermetic benchmark suite (mise run bench:bump) and pins benchmarks/results.json to the current workspace release (aube 1.22.0), along with bumped comparator versions (pnpm 11.7.0, yarn 4.17.0). Timings and mean/stddev stats are replaced for warm/cold fresh installs and the repeat install && test scenario.

The README.md BENCH_RATIOS block is regenerated from the new JSON: warm vs Bun is now ~1.8x (was ~2x), repeat test vs pnpm up to 42x (was 30x) and vs Bun up to 8x (was 5x); warm vs pnpm stays ~3x.

^{Reviewed by Cursor Bugbot for commit 3c037b0. Bugbot is set up for automated code reviews on this repo. Configure here.}

Summary by CodeRabbit

• Documentation

  • Updated "Fast installs" benchmark claims with refreshed performance measurements.

• Chores

  • Refreshed benchmark results with latest toolchain versions and performance metrics.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Central YAML (base), Organization UI (inherited)

Review profile: CHILL

Plan: Pro Plus

Run ID: 988f4945-283c-4914-aeaf-c4ee55f286e7

📥 Commits

Reviewing files that changed from the base of the PR and between 2bcf93a and 3c037b0.

📒 Files selected for processing (2)

• README.md
• benchmarks/results.json

---

📝 Walkthrough

Walkthrough

Benchmark measurements in benchmarks/results.json were re-run: the updated timestamp, toolchain version strings, and all per-manager values/stats for the gvs-warm, gvs-cold, and install-test rows were replaced with new numbers. The README "Fast installs" speedup sentence was updated to reflect the new results.

Changes

Benchmark Refresh

Layer / File(s)	Summary
Benchmark data refresh benchmarks/results.json	Top-level updated timestamp and versions object refreshed; all rows entries (gvs-warm, gvs-cold, install-test) have new per-manager values arrays and stats (mean/stddev/min/max) replaced with new measurements.
README speedup claims README.md	"Fast installs" sentence updated with revised warm-install multiplier vs Bun and increased repeat-command multipliers vs pnpm and Bun.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Possibly related PRs

• jdx/aube#820: Updates the same README.md benchmark ratios and benchmarks/results.json measurements in a prior benchmark refresh round.
• jdx/aube#890: Refreshes the same benchmark artifacts — README.md speedup claims and benchmarks/results.json values/stats for all three benchmark keys.
• jdx/aube#866: Directly overlapping changes: same README.md ratio update and benchmarks/results.json regeneration with new timestamps, versions, and measurement data.

Poem

🐇 Hop hop, the numbers flew,
Fresh benchmarks baked and hot and new,
The JSON ticked its timestamp right,
And README shone with speedup light,
Faster than Bun on a warm spring day!

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'chore: refresh benchmarks for v1.22.0' accurately describes the main change: refreshing benchmark data and README content for a new version release.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[greptile-apps]

Greptile Summary

This PR refreshes the benchmark data for the aube 1.22.0 release, bumping pnpm to 11.7.0 and yarn to 4.17.0, and regenerates the README BENCH_RATIOS blurb from the new results.

• The repeat-test speedup claims (42x vs pnpm, 8x vs Bun) and the ~1.8x Bun warm-install claim all check out against the JSON means.
• aube's warm-cache vs pnpm ratio slipped from 2.84x to 2.63x; the README still says "about 3x," which was plausible before but is now a stretch.
• aube's cold-cache install regressed ~16% (5.86s → 6.82s) with tight stddev, suggesting the result is stable rather than noise — worth a sanity check before landing.

Confidence Score: 4/5

Safe to merge after confirming the aube cold-cache regression is expected for 1.22.0 and the pnpm warm-install ratio wording is acceptable.

Both findings are non-blocking: one is a wording accuracy concern in the README (2.63x described as "about 3x"), the other is a +16% cold-cache regression that the PR description itself flags as needing review but that appears stable rather than jitter.

benchmarks/results.json — the cold-cache row for aube and the warm-cache row for yarn both show notable regressions that deserve a second look before merging.

Important Files Changed

Filename	Overview
README.md	BENCH_RATIOS block updated; the ~1.8x Bun, 42x pnpm, and 8x Bun repeat-test claims are accurate, but "about 3x faster than pnpm" (warm) overstates the actual 2.63x ratio.
benchmarks/results.json	Versions bumped (aube 1.22.0, pnpm 11.7.0, yarn 4.17.0) and all timing/stats data refreshed; values are consistent with the rounded means, but aube shows a notable +16% cold-cache regression and yarn warm-install regressed +31%.

_{Reviews (1): Last reviewed commit: "chore: refresh benchmarks" | Re-trigger Greptile}

[greptile-apps]

Warm-install pnpm ratio is now 2.63x, not ~3x

The updated JSON gives aube=787ms vs pnpm=2074ms, a ratio of 2074/787 ≈ 2.63x. The previous run was 2512/886 ≈ 2.84x, which made rounding to "about 3x" defensible. At 2.63x the claim has slipped noticeably; "about 2.6x" or "nearly 3x" would be more accurate and harder to dispute in benchmarks discussions.

Note: If this suggestion doesn't match your team's coding style, reply to this and let me know. I'll remember it for next time!

[greptile-apps]

aube cold-cache regression worth investigating before merge

The PR description explicitly asks reviewers to flag anything that "looks wildly off." aube's cold-cache install went from 5860ms → 6821ms (+16%), while every other manager either improved (pnpm: -13%) or stayed flat. This is the only scenario where aube now trails Bun (6821ms vs 5429ms). Worth confirming this is expected for 1.22.0 rather than hermetic-proxy jitter — given the tight stddev (±0.236s) the result appears stable, which makes the regression more likely real.