Repository navigation
Conversation
`fetch_full_manifest` and `fetch_version_manifest` previously ran simd_json on the tokio runtime: every CPU-bound parse blocked the FuturesUnordered task from polling sibling fetches' network IO, so manifest preload effectively serialised parse work even with 64 in-flight requests. Move the parse step to rayon's global CPU thread pool via `rayon::spawn` + `tokio::sync::oneshot`. Tokio is now responsible for fetch only; rayon handles all JSON parsing in parallel across num_cpus threads with work-stealing. No trait surface change. No preload.rs change. No `Send`/`Sync` propagation. Wasm32 falls back to inline parse. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
… CI)
A/B-verified perf fix isolated as its own PR. Detect stderr's TTY
status once at startup (`IS_TTY: Lazy<bool>`) and short-circuit the
entire `ProgressReceiver::on_event` path on non-TTY environments.
## Why this is needed
indicatif's `inc()` / `set_message()` always acquire the internal
`Mutex<ProgressState>` write-lock and mutate `pos` / `est.buf` /
`message`, regardless of TTY status — only the *rendering* short-
circuits, not the state mutation. On the worker-pool resolve hot path
that's 9000+ lock acquisitions per phase contending across multiple
workers, costing measurable wall time on CI.
## A/B/A/A' bench evidence (CI Linux, ant-design / npmjs)
| Round | indicatif config | Lock count | p1_resolve | σ |
|---|---|---|---|---|
| A | drop entirely | 0 | 2.50s | 0.10 |
| B | full restore | 13713 | 2.73s | 0.07 |
| A' | drop again (verify) | 0 | 2.58s | 0.15 |
| A'' | only "resolved" event | 9142 | 2.65s | 0.05 |
| this PR | TTY gate at startup | 0 in CI | 2.58s | 0.07 |
A→B Δ = +0.23s (+9%); B→A Δ = −0.15s. Three independent runs across
different GH Actions runners and Cloudflare egress conditions.
Amortized cost is approximately linear in lock count
(0.23 × 9142/13713 ≈ 0.15 ≈ predicted; observed 0.15).
## Why this approach
- **Drop-only solutions** (A round) lose the local-dev progress UX.
- **Partial drops** (A'' round, only-resolved) still pay 9142 locks.
- **TTY gate** is the only solution that:
- Pays zero overhead on CI / piped (`*IS_TTY = false` → match never enters)
- Preserves full per-package "resolved {name}" UX on local dev (TTY)
- Is a 1-file, 3-call-site change
## Counter-intuitive finding
The first instinct ("CI has no TTY, indicatif must already short-
circuit internally") is wrong: indicatif separates *rendering*
(skipped on non-TTY / hidden draw target) from *state mutation*
(always runs, always locks). Round A'' confirmed this empirically —
9142 locks with no rendering still cost +0.15s on p1_resolve.
General rule: gate hot-path library calls at the call-site; never
assume the library will short-circuit internally.
## Test plan
- [x] `cargo build --release -p utoo-pm` clean
- [x] `cargo clippy -p utoo-pm --all-targets -- -D warnings --no-deps` clean
- [x] CI bench-phases-linux on PR #2818 round-10 (`f122478b`): p1_resolve
2.58s ±0.07 — matches "indicatif fully dropped" baseline (round-6/8)
- [ ] Local dev TTY: `cargo run --profile release-local --bin utoo -- install`
should show full per-package progress
- [ ] Non-TTY: `… 2>/tmp/stderr.log` should produce zero indicatif ANSI
sequences in the log
`resolve_via_full_manifest` now checks the version-manifest cache by the resolved version BEFORE the rayon `extract_version` reparse. Whenever multiple specs (`react@^18`, `react@^18.1`, `react@18.x`) resolve to the same concrete version (`18.2.0`), only the first call pays the parse cost; siblings reuse the cached `Arc<CoreVersionManifest>`. The cache itself was already populated under both `(name, spec)` and `(name, resolved_version)` keys (line 373-378), so the second key becomes the natural short-circuit point. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
`resolve_target_version` takes `&[String]`, so the four call sites that prefixed it with `let version_list: Vec<String> = manifest.versions.clone();` were paying for a pointless deep copy of the entire version list per resolve (50-200 entries on popular packages, ~600 packages on ant-design). Pass `&manifest.versions` / `&full_manifest.versions` directly. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Previously the 200 path of `resolve_full_manifest` deep-cloned `manifest.versions` (Vec<String>, ~50-200 entries on popular packages) and `manifest.dist_tags` (HashMap) per package fetch to fill the `versions_info` cache slot, while also moving the same fields into `Arc<FullManifest>` for the `full_manifests` slot. Use `std::mem::take` to MOVE both fields out of `manifest` directly into the new `VersionsInfo`. The cached `Arc<FullManifest>` keeps `raw` + scalar metadata; `version_list` / `dist_tags` now live only in the `versions_info` cache slot. `resolve_via_full_manifest`'s Full branch reads them via `cache.get_versions(name)` (same code path that filled the slot, atomic w.r.t. the inflight gate). ant-design fetch ≈ 600 packages × ~80 String entries; eliminates one deep copy per fetch. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Manifest preload is currently capped at 64 in-flight requests. With ~600 deps on ant-design completing in ~3.5s, observed throughput is ~170 req/s — far below what 64 in-flight × 10ms RTT would saturate (~6k req/s). Cloudflare per-IP throttle is the suspected ceiling, but GH Actions egress IP space may not hit the throttle. Experiment to measure: does doubling concurrency convert into wall-clock on CI?
Rayon's global pool absorbs simd_json manifest parses, extract_version reparses, and tarball gzip+tar+write. Default `num_cpus` workers leaves work-stealing idle when any single worker stalls on `fs::File::create` inside extract_tarball — siblings can't drain. 2× lets blocked-on-IO workers be backstopped without oversubscribing CPU during pure-CPU bursts (rayon's work-stealing self-tunes). Experiment to measure on top of A (preload=128).
This reverts commit 19ce70e.
H2 multiplexes 100+ streams per TCP connection. Cloudflare's per-IP rate limit can be paired with a per-connection stream-rate cap that silently serialises concurrent fetches even when the client opens parallel streams. H/1.1 spreads concurrency across multiple TCP connections — each request gets its own connection slot, hitting only the per-IP cap (which appears higher in absolute terms). Experiment to measure on top of preload=128.
This reverts commit c41407a.
…nt E) Two changes on the tarball download path: 1. `DOWNLOAD_SEMAPHORE` now reads `DEFAULT_CONCURRENCY` (128, shared with manifest preload) instead of `manifests-concurrency-limit` (64). Same per-IP throttle ceiling applies to manifest and tarball fetches; capping download tighter than resolve underutilises the network budget. 2. `build_dns_cached_client` now requests HTTP/2 with prior knowledge. Tarball CDN previously default-negotiated HTTP/1.1 — 128 concurrent downloads opened 128 separate TCP/TLS connections. H/2 multiplexes onto a few persistent connections, but exposes us to the per-stream rate limit if the CDN has one (mirror of experiment D's hypothesis). Drop the now-unused `get_manifests_concurrency_limit_sync`.
…experiment E)" This reverts commit 036f479.
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request introduces performance optimizations by offloading CPU-intensive JSON parsing to a Rayon thread pool, which prevents the Tokio runtime from being blocked during manifest processing. Additionally, it increases the default concurrency for manifest fetching from 64 to 256 and reduces memory overhead by utilizing std::mem::take and Arc to avoid unnecessary clones of manifest data. The review feedback focuses on ensuring that panics occurring within the spawned Rayon tasks are properly propagated to the caller rather than being swallowed or converted into recoverable errors, adhering to the principle that panics represent unrecoverable bugs.
| rayon::spawn(move || { | ||
| let _ = tx.send(full.get_core_version(&version).map(std::sync::Arc::new)); | ||
| }); | ||
| rx.await.ok().flatten() |
Contributor
There was a problem hiding this comment.
The current implementation with .ok().flatten() silently swallows panics that might occur within the rayon::spawn task (e.g., in full.get_core_version). This can lead to confusing behavior, as a panic during parsing would manifest as a "version not found" error downstream, which is misleading.
According to the repository's general rules, panics should be treated as unrecoverable bugs and not be silently recovered from. To adhere to this, you should propagate the panic to the calling tokio task.
You can achieve this by replacing .ok().flatten() with .expect(...) to ensure that a panic in the spawned task will cause the awaiting task to panic as well.
Suggested change
| rx.await.ok().flatten() | |
| rx.await.expect("rayon task for extract_core_version_off_runtime panicked") |
References
- Do not implement recovery logic for panics. Panics should be treated as unrecoverable bugs that need to be fixed, not as transient, recoverable errors.
Comment on lines
+35
to
+36
| rx.await | ||
| .map_err(|e| anyhow!("rayon parse channel closed: {e}"))? |
Contributor
There was a problem hiding this comment.
If a panic occurs within the rayon::spawn task, the current implementation treats it as a recoverable error by mapping it to an anyhow! result. According to the repository rules, panics should be treated as unrecoverable bugs. To adhere to this, you should propagate the panic using .expect() instead of converting it into a Result error.
Suggested change
| rx.await | |
| .map_err(|e| anyhow!("rayon parse channel closed: {e}"))? | |
| rx.await.expect("rayon task panicked during JSON parsing") |
References
- Do not implement recovery logic for panics. Panics should be treated as unrecoverable bugs that need to be fixed, not as transient, recoverable errors.
elrrrrrrr
added
benchmark
📊 pm-bench-phases ·
|
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 9.15s | 0.31s | 10.16s | 10.09s | 718M | 333.2K |
| utoo-next | 9.48s | 0.15s | 11.30s | 13.02s | 1.26G | 174.7K |
| utoo-npm | 10.53s | 0.87s | 11.45s | 13.11s | 1.16G | 154.5K |
| utoo | 9.82s | 1.77s | 11.32s | 12.83s | 1.38G | 187.1K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 14.3K | 17.9K | 1.16G | 6M | 1.85G | 1.73G | 1M |
| utoo-next | 161.9K | 140.7K | 1.13G | 4M | 1.68G | 1.67G | 2M |
| utoo-npm | 168.1K | 145.8K | 1.14G | 4M | 1.68G | 1.68G | 2M |
| utoo | 138.0K | 83.1K | 1.13G | 3M | 1.68G | 1.68G | 2M |
p1_resolve
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 1.91s | 0.07s | 3.94s | 1.03s | 490M | 167.0K |
| utoo-next | 4.85s | 0.03s | 5.87s | 1.38s | 426M | 70.1K |
| utoo-npm | 5.09s | 0.06s | 5.87s | 1.04s | 431M | 69.1K |
| utoo | 3.58s | 0.14s | 5.62s | 2.13s | 608M | 84.9K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 8.2K | 4.3K | 203M | 3M | 106M | - | 1M |
| utoo-next | 93.8K | 8.8K | 197M | 2M | 7M | 3M | 2M |
| utoo-npm | 65.1K | 2.9K | 202M | 2M | 9M | 5M | 2M |
| utoo | 69.0K | 123.5K | 197M | 1M | 7M | 3M | 2M |
p3_cold_install
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 7.35s | 0.86s | 6.11s | 9.86s | 639M | 207.3K |
| utoo-next | 8.43s | 3.39s | 5.43s | 11.47s | 943M | 117.6K |
| utoo-npm | 7.98s | 1.35s | 5.49s | 11.27s | 909M | 117.8K |
| utoo | 6.11s | 0.05s | 5.27s | 10.91s | 998M | 114.6K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 3.7K | 7.0K | 993M | 3M | 1.73G | 1.73G | 1M |
| utoo-next | 120.5K | 72.4K | 964M | 3M | 1.67G | 1.67G | 2M |
| utoo-npm | 109.9K | 61.9K | 964M | 3M | 1.67G | 1.67G | 2M |
| utoo | 94.2K | 65.8K | 964M | 2M | 1.67G | 1.67G | 2M |
p4_warm_link
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 3.53s | 0.15s | 0.20s | 2.39s | 134M | 30.9K |
| utoo-next | 2.48s | 0.08s | 0.59s | 3.90s | 82M | 19.3K |
| utoo-npm | 2.50s | 0.08s | 0.61s | 3.88s | 83M | 19.1K |
| utoo | 2.38s | 0.07s | 0.64s | 3.92s | 82M | 19.5K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 465 | 21 | 3K | 15K | 1.84G | 1.73G | 1M |
| utoo-next | 50.1K | 21.4K | 21K | 12K | 1.67G | 1.67G | 2M |
| utoo-npm | 50.0K | 20.6K | 21K | 28K | 1.67G | 1.67G | 2M |
| utoo | 49.8K | 22.6K | 21K | 28K | 1.68G | 1.67G | 2M |
npmmirror.com
p0_full_cold
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 25.97s | 6.72s | 8.98s | 9.52s | 532M | 426.9K |
| utoo-next | 20.17s | 2.67s | 7.95s | 13.85s | 868M | 116.9K |
| utoo-npm | 18.19s | 0.69s | 8.22s | 13.93s | 1012M | 117.8K |
| utoo | 22.26s | 4.31s | 8.01s | 14.56s | 901M | 125.8K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 52.8K | 4.2K | 1.12G | 9M | 1.84G | 1.72G | 2M |
| utoo-next | 227.9K | 116.3K | 982M | 8M | 1.67G | 1.68G | 2M |
| utoo-npm | 223.0K | 118.1K | 980M | 7M | 1.67G | 1.68G | 2M |
| utoo | 228.7K | 169.2K | 999M | 8M | 1.67G | 1.68G | 2M |
p1_resolve
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 2.83s | 0.40s | 3.84s | 1.25s | 531M | 197.5K |
| utoo-next | 4.97s | 0.13s | 1.98s | 0.64s | 82M | 19.8K |
| utoo-npm | 4.95s | 0.29s | 1.96s | 0.58s | 74M | 16.3K |
| utoo | 11.73s | 11.81s | 2.07s | 1.33s | 83M | 19.8K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 16.4K | 4.0K | 151M | 3M | 106M | - | 2M |
| utoo-next | 47.0K | 734 | 14M | 2M | - | 3M | 2M |
| utoo-npm | 46.4K | 702 | 14M | 2M | - | 4M | 2M |
| utoo | 58.9K | 67.6K | 13M | 2M | - | 3M | 2M |
p3_cold_install
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 16.87s | 0.31s | 5.68s | 8.88s | 296M | 153.4K |
| utoo-next | 20.58s | 0.78s | 5.99s | 12.40s | 712M | 92.4K |
| utoo-npm | 20.16s | 0.28s | 5.98s | 12.50s | 551M | 93.5K |
| utoo | 20.53s | 0.38s | 6.09s | 12.56s | 585M | 87.6K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 22.8K | 3.1K | 998M | 6M | 1.73G | 1.73G | 2M |
| utoo-next | 174.3K | 105.7K | 966M | 5M | 1.67G | 1.67G | 2M |
| utoo-npm | 177.4K | 105.8K | 967M | 6M | 1.67G | 1.67G | 2M |
| utoo | 179.7K | 104.2K | 967M | 5M | 1.67G | 1.67G | 2M |
p4_warm_link
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 3.46s | 0.04s | 0.20s | 2.36s | 136M | 31.7K |
| utoo-next | 2.51s | 0.26s | 0.56s | 3.89s | 82M | 18.7K |
| utoo-npm | 2.42s | 0.08s | 0.56s | 3.85s | 83M | 18.6K |
| utoo | 2.42s | 0.11s | 0.57s | 3.80s | 82M | 18.9K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 526 | 22 | 7M | 26K | 1.88G | 1.72G | 2M |
| utoo-next | 49.7K | 21.6K | 22K | 39K | 1.67G | 1.67G | 2M |
| utoo-npm | 50.3K | 20.5K | 21K | 10K | 1.67G | 1.67G | 2M |
| utoo | 47.8K | 19.5K | 23K | 10K | 1.67G | 1.67G | 2M |
📊 pm-bench-phases ·
|
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 23.29s | 1.08s | 7.15s | 23.84s | 783M | 50.5K |
| utoo-npm | 20.64s | 4.51s | 9.87s | 21.59s | 1.12G | 107.4K |
| utoo | 18.34s | 3.26s | 8.49s | 18.96s | 1002M | 98.7K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 15.6K | 140.6K | - | - | 1.79G | 1.90G | 1M |
| utoo-npm | 13.6K | 362.5K | - | - | 1.63G | 1.87G | 2M |
| utoo | 13.5K | 328.1K | - | - | 1.63G | 1.84G | 2M |
p1_resolve
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 2.52s | 0.07s | 2.77s | 1.32s | 491M | 31.8K |
| utoo-npm | 5.30s | 0.30s | 4.23s | 2.21s | 545M | 36.9K |
| utoo | 4.01s | 0.06s | 4.41s | 2.61s | 603M | 40.3K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 32 | 19.8K | - | - | 110M | - | 1M |
| utoo-npm | 12 | 72.4K | - | - | 28M | 5M | 2M |
| utoo | 28 | 154.9K | - | - | 27M | 3M | 2M |
p3_cold_install
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 20.82s | 7.13s | 4.03s | 21.23s | 533M | 34.7K |
| utoo-npm | 18.86s | 2.98s | 4.85s | 22.53s | 809M | 80.8K |
| utoo | 17.85s | 0.82s | 5.03s | 23.33s | 677M | 76.3K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 4.3K | 136.1K | - | - | 1.70G | 1.94G | 1M |
| utoo-npm | 1.7K | 241.5K | - | - | 1.61G | 1.83G | 2M |
| utoo | 1.3K | 238.2K | - | - | 1.61G | 1.83G | 2M |
p4_warm_link
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 8.99s | 0.33s | 0.16s | 3.50s | 51M | 3.9K |
| utoo-npm | 7.68s | 0.88s | 0.94s | 5.17s | 93M | 6.9K |
| utoo | 6.88s | 1.34s | 0.79s | 4.59s | 97M | 7.2K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 19.9K | 1.3K | - | - | 1.86G | 1.91G | 1M |
| utoo-npm | 13.3K | 82.7K | - | - | 1.61G | 1.83G | 2M |
| utoo | 14.0K | 76.4K | - | - | 1.63G | 1.83G | 2M |
npmmirror.com
p0_full_cold
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 31.93s | 5.75s | 7.86s | 26.02s | 577M | 37.3K |
| utoo-npm | 26.64s | 1.55s | 8.29s | 25.69s | 802M | 80.1K |
| utoo | 30.96s | 4.31s | 8.74s | 27.78s | 757M | 77.5K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 15.5K | 149.4K | - | - | 1.76G | 1.91G | 2M |
| utoo-npm | 1.0K | 338.5K | - | - | 1.61G | 1.87G | 2M |
| utoo | 4.9K | 386.6K | - | - | 1.61G | 1.88G | 2M |
p1_resolve
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 2.86s | 0.11s | 3.12s | 1.80s | 530M | 34.5K |
| utoo-npm | 5.83s | 0.31s | 1.95s | 1.05s | 78M | 5.7K |
| utoo | 11.57s | 12.26s | 2.26s | 1.50s | 94M | 6.7K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 15 | 21.8K | - | - | 112M | - | 2M |
| utoo-npm | 6 | 36.6K | - | - | - | 4M | 2M |
| utoo | 26 | 190.5K | - | - | - | 3M | 2M |
p3_cold_install
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 53.56s | 34.43s | 4.00s | 17.76s | 235M | 15.6K |
| utoo-npm | 76.36s | 88.80s | 5.56s | 20.30s | 720M | 76.6K |
| utoo | 39.88s | 23.79s | 5.97s | 21.37s | 693M | 74.0K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 1.8K | 109.0K | - | - | 1.65G | 1.91G | 2M |
| utoo-npm | 2.0K | 382.8K | - | - | 1.60G | 1.87G | 2M |
| utoo | 1.6K | 373.5K | - | - | 1.60G | 1.87G | 2M |
p4_warm_link
| PM | wall | ±σ | user | sys | RSS | pgMinor |
|---|---|---|---|---|---|---|
| bun | 4.82s | 0.53s | 0.09s | 2.11s | 44M | 3.4K |
| utoo-npm | 5.36s | 1.39s | 0.83s | 4.23s | 84M | 6.3K |
| utoo | 6.67s | 0.57s | 0.89s | 4.99s | 99M | 7.4K |
| PM | vCtx | iCtx | netRX | netTX | cache | node_mod | lock |
|---|---|---|---|---|---|---|---|
| bun | 13.6K | 638 | - | - | 1.78G | 1.90G | 2M |
| utoo-npm | 12.6K | 82.6K | - | - | 1.60G | 1.86G | 2M |
| utoo | 12.5K | 83.3K | - | - | 1.60G | 1.86G | 2M |
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.
Reproduction bench for f5006bb pre-rebase state (PR #2874 best-perf snapshot).
State:
Goal: confirm whether the 3.77s p1_resolve / 6.60s p3_cold_install observed at f5006bb is reproducible, to disambiguate CI variance from regression introduced by post-rebase changes.
DO NOT MERGE.