feat(watcher): port watchpack safeTime/mtime-accuracy to native watcher

[stormslowly]

Why

The native watcher returned empty fileTimeInfoEntries / contextTimeInfoEntries (left as a // TODO), so compiler.fileTimestamps / contextTimestamps were empty. Persistent-cache snapshot validation then degrades to "trust the changedFiles set only", which can miss or over-invalidate across processes — the JS (watchpack) path does populate these via collectTimeInfoEntries, so this closes a parity gap.

What

Port watchpack's safeTime / mtime-accuracy mechanism into rspack_watcher:

• FS_ACCURACY global (2000ms, monotonically ratchets down) + ensureFsAccuracy resolution detection
• per-entry { safe_time, timestamp, accuracy } with the initial-scan regime (min(now, mtime) + FS_ACCURACY) and the real-time regime (now, accuracy 0), including the attribute-only within-window skip
• fixupEntryAccuracy lazy tightening on read + directory safe_time aggregated from descendant files
• expose via napi getTimeInfo(); NativeWatchFileSystem now fills both maps and getInfo()

accuracy is kept internal — rspack's FileSystemInfoEntry exposes only { safeTime, timestamp }, matching webpack's public contract.

Before / After

// Before
callback(err, new Map(), new Map(), changed, removed); // TODO

// After
const timeInfo = nativeWatcher.getTimeInfo();
callback(err, toTimeInfoMap(timeInfo.fileTimeInfoEntries),
         toTimeInfoMap(timeInfo.contextTimeInfoEntries), changed, removed);

Test

• cargo test -p rspack_watcher — ported watchpack safeTime characterization tests (FS_ACCURACY ladder, fixup, initial/event regimes, within-window skip, dir aggregation)
• watchpack DirectoryWatcherExtra suite green (reference parity)
• NativeWatcher.part1/2/3 JS suites green (230 cases)

Draft: opened to produce a canary for rstest native-watcher stress testing.

[codspeed-hq]

Merging this PR will degrade performance by 2.4%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

❌ 1 regressed benchmark
✅ 51 untouched benchmarks
⏩ 40 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	js@collect imported identifiers	225 µs	230.5 µs	-2.4%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing stormslowly:feat/native-watcher-safetime (4f1743a) with main (ac43bc2)}

Footnotes

1. 40 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[codspeed-hq]

Unable to generate the flame graphs

The performance report has correctly been generated, but there was an internal error while generating the flame graphs for this run. We're working on fixing the issue. Feel free to contact us on Discord or at support@codspeed.io if the issue persists.