[Superseded by #2327] meta: map package metadata on Windows

[cpunion]

Superseded by #2327.

The Windows package-metadata mapping from this PR has been replayed as its own commit on the latest main in #2327, together with the related target identity, cache locking, and native Windows smoke coverage.

This PR remains Draft only as historical review context. New Windows work should depend on #2327 rather than this PR.

[github-actions]

LLGo baseline benchmarks

35dd92b97a14 | workflow run | long-term charts

Program measurements

Platform	Workload	File size	vs base	Build	vs base	Run	vs base
Linux	cprintf	18656 B	+0.0%	240.247 ms	-1.2% (better)	1.077 ms	+0.6% (worse)
Linux	fmtprintf	1874264 B	+0.0%	2.067 s	-1.1% (better)	2.800 ms	+0.2% (worse)
Linux	println	67992 B	+0.0%	237.254 ms	-1.6% (better)	1.322 ms	-4.3% (better)
macOS	cprintf	84624 B	+0.0%	281.939 ms	-12.5% (better)	2.442 ms	+4.1% (worse)
macOS	fmtprintf	1889968 B	+0.0%	1.995 s	-11.0% (better)	10.610 ms	-11.5% (better)
macOS	println	121168 B	+0.0%	277.214 ms	-2.0% (better)	3.528 ms	+9.7% (worse)

Core language and compiler benchmarks

Platform	Benchmark	ns/op	vs base
Linux	BenchmarkLookupPCRandom	9.552 ns/op	-3.9% (better)
Linux	BenchmarkMergeCompilerFlags	111.200 ns/op	-0.4% (better)
Linux	BenchmarkMergeLinkerFlags	72.470 ns/op	-0.9% (better)
Linux	BenchmarkChannelBuffered	28.410 ns/op	-0.1% (better)
Linux	BenchmarkChannelHandoff	19940 ns/op	-2.4% (better)
Linux	BenchmarkDefer	35.100 ns/op	+1.2% (worse)
Linux	BenchmarkDirectCall	1.364 ns/op	+0.1% (worse)
Linux	BenchmarkGlobalRead	1.363 ns/op	+0.0%
Linux	BenchmarkGlobalWrite	2.176 ns/op	+0.0% (worse)
Linux	BenchmarkGoroutine	24742 ns/op	-3.6% (better)
Linux	BenchmarkInterfaceCall	6.549 ns/op	-2.5% (better)
Linux	BenchmarkRuntimeGetG	1.909 ns/op	-0.1% (better)
macOS	BenchmarkLookupPCRandom	10.650 ns/op	+0.9% (worse)
macOS	BenchmarkMergeCompilerFlags	97.260 ns/op	+4.6% (worse)
macOS	BenchmarkMergeLinkerFlags	64.050 ns/op	+8.2% (worse)
macOS	BenchmarkChannelBuffered	20.500 ns/op	-2.8% (better)
macOS	BenchmarkChannelHandoff	6500 ns/op	+0.2% (worse)
macOS	BenchmarkDefer	24.730 ns/op	-0.0% (better)
macOS	BenchmarkDirectCall	0.947 ns/op	+0.5% (worse)
macOS	BenchmarkGlobalRead	0.942 ns/op	+0.1% (worse)
macOS	BenchmarkGlobalWrite	0.941 ns/op	-0.2% (better)
macOS	BenchmarkGoroutine	23657 ns/op	-11.7% (better)
macOS	BenchmarkInterfaceCall	4.079 ns/op	-0.1% (better)
macOS	BenchmarkRuntimeGetG	1.883 ns/op	-1.2% (better)

Compared with f65ddd1b0b4e measured in the same runner job.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!