[Based on #135] debug: add Chrome WebAssembly source sessions

[cpunion]

Implements Task 6 of xgo-dev#2164 on top of #135.

• adds standard Wasm build IDs and validates embedded/external DWARF identity
• builds a DWARF-derived browser index with source remapping, scopes, variables, and types
• adds an LLGo Chrome Language Extension and llgo debug -target=wasm launcher
• validates embedded DWARF, URL-escaped external sidecars, real LLGo artifacts, and no-extension fallback in pinned headless Chrome
• documents launch, remapping, sidecar, and optimized-out behavior

Local validation:

• Node Language Extension tests pass
• relevant Go packages pass
• Chrome for Testing 151 passes embedded/external fixtures, real embedded/external LLGo Wasm, and fallback

Known upstream/final-artifact caveat: the current LLGo Asyncify output still produces llvm-dwarfdump range verification diagnostics; this PR does not hide them or claim the umbrella issue is complete.

[gemini-code-assist]

Caution

The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased.

[github-actions]

LLGo baseline benchmarks

0c046904739e | workflow run | long-term charts

Program measurements

Platform	Workload	File size	vs main	Build	vs main	Run	vs main
Linux	cprintf	19344 B	+4.3% (worse)	358.064 ms	-13.2% (better)	1.272 ms	-18.0% (better)
Linux	fmtprintf	2218832 B	+0.2% (worse)	3.204 s	-16.9% (better)	2.394 ms	-27.2% (better)
Linux	println	73832 B	+1.6% (worse)	357.620 ms	-10.6% (better)	1.562 ms	-6.7% (better)
macOS	cprintf	84672 B	+0.0%	350.694 ms	-15.4% (better)	2.883 ms	-19.7% (better)
macOS	fmtprintf	2361696 B	+0.7% (worse)	3.549 s	-9.9% (better)	26.342 ms	+12.7% (worse)
macOS	println	125888 B	-0.9% (better)	355.164 ms	-14.3% (better)	4.257 ms	-41.4% (better)

Core language and compiler benchmarks

Platform	Benchmark	ns/op	vs main
Linux	BenchmarkLookupPCRandom	13.430 ns/op	-0.4% (better)
Linux	BenchmarkMergeCompilerFlags	149.800 ns/op	-3.0% (better)
Linux	BenchmarkMergeLinkerFlags	94.910 ns/op	-6.3% (better)
Linux	BenchmarkChannelBuffered	35.790 ns/op	+3.0% (worse)
Linux	BenchmarkChannelHandoff	25307 ns/op	-21.0% (better)
Linux	BenchmarkDefer	45.460 ns/op	-19.3% (better)
Linux	BenchmarkDirectCall	1.556 ns/op	-0.4% (better)
Linux	BenchmarkGlobalRead	1.556 ns/op	-0.3% (better)
Linux	BenchmarkGlobalWrite	2.479 ns/op	-0.6% (better)
Linux	BenchmarkGoroutine	41323 ns/op	-45.3% (better)
Linux	BenchmarkInterfaceCall	8.090 ns/op	+3.8% (worse)
Linux	BenchmarkRuntimeGetG	2.178 ns/op	-59.0% (better)
macOS	BenchmarkLookupPCRandom	13.680 ns/op	+10.0% (worse)
macOS	BenchmarkMergeCompilerFlags	129.100 ns/op	+3.7% (worse)
macOS	BenchmarkMergeLinkerFlags	72.450 ns/op	+4.5% (worse)
macOS	BenchmarkChannelBuffered	22.920 ns/op	-17.3% (better)
macOS	BenchmarkChannelHandoff	8329 ns/op	+18.8% (worse)
macOS	BenchmarkDefer	30.600 ns/op	-18.2% (better)
macOS	BenchmarkDirectCall	1.160 ns/op	-20.2% (better)
macOS	BenchmarkGlobalRead	1.029 ns/op	-15.0% (better)
macOS	BenchmarkGlobalWrite	1.070 ns/op	-28.0% (better)
macOS	BenchmarkGoroutine	37312 ns/op	+17.9% (worse)
macOS	BenchmarkInterfaceCall	4.650 ns/op	-13.9% (better)
macOS	BenchmarkRuntimeGetG	2.443 ns/op	-5.2% (better)

Compared only with the latest matching platform in the main series.

Warning

• Persistent publishing is unavailable because no data token is configured.