[Based on #99] debug: add LLDB map and channel views

[cpunion]

Summary

• emit typed DWARF map buckets and channel wait queues from the existing compiler/runtime layouts without changing the runtime ABI
• show map length and typed key/value children, including overflow, growth, pointer, and indirect large-entry storage
• show channel length, capacity, closed state, and buffered values in receive order
• keep map growth and evacuation constants in the versioned debugger schema instead of hardcoding them in the LLDB adapter
• verify real map lookup and channel send/receive/close results in the LLDB fixture

Advances xgo-dev#2164 and the map/channel scope of xgo-dev#2154. Based on #99.

Tests

• macOS arm64, Apple LLDB 1703: 273/273
• Linux amd64, LLVM LLDB 19: 273/273
• TestStandardDWARF at O0/O1/O2/O3/Os/Oz
• macOS: go test ./ssa ./cmd/internal/lldb ./cmd/llgo ./internal/build
• macOS coverage: ssa 93.3%, cmd/internal/lldb 87.5%
• Linux coverage: internal/debugabi 93.8%, cmd/internal/lldb 87.5%
• Python syntax check and git diff --check

[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

d25ed330d334 | workflow run | long-term charts

Program measurements

Platform	Workload	File size	vs main	Build	vs main	Run	vs main
Linux	cprintf	18544 B	+0.0%	342.103 ms	-17.0% (better)	1.370 ms	-11.7% (better)
Linux	fmtprintf	2217600 B	+0.2% (worse)	3.357 s	-12.9% (better)	2.449 ms	-25.5% (better)
Linux	println	71504 B	-1.6% (better)	336.012 ms	-16.0% (better)	1.583 ms	-5.4% (better)
macOS	cprintf	84672 B	+0.0%	413.027 ms	-0.3% (better)	3.718 ms	+3.6% (worse)
macOS	fmtprintf	2361520 B	+0.7% (worse)	4.600 s	+16.8% (worse)	20.456 ms	-12.5% (better)
macOS	println	125712 B	-1.0% (better)	395.280 ms	-4.7% (better)	5.158 ms	-29.0% (better)

Core language and compiler benchmarks

Platform	Benchmark	ns/op	vs main
Linux	BenchmarkLookupPCRandom	13.310 ns/op	-1.3% (better)
Linux	BenchmarkMergeCompilerFlags	151.100 ns/op	-2.2% (better)
Linux	BenchmarkMergeLinkerFlags	95.070 ns/op	-6.2% (better)
Linux	BenchmarkChannelBuffered	34.010 ns/op	-2.2% (better)
Linux	BenchmarkChannelHandoff	26071 ns/op	-18.6% (better)
Linux	BenchmarkDefer	44.870 ns/op	-20.3% (better)
Linux	BenchmarkDirectCall	1.557 ns/op	-0.4% (better)
Linux	BenchmarkGlobalRead	1.875 ns/op	+20.2% (worse)
Linux	BenchmarkGlobalWrite	2.486 ns/op	-0.3% (better)
Linux	BenchmarkGoroutine	32902 ns/op	-56.5% (better)
Linux	BenchmarkInterfaceCall	8.100 ns/op	+3.9% (worse)
Linux	BenchmarkRuntimeGetG	1.868 ns/op	-64.8% (better)
macOS	BenchmarkLookupPCRandom	15.260 ns/op	+22.7% (worse)
macOS	BenchmarkMergeCompilerFlags	133.200 ns/op	+7.0% (worse)
macOS	BenchmarkMergeLinkerFlags	99.830 ns/op	+44.0% (worse)
macOS	BenchmarkChannelBuffered	27.520 ns/op	-0.8% (better)
macOS	BenchmarkChannelHandoff	9913 ns/op	+41.4% (worse)
macOS	BenchmarkDefer	34.530 ns/op	-7.7% (better)
macOS	BenchmarkDirectCall	1.329 ns/op	-8.5% (better)
macOS	BenchmarkGlobalRead	1.119 ns/op	-7.6% (better)
macOS	BenchmarkGlobalWrite	1.169 ns/op	-21.4% (better)
macOS	BenchmarkGoroutine	32075 ns/op	+1.3% (worse)
macOS	BenchmarkInterfaceCall	5.925 ns/op	+9.8% (worse)
macOS	BenchmarkRuntimeGetG	2.390 ns/op	-7.3% (better)

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

Warning

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