chore: Polish the debug output of capability

[Xuanwo]

Which issue does this PR close?

Nope

Rationale for this change

The old debug output of capability requires multiple alloc.

What changes are included in this PR?

This PR makes it zero cost and removes some not useful debug output.

Are there any user-facing changes?

[codspeed-hq[bot]]

CodSpeed Performance Report

Merging #5315 will degrade performances by 15.89%

_{Comparing polish-capability (7310ef0) with main (7b18680)}

Summary

❌ 8 regressions ✅ 65 untouched benchmarks

:warning: Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	main	polish-capability	Change
❌	buffer 256 KiB * 32 chunk	183.6 ns	212.8 ns	-13.71%
❌	buffer 256 KiB * 4 chunk	154.4 ns	183.6 ns	-15.89%
❌	buffer 4.00 MiB * 32 chunk	154.4 ns	183.6 ns	-15.89%
❌	buffer 4.00 MiB * 4 chunk	154.4 ns	183.6 ns	-15.89%
❌	256 KiB * 1000k truncate	245.8 ns	275 ns	-10.61%
❌	256 KiB * 100k truncate	245.8 ns	275 ns	-10.61%
❌	256 KiB * 10k truncate	245.8 ns	275 ns	-10.61%
❌	256 KiB * 1k truncate	245.8 ns	275 ns	-10.61%