ptx: add benchmark

[Aisha630]

ptx: add benchmark

Summary

I'm adding Divan benchmarks for ptx.

This is a benchmark-only change. No utility code is touched.

Why

ptx had no benchmark. Its cost is dominated by converting keyword byte offsets to character offsets while building the output, which grows with the line length. So the benchmarks hold the total input size roughly constant and vary the line length, which is what surfaces that behavior. It also gives us a baseline to measure the offset-conversion optimization I'm proposing separately.

Benchmarks

Bench	Args	Covers
ptx_short_lines	100_000 lines	many short lines
ptx_long_lines	100 lines	fixed ~1 MiB spread over fewer, longer lines
ptx_input_references_short_lines	100_000 lines	-r , short lines
ptx_input_references_long_lines	100 lines	-r on fewer, longer lines

I followed the guidelines in the performance.md file and drew inspiration from the tr, nl, cat benchmarks that were already set up.

[oech3]

Can we use file for benchmark and use std instead of rustix?

[Aisha630]

Can we use file for benchmark and use std instead of rustix?

Done. Removed the extra dependencies as well. I did the redirection earlier because running the benchmark locally was flooding the terminal. But I suppose that's a non-issue.

[github-actions]

GNU testsuite comparison:

Skip an intermittent issue tests/date/resolution (fails in this run but passes in the 'main' branch)
Skip an intermittent issue tests/tail/tail-n0f (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/cut/bounded-memory (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/date/date-locale-hour (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/seq/seq-epipe is now passing!