Add benchmarks for table rendering

[majiayu000]

Fixes #7727

Description

Add table rendering benchmarks to benches/benchmarks.rs using the existing tango-bench framework. These benchmarks measure NuTable::draw() performance — the actual rendering pipeline that converts structured data into formatted terminal output.

This fills the gap between existing table data-operation benchmarks (create/get/select/insert) and actual display rendering.

New benchmarks:

• bench_table_render(rows, cols) — Core rendering at varying dimensions: (10,3), (100,5), (1000,5), (1000,15)
• bench_table_render_with_theme(name, theme) — Rendering with different themes (basic, rounded, thin, heavy, none) at 100x5
• bench_table_render_wide(termwidth) — Rendering at varying terminal widths (40, 80, 120, 200) at 100x10

All benchmarks use NuTable directly (not through the eval engine) to isolate rendering from parsing overhead.

Release notes summary - What our users need to know

Added Criterion-style benchmarks for table rendering (NuTable::draw()) covering different table sizes, themes, and terminal widths.

Tasks after submitting

• Update the documentation

[fdncred]

I'm up for this but there's a lot of changes to the cargo.lock file. i'd like those changes reverted except for the changes related to this PR. it looks like cargo update was ran by mistake.

[flinesse]

Careful #17240 (comment)

[fdncred]

Careful #17240 (comment)

what's your point? are you saying @majiayu000 is an AI bot?

[flinesse]

I can't claim with certainty but the comment on the linked issue and the prior MR attempt raises all kinds of flags.

Just sending up a flare so you can decide how much time you want to invest. Sorry for the vague comment prior.

[fdncred]

I can't claim with certainty but the comment on the linked issue and the prior MR attempt raises all kinds of flags.

Just sending up a flare so you can decide how much time you want to invest. Sorry for the vague comment earlier.

yes, indeed. I just saw that just now. it does raise flags. @majiayu000 who is doing your coding for you?

[majiayu000]

@fdncred I use claude code for planing issue, codex + claude code for cross-review , but this may not slove some issue, I am trying to improve my system. Feel free to close my pr if you feel bother or pr is not a right way.

[cptpiepmatz]

This still has many changes in Cargo.lock for no reason. Also while I like having more benchmarks, I feel like we should first look into making our benchmarking system more useful by reducing as much overhead as possible. I think we can go back to this, once we have a better benchmarking system in place.