refactor(rome_formatter): new format API, part 2

[ematipico]

Summary

Part of #1996

This PR refactors more code to use the new format API.

While doing so, I encountered new edge cases, because of that I created new APIs for these cases:

• try_format_with_or which allow the use of the try operator inside the closures
• try_format_with which allow the use of the try operator inside the closures

I've also fixed some bug in the process.

Test Plan

Existing tests should not break

[github-actions]

Parser conformance results on ubuntu-latest

T262

Test result	main count	This PR count	Difference
Total	45250	45250	0
Passed	44130	44130	0
Failed	1120	1120	0
Panics	0	0	0
Coverage	97.52%	97.52%	0.00%

TS

Test result	main count	This PR count	Difference
Total	15976	15976	0
Passed	11126	11126	0
Failed	4848	4848	0
Panics	2	2	0
Coverage	69.64%	69.64%	0.00%

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Pages

Latest commit:	29f7641
Status:	✅  Deploy successful!
Preview URL:	https://b44bc198.tools-8rn.pages.dev

View logs

[ematipico]

This case was uncovered by our formatter and it was buggy. The new trait system worked quite well.

[ematipico]

!bench_formatter

[github-actions]

Bench results on ubuntu-latest

group    main    pr
-----    ----    --
formatter
/compiler.js    1.00    298.9±0.89ms     3.5 MB/sec    1.00    299.4±1.14ms     3.5 MB/sec
formatter
/d3.min.js    1.00    209.7±1.09ms  1280.0 KB/sec    1.01    211.3±1.52ms  1270.4 KB/sec
formatter
/dojo.js    1.00     16.1±0.09ms     4.3 MB/sec    1.01     16.2±0.13ms     4.2 MB/sec
formatter
/jquery.min.js    1.00     61.4±0.44ms  1379.3 KB/sec    1.00     61.4±0.33ms  1378.0 KB/sec
formatter
/pixi.min.js    1.00    248.1±2.47ms  1811.5 KB/sec    1.00    248.7±0.95ms  1807.0 KB/sec
formatter
/react-dom.production.min.js    1.00     82.9±0.35ms  1421.1 KB/sec    1.01     83.7±0.53ms  1408.5 KB/sec
formatter
/react.production.min.js    1.00      3.5±0.00ms  1784.3 KB/sec    1.01      3.6±0.02ms  1772.9 KB/sec
formatter
/tex-chtml-full.js    1.00    656.6±1.75ms  1421.2 KB/sec    1.00    659.5±1.82ms  1414.9 KB/sec
formatter
/three.min.js    1.00    296.3±1.80ms  2029.3 KB/sec    1.00    297.4±1.37ms  2021.3 KB/sec
formatter
/vue.global.prod.js    1.00    101.9±0.50ms  1211.1 KB/sec    1.01    102.7±0.90ms  1201.7 KB/sec

[MichaReiser]

@ematipico the benchmark output is no longer printed as a table which makes it hard to read the results. Would you mind having a look?

[ematipico]

@ematipico the benchmark output is no longer printed as a table which makes it hard to read the results. Would you mind having a look?

It was never printed as a markdown table (check the code), but just as simple block and a mix of tabs/spaces to make the formatting decent. I can change it make a markdown table if you want

[MichaReiser]

@ematipico the benchmark output is no longer printed as a table which makes it hard to read the results. Would you mind having a look?

It was never printed as a markdown table (check the code), but just as simple block and a mix of tabs/spaces to make the formatting decent. I can change it make a markdown table if you want

You're right. It also prints somewhat misaligned locally. I wonder why that is. Maybe there's some way to get the nice formatting back on console & PR comment.

[ematipico]

@ematipico the benchmark output is no longer printed as a table which makes it hard to read the results. Would you mind having a look?

It was never printed as a markdown table (check the code), but just as simple block and a mix of tabs/spaces to make the formatting decent. I can change it make a markdown table if you want

You're right. It also prints somewhat misaligned locally. I wonder why that is. Maybe there's some way to get the nice formatting back on console & PR comment.

The output comes form critcmp, which is installed and run only during the CI. Unfortunately I don't see any markdown option. I guess the only solution would be to build our own.