feat(rome_css_syntax): codegen for CSS grammar

[ematipico]

Summary

This PR is part of #2350

It starts to spin the ball around multi language support, by generating the code for the CSS grammar. The scope of this PR is exclusively having a code generation that works and emits code that compiles.

There are few things won't be tackled in this PR:

• repeated code between rome_js_syntax and rome_css_syntax; for AstNode is repeated at the moment and it's intentional for now;
• how valid is the css.ungram; it is still incomplete and it doesn't hold the grammar, the code was shown to Nicholas and Micha few weeks back to show them the differences from the JS grammar and if we can achieve that;

What's been done:

• created a new rome_css_syntax crate that contains the generated code and some other code related to the syntax. Some code is copied and repeated;
• the code generation now uses a LanguageKind to generate different code accordingly to the value of the variant;
• handled some edge cases that we now have with CSS. We have the $= which causes issues inside Rust ($ is mistaken as a pattern of a macro), so I had to play with proc_macro2 to generate the correct code;
• created an empty crate called rome_syntax where we will have our shared syntax, such as AstNode, AstSeparatedList, etc.

Test Plan

cargo check should compile

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Pages

Latest commit:	ea41599
Status:	✅  Deploy successful!
Preview URL:	https://18da30b6.tools-8rn.pages.dev

View logs

[github-actions]

Parser conformance results on ubuntu-latest

js/262

Test result	main count	This PR count	Difference
Total	45250	45250	0
Passed	44310	44310	0
Failed	940	940	0
Panics	0	0	0
Coverage	97.92%	97.92%	0.00%

jsx/babel

Test result	main count	This PR count	Difference
Total	39	39	0
Passed	36	36	0
Failed	3	3	0
Panics	0	0	0
Coverage	92.31%	92.31%	0.00%

ts/babel

Test result	main count	This PR count	Difference
Total	584	584	0
Passed	508	508	0
Failed	76	76	0
Panics	0	0	0
Coverage	86.99%	86.99%	0.00%

ts/microsoft

Test result	main count	This PR count	Difference
Total	15983	15983	0
Passed	12167	12167	0
Failed	3816	3816	0
Panics	0	0	0
Coverage	76.12%	76.12%	0.00%

[xunilrj]

Equals here make sense? Given that __LAST is just a marker and not a node.

[MichaReiser]

Using equals ensures it's symmetric with the impl From<CssSyntaxKind> for u16

[ematipico]

This was a copy paste from rome_js_syntax. I think it would make senses to keep debug_assert, or remove it. Although the same change should be applied to the other crate too. I would leave it out of the scope for now, for this PR.

[leops]

This should stay as an assert! since it's guarding unsafe code, if something goes wrong with this conversion a panic will be a lot easier to debug than some random undefined behavior down the line when the code matches on an invalid enum

[MichaReiser]

Using equals ensures it's symmetric with the impl From<CssSyntaxKind> for u16