Add VHDL extension

[rapgenic]

This PR implements an extension for the VHDL language using

• https://github.com/alemuller/tree-sitter-vhdl for the tree-sitter grammar
• https://github.com/VHDL-LS/rust_hdl as the language server
• the zed csharp extension as a reference

This is my first attempt at creating an extension, so any feedback is appreciated!

I'm also marking this as a draft, because I cannot test MacOS and Windows, plus I'd like to use it for a few days to iron out any problems (I wrote it in a few hours, so there might be some rough mistakes...).

[cla-bot]

We require contributors to sign our Contributor License Agreement, and we don't have @rapgenic on file. You can sign our CLA at https://zed.dev/cla. Once you've signed, post a comment here that says '@cla-bot check'.

[rapgenic]

@cla-bot check

[cla-bot]

The cla-bot has been summoned, and re-checked this pull request!

[rapgenic]

I am having some problems with symbol outline and multiple variable definitions in a single statement, as per the following example:

variable a, b : Integer;

When showing the outline, the b variable appears as a child of the a variable, see the following picture:

The current tree-sitter query for that node is:

(variable_declaration
    "variable" @context
    (identifier_list (identifier) @name)
) @item

I cannot find a way to make them appear as two distinct variables, with the correct alignment...

[NickelLiang]

Propose to use a newer VHDL tree-sitter implementation here, instead of the one that seems to be unmaintained in the past years.

[rapgenic]

Thank you for the suggestion! I will have a look at it, but since it looks like it would be a complete rewrite of all the scm files, I am not sure it is worth the time! (The extension is basically finished, just waiting for zed-industries/zed#12739)

Is it solving some problem with the other grammar or better in some way?

Anyway I'm not against changing it, I just don't have much time ATM. If someone wanted to do the work I'd gladly accept a PR unless there is some regression

[rapgenic]

As a quick note, I have done a small test with the suggested new tree-sitter grammar and ATM I wasn't able to obtain an acceptable result. These are the problems I found:

• Grammar does not load properly in zed as is due to the use of the printf function:

  [2024-06-27T22:15:26+02:00 ERROR language::language_registry] failed to load language VHDL:
  Failed to instantiate wasm module: invalid import 'printf'
  

• Even removing all the printf calls from the source code, the parser seems to be a quite buggy (90% of source codes that worked with the previous grammar did not work with the suggested one, i.e. the tree view shows an ERROR node)

I did not investigate further, but since I have not found limitations with the older grammar and since the proposed grammar seems incomplete, I am going to keep using the old one for now.

[maxdeviant]

@rapgenic Since there hasn't been any activity on this PR in a few months, I'm going to close it, for now.

Feel free to reopen if you pick it up again.