Tree-sitter fixes (1.116 edition) #968
Conversation
(Apologies for this oversight.)
…when defining hyperlink injection.
…with one based on a better parser. The `ikatyang` parser is great… until multi-byte input comes into play, or some other construct it doesn't understand. It doesn't have any way to recover when it hits those cases; it just throws exceptions, and WASM can't catch them. The `MDeiml` parser for Markdown had one deal-breaking issue (tree-sitter-grammars/tree-sitter-markdown#92), and once that got solved, it emerged as the clear victor. Several people have been using this grammar as `language-markdown-alpha` on the repository, and none of them have encountered any problems of the sort they did with the `ikatyang` parser.
…like `export * from 'foo'`.
…and a `meta` scope to every type annotation.
…to match the names we use in JavaScript.
…so they don’t also apply to namespaced TSX tags.
|
Taking this one out of draft a week early because of how big it is (the replacement Markdown grammar). |
There was a problem hiding this comment.
Without testing, LGTM - when reading through there were no obvious "WTF"s
I saw the comments in src/wasm-tree-sitter-grammar, do we know why that file has an activation/deactivation? It's not like that file is a grammar unto itself, so there shouldn't be any need for [de]activations, right?
It's a grammar, but it's not a package, so you're right that it's not entirely clear why it has |
|
Hmm...do we know what happens if we remove them...? |
No, but I'll leave that until next month's PR. :) |
This is a great one! I'm declaring victory in our experiment from #878. We are switching Markdown parsers. The newer one is actively maintained and has fixed the one bug that was standing in the way of our original adoption.
I've also thrown in some TypeScript fixes, and I expect for there to be a few more by the time I'm done.