-
-
Notifications
You must be signed in to change notification settings - Fork 90
Description
Hey folks!
I'm (very slowly) working on a rust version of textmate-based highlighting, and your collection of grammars is hugely helpful. However, since my parser is a bit more strict than vscode's, I found a few grammars that have issues in them. Here's an example from xml.json:
{
"begin": "<%--",
"captures": {
"0": {
"name": "punctuation.definition.comment.xml"
},
"end": "--%>",
"name": "comment.block.xml"
}
},It's obviously a typo, end and name should be one level up.
information_for_contributors (taken from https://github.com/microsoft/vscode/blob/main/extensions/xml/syntaxes/xml.tmLanguage.json I presume) suggests upstreaming fixes, but the original grammar https://github.com/atom/language-xml/blob/master/grammars/xml.cson is now archived, so it can't be fixed.
I found similar issues in wikitext.json, stata.json, and racket.json. Additionally, jinja.json uses arrays for captures, which is a bit funny. I get why it works in JS, but stronger typed languages fail on that.
It'd be lovely to fix those grammars, so what would be your preferred strategy?
- should I submit a fix in vscode for them?
- should I fork archived grammars, introduce fixes, and propose new upstream here (and maybe in vscode)?
- should I submit a PR with "fixups" for those grammars (maybe using
patch)? - would you like to have some form of validation for grammars, maybe a JSON schema?