Christian Clason

Also, can you pull out the unrelated changes to a separate PR? The `@nospell` for `link_label` is also debatable, as the label often _is_ normal text that should be spell-checked.

Closing since I still think we need to be careful about default conceal; I'd prefer to leave this driven by upstream. Thanks, though!

The LaTeX parser is currently being heavily refactored in a separate branch; I'd suggest waiting for that to get merged before looking at this. (It might be that this requires...

You can also just enable this manually ```lua vim.bo.indentexpr = "nvim_treesitter#indent()" ``` in a filetype plugin or filetype autocommand (this is all `enable=true` does, and in fact the recommended way...

@ribru17 What's the status here? We "own" our queries (which are editor-specific) so we don't want to get blocked by slow upstream maintenance unless absolutely necessary.

I worry a bit that these queries are overly specific at the cost of performance -- is this really necessary for acceptable highlighting?

Testing `ts g` with HEAD CLI gives the dreaded ``` Warning: rule number is just a `seq` or `choice` rule with a single element. This is unnecessary. ``` This often...

> Based on the grammar and the json file produced it seems like there are no states where number has a redundant seq or choice There are no false positive...

> Unless just the quantity of queries themselves has a non-negligible effect It very much does, but I don't think these are excessive (but I haven't tested these myself). `check-queries`...

Honestly, I think this would be better solved on the level of the parser, which could parse align-style equations. Mind asking there first? (regex or pattern matches can be quite...