Re: Disable tree-sitter font-locking for smaller ranges

[Theodor Thornhill]

Yuan Fu <casouri@gmail.com> writes:

>
> Actually, thinking more of it, it working relies on the exact order of which 
> these rules are applied:
> 1. Outer template_string
> 2. Outer template_substitution
> 3. Inner template_string
> 4. Inner template_substitution

That's correct.

>
> Relying on such un-guaranteed fact is a bit uncomfortable, so I just
> wrote a function that did what you suggested initially: to fontify
> parts of template_string that’s not a template_substitution. WDYT? (I
> just pushed the change.)
>

I agree! The change looks good to me. Is there any change in performance
doing this? I'm not yet 100% on the inner workings, but if this would
slow things down that wouldn't be so nice.

> BTW, if you have time and energy, could you look into separating the
> queries into roughly three levels: minimum, moderate, and full
> fontification, and mark each with the new :feature flag? Or even
> better, separate them into different features (as suggested in the
> docstring of treesit-font-lock-feature-list).
>

I will do that promptly and attach patches for ts-mode and js-mode to
this thread.

Theo