The code below (Javascript) doesn’t get highlighted properly. We should have all stringN be highlighted with string color, but only string1 and string4 are.
const a = `string1 ${var1 + `string2${`string3` + var2}`}string4`;
What is supposed to be happening is that the whole string gets the string face, then the template substitution gets the embedded face, then the nested string gets (again) the string face, the nested substitution gets the embedded face, and so on.
The problem is that (as far as I understand) the priority of the queries is defined by the ordering in the queries file, but here we want both queries to have the same priority (more exactly we want shorter matches to have priority above longer matches). I'm not even sure if that's something that is supposed to work in upstream tree-sitter?
Anyway, I managed to fix it somehow (not tested very much) by ordering the matches by length (using cl-stable-sort) before applying the text properties, but I'm not sure if that's what we're supposed to do?
There could also be performance issues by sorting the matches at every keystroke (although I tried a bit on a 3500 lines file and couldn’t see any issue so far).
The code below (Javascript) doesn’t get highlighted properly. We should have all
stringNbe highlighted with string color, but onlystring1andstring4are.What is supposed to be happening is that the whole string gets the
stringface, then the template substitution gets theembeddedface, then the nested string gets (again) thestringface, the nested substitution gets theembeddedface, and so on. The problem is that (as far as I understand) the priority of the queries is defined by the ordering in the queries file, but here we want both queries to have the same priority (more exactly we want shorter matches to have priority above longer matches). I'm not even sure if that's something that is supposed to work in upstreamtree-sitter?Anyway, I managed to fix it somehow (not tested very much) by ordering the matches by length (using
cl-stable-sort) before applying the text properties, but I'm not sure if that's what we're supposed to do? There could also be performance issues by sorting the matches at every keystroke (although I tried a bit on a 3500 lines file and couldn’t see any issue so far).Any thoughts?