EndMarker and how to handle it

[KaranAhlawat]

So, for a while, I've been trying to bend EndMarkers to my will, since I opened #275. But, unfortunately, it has escaped me how we can group them. I think, looking at the EBNF grammar, it's not really possible to group them with their corresponding openings.

At least not if we implement the tree-sitter grammar according to the EBNF. But, even if we do deviate from it, we still run into issues because we can't really figure out which statement an end marker might belong to, without the rest of the parse tree.

So, my question is, should I just fall back to using regexp to determine the correct indentation for a end marker? Or is there something we can do on the parser level for this?

[susliko]

Deviations from the ebnf are okay.
I believe, that we can handle them with tree-sitter. I did some digging and was able to parse a majority of end marker use cases, but some tests for other cases broke, and I haven't spent enough time to fix them

[KaranAhlawat]

I haven't been able to make them optional and parse them correctly. Also, wasn't able to figure out how to capture those correctly for given_expressions. If you guys have any pointers, I'd love to go back to this once again and try to fix it.