bug#59633: 29.0.50; tree-sitter-bash: Incorrect behaviour of heredocs with expansions

[Yuan Fu <casouri@gmail.com>]

<miha@kamnitnik.top> writes:

> Yuan Fu <casouri@gmail.com> writes:
>
>> What’s the return value of (treesit-node-at (point)) at that point?
>> Probably the command_substitution node, and since it doesn’t have any
>> children, treesit-node-first-child-for-pos would return nil.
>
> (treesit-node-at (point) 'bash)
> => #<treesit-node (heredoc_body) in 12-56>
>
> I forgot to say that point is before "text2", outside of the
> command_substitution node.
>
>> Yuan
>
> tee << EOF
> text1 $var
> text2 $(echo cmd)
> text3
> $var2
> EOF
>
>
> Filling a test buffer with these contents and placing point before
> "text2", we get
>
>     (treesit-node-at (point) 'bash)
>     => #<treesit-node (heredoc_body) in 12-56>
>
>     (treesit-node-children (treesit-node-at (point) 'bash))
>     => (#<treesit-node (simple_expansion) in 18-22>
>         #<treesit-node (command_substitution) in 29-40>
>         #<treesit-node (simple_expansion) in 47-52>)
>
> This is now expected.
>
>     (point)
>     => 23
>
>     (treesit-node-first-child-for-pos (treesit-node-at (point) 'bash) (point))
>     => nil
>
> This is not expected. One would expect this to return
> command_substitution node, since it follows the position 23.

Ah, I (finally) see what you mean. I agree that it should return the
command_substitution node. It seems that tree-sitter returns nil
whenever POS is not in a child node. I need to see what’s the problem
here since treesit-node-first-child-for-pos is just a think wrapper
around the C counterpart.

Yuan