Re: [SPAM UNSURE] Re: Plug treesit.el into other emacs constructs

["Kévin Le Gouguec" <kevin.legouguec@gmail.com>]

Stephen Leake <stephen_leake@stephe-leake.org> writes:

> Kévin Le Gouguec <kevin.legouguec@gmail.com> writes:
>
>> Stephen Leake <stephen_leake@stephe-leake.org> writes:
>>
>>> ada-mode has something similar; forward-sexp goes to the next labeled
>>> keyword. That's more useful than stopping at every little AST node. It
>>> relies on markup in the grammar to label the keywords; tree-sitter would
>>> need another markup similar to the current indent markup.
>>
>> Chiming in since I got bitten by this in ada-mode right now: as someone
>> used to mark long_identifiers_with_underscores with C-M-SPC in other
>> modes, it's tripping me up that mark-sexp (and sexp movement in general)
>> "overshoots" and goes over much bigger expressions in ada-mode.
>
> Sounds like we need an option for this in ada-mode?

That'd be very welcome indeed, as far as I am concerned (no rush though;
I can take a stab at it on a rainy day if no-one else is clamoring for
it).

> My solution to this is to bind C-<right> to (forward-symbol 1), and
> C-<left> to (forward-symbol -1), so these keys move over whole
> identifiers.

Interesting; I don't think the *-symbol commands were on my radar (no
mark-symbol though AFAICT).

(I see now we also have {forward,backward}-list, bound to C-M-{n,p};
that's another pair I don't often use; I do use C-M-{d,u} a lot, but for
"horizontal" traversal I usually rely on C-M-{f,b})