Re: Average-user-facing interface for tree-sitter

[Stephen Leake]

Fu Yuan <casouri@gmail.com> writes:

>> 
>>> 
>>>>> Yeah.  Shouldn't it be possible to just have a global var instead of a
>>>>> mode?  That way we can just look for that variable when enabling the
>>>>> mode, and avoid calling anything other than what we want.  At least for
>>>>> the foreseeable future, enabling these per mode in the init file
>>>>> shouldn't really be too much of a problem, IMO.  When more users
>>>>> actually get to try this we can get a feel for how the init should best
>>>>> be handled.
>>>>> 
>>>>> To me the '*-use-tree-siter' defcustoms was beautiful :)
>>>> 
>>>> Back to centralized variable, perhaps?
>>> 
>>> I’ve thought really hard but didn’t come up with any brilliant ideas, so I’m
>>> going with centralized variable. Now, should I throw away recent commits and
>>> create a new branch so we don’t have so many changes back and forth on js.el
>>> and python.el? Plus feature/tree-sitter wouldn’t be used for long anyway
>>> since it’s merging into master soon.
>> 
>> I'd wait a bit more to see if some other ideas come up first.
>
> Here’s my thought (that didn’t go anywhere): since major modes sets a
> plethora of local hooks and variables, only the major mode itself
> knows how to reverse them. The cleanest way is probably to clear all
> the local variables and hooks and re-run the major mode setup, which
> suggests we should let major mode branch on whether to enable
> tree-sitter during initialization. 

That's what I ended up with in ada-mode; it decides on xref, indent,
face backend in the major-mode function based on ada-*-backend
variables. Actually, in a post-local-vars function; we need to allow
dir-local and file-local hooks to set those vars.

> I wonder if minor modes can somehow work with this model?

It seems to me minor modes are only needed if you want to have different
menus and keybindings for each minor mode. In ada-mode, I just
enable menu entries based on the current backends; I left keybindings
alone for now, which means many of them are undefined or non-functional.
Not the ideal solution.

Is there something written on why Emacs prefers minor-modes for choices
like this?

> It would be also nice to leave room for inclusion of other “backends”
> besides elisp and tree-sitter in the future.

Right. wisi is one, available now.

-- 
-- Stephe