bug#59763: 29.0.60; Filling for c-ts-mode

[Yuan Fu <casouri@gmail.com>]

Theodor Thornhill <theo@thornhill.no> writes:

> Yuan Fu <casouri@gmail.com> writes:
>
>>> On Dec 26, 2022, at 2:03 PM, Theodor Thornhill <theo@thornhill.no> wrote:
>>> 
>>> Yuan Fu <casouri@gmail.com> writes:
>>> 
>>>> Theodor Thornhill <theo@thornhill.no> writes:
>>>> 
>>>>> On 25 December 2022 02:30:35 CET, Yuan Fu <casouri@gmail.com> wrote:
>>>>>> 
>>>>>> Theodor Thornhill <theo@thornhill.no> writes:
>>>>>> 
>>>>>>> On 24 December 2022 09:36:21 CET, Yuan Fu <casouri@gmail.com> wrote:
>>>>>>>> 
>>>>>>>> Theodor Thornhill <theo@thornhill.no> writes:
>>>>>>>> 
>>>>>>>>> Yuan Fu <casouri@gmail.com> writes:
>>>>>>>>> 
>>>>>>>>>>> On Dec 2, 2022, at 6:58 AM, Eli Zaretskii <eliz@gnu.org> wrote:
>>>>>>>>>>> 
>>>>>>>>>>>> From: Yuan Fu <casouri@gmail.com>
>>>>>>>>>>>> Date: Thu, 1 Dec 2022 21:33:06 -0800
>>>>>>>>>>>> 
>>>>>>>>>>>> 
>>>>>>>>>>>> IMO For c-ts-mode to be usable we need to have at least a basic filling
>>>>>>>>>>>> function. Below is the function I have in my init.el, could someone have
>>>>>>>>>>>> a look and see if it’s good? Alternatively we could copy out the comment
>>>>>>>>>>>> and fill it in a temp buffer with c-mode, but I didn’t have the time to try
>>>>>>>>>>>> it out and see how well it works.
>>>>>>>>>>> 
>>>>>>>>>>> Thanks.
>>>>>>>>>>> 
>>>>>>>>>>> From quick testing, I see a problem:
>>>>>>>>>>> 
>>>>>>>>>>> . Visit dispnew.c and go to the comment that starts on line 324.  Delete
>>>>>>>>>>>   the newline between the two lines of the comment, and invoke the
>>>>>>>>>>>   function.  Observe how the first non-blank character of the comment's
>>>>>>>>>>>   second line is aligned with the "/*" on the previous line, not with the
>>>>>>>>>>>   text after "/*" as I'd expect.
>>>>>>>>>> 
>>>>>>>>>> I see. I’ll need to look at how cc-mode fill comments.
>>>>>>>>>> 
>>>>>>>>>>> 
>>>>>>>>>>> Btw, this command should be bound to M-q in ts-c-mode.
>>>>>>>>>> 
>>>>>>>>>> Will do, once our fill function works well. BTW, Theo, if you have any
>>>>>>>>>> idea, don’t hesitate to go ahead :-) No obligations, of course.
>>>>>>>>>> 
>>>>>>>>>> Yuan
>>>>>>>>> 
>>>>>>>>> Sure!  Added to my list :)  I had a function at some point that used
>>>>>>>>> c-mode to do this. I'll see if I can polish it a little.
>>>>>>>> 
>>>>>>>> I did some work in filling, it should work like cc-mode in like 90% of
>>>>>>>> the cases now, yay!
>>>>>>>> 
>>>>>>>> Yuan
>>>>>>> 
>>>>>>> Nice! For all cc-ts-modes?
>>>>>> 
>>>>>> I only added for c and c++, but support for other modes should be
>>>>>> identical. And I think we should have something equivalent to cc-mode’s
>>>>>> init which sets up things that are the same in all C-like languages,
>>>>>> basically comments and filling.
>>>> 
>>>> I added indent and filling for other C-like modes.
>>>> 
>>>>>> But I wonder where should we put it, I guess it’s fine to leave it in
>>>>>> c-ts-mode, since there really isn’t much code. Having other modes to
>>>>>> require c-ts-mode shouldn’t be a big problem, I think?
>>>>>> 
>>>>>> Yuan
>>>>> 
>>>>> How about just having treesit-utils.el, or something like that? There
>>>>> are probably many things in the future that will be common among
>>>>> modes, yet won't really warrant inheritance. I think we have such an
>>>>> example in js/typescript too, iirc.
>>>> 
>>>> If it’s shared across all tree-sitter modes, it should be in treesit.el,
>>>> of course. We are talking about things shared by tree-sitter C-like
>>>> modes, so the scope is smaller.
>>>> 
>>>> Since right now it’s only a handful functions, I made other modes
>>>> require c-ts-mode.el. In the future if things accumulate, we can put
>>>> things into a separate file (c-ts-mode-common.el or something).
>>>> 
>>> 
>>> Sure.  I just don't like it when these namespaces blend too much.  But
>>> your call :-)
>>
>> It can be treesit-c-common.el or something. We’ll figure it out when the time comes ;-)
>>
>
> Yep :)
> Theo

I’ve added this to other modes, closing this report.

Yuan