--- Begin Message ---
Subject: |
29.0.50; Compilation warnings without tree-sitter |
Date: |
Mon, 21 Nov 2022 14:01:14 -0800 |
[Continuing from bug#59426]
Those warnings are due to no-tree-sitter build not having functions
defined in treesit.c. Eli fixed those warnings by adding
declare-function’s in every file using those functions. Can we make it a
bit nicer for lisp developers? Maybe defining those functions with
dummies with something like
(defun xxx
(error ’treesit "Tree-sitter not available"))
?
Yuan
--- End Message ---
--- Begin Message ---
Subject: |
Re: bug#59459: 29.0.50; Compilation warnings without tree-sitter |
Date: |
Tue, 22 Nov 2022 15:22:10 -0800 |
> On Nov 22, 2022, at 4:21 AM, Eli Zaretskii <eliz@gnu.org> wrote:
>
>> From: Yuan Fu <casouri@gmail.com>
>> Date: Mon, 21 Nov 2022 14:01:14 -0800
>>
>> [Continuing from bug#59426]
>>
>> Those warnings are due to no-tree-sitter build not having functions
>> defined in treesit.c. Eli fixed those warnings by adding
>> declare-function’s in every file using those functions. Can we make it a
>> bit nicer for lisp developers? Maybe defining those functions with
>> dummies with something like
>>
>> (defun xxx
>> (error ’treesit "Tree-sitter not available"))
>>
>> ?
>
> I'm not sure this is justified. The cases where a build lacks a very large
> group of primitives references in *.el files are quite rare in our practice;
> about the only two examples I know of are xwidgets.el and treesit.el (plus
> modes which use tree-sitter). It's easy enough (albeit annoying) to add a
> few declare-function's, so unless we are going to have more and more of
> these cases, adding a whole new infrastructure, let alone tricks like the
> above, which will raise a lot of brows, sound excessive to me.
If it cannot be easily done with existing tools, then I agree declare-function
is good enough.
Thanks,
Yuan
--- End Message ---