--- Begin Message ---
|
Subject: |
[PATCH] Don't fill yaml except comments and block scalars. |
|
Date: |
Sun, 28 Jan 2024 14:15:58 +0100 |
Tags: patch
Hi,
Currently, yaml-ts-mode fills comments and block scalars (multi-line
text literals) as expected, but re-fills the whole file when point is
outside of either of these constructs. Since yaml line breaks and
whitespace are significant, I'd say that this is never the correct
behavior.
This patch against current master inhibits M-q (fill-paragraph) outside
of comments and block scalars. In my tests default fill-paragraph
worked as expected both with and without justify, correctly detecting
comment and block literal boundaries, so I did not preserve the previous
code in `yaml-ts-mode--fill-paragraph'.
0001-Don-t-fill-yaml-except-comments-and-block-scalars.patch
Description: Text Data
--- End Message ---
--- Begin Message ---
|
Subject: |
Re: bug#68781: [PATCH] Don't fill yaml except comments and block scalars. |
|
Date: |
Thu, 01 Feb 2024 12:31:30 +0200 |
> Cc: Rudolf Schlatte <rudi@constantly.at>, 68781@debbugs.gnu.org
> Date: Tue, 30 Jan 2024 19:25:07 +0000
> From: Randy Taylor <dev@rjt.dev>
>
> On Monday, January 29th, 2024 at 20:20, Graham Marlow <graham@mgmarlow.com>
> wrote:
> >
> >
> > > Hello Randy, thanks for having a look! Could you tell me which
> >
> > > tree-sitter grammar you are using? I'm asking because with the grammar
> > > from https://github.com/ikatyang/tree-sitter-yaml I don't see what you
> > > describe.
> >
> >
> > For the record I'm also using this grammar.
> >
> > Looking at the patch, what do you think about retaining the existing
> > behavior (so block_scalars still fill correctly) while inhibiting
> > fill_paragraph for everything else as suggested? Originally I retained
> > the existing behavior of fill-paragraph just to limit the number of
> > things changed by the patch, not because it was working properly. I
> > think blocking the call to fill-paragraph for non-block/comment nodes
> > makes sense.
> >
> > I attached a patch w/ my edits, but it just swaps the when to and if,
> > accepts the comment node type for filling, and returns t to avoid
> > calling fill-paragraph for other nodes.
>
> Thanks Graham, the patch looks good to me.
>
> Would someone please install it on master? Thanks in advance.
Thanks, done, and closing the bug.
--- End Message ---