bug#59810: [PATCH] Fix rcirc-markup-text-functions applying to the whole message

bug#59810: [PATCH] Fix rcirc-markup-text-functions applying to the whole message

[Thuna]

[-- Attachment #1: Type: text/plain, Size: 628 bytes --]

Since `rcirc-markup-text-functions' is meant to apply only to the text,
the changed code, which applied it to the entirety of the message was
bugged.  The simplest example is `\me NICK' with `rcirc-color', where
NICK is supposed to be colored but isn't because the message reads
`[... NICK]' and the `]' is recognized as a part of the nick.

The patch should fix the issue, although it is possible that some
already existing functions put to `rcirc-markup-text-functions' will
break due to them expecting the previous behavior.  I still propose the
patch is applied as the issue is a bug on those functions' end and not
rcirc.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: the patch --]
[-- Type: text/x-patch, Size: 1468 bytes --]

From ae0334e422f84941218acdd0a953b6f1f2d34fac Mon Sep 17 00:00:00 2001
From: Thuna <thuna.cing@gmail.com>
Date: Sun, 4 Dec 2022 00:42:37 +0100
Subject: [PATCH] Fix `rcirc-markup-text-functions' applying to the whole
 message

* lisp/net/rcirc.el (rcirc-print): Apply `rcirc-markup-text-functions'
only to the text part of the message.
---
 lisp/net/rcirc.el | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/lisp/net/rcirc.el b/lisp/net/rcirc.el
index 1fdf41a35ee..96109dcd5c9 100644
--- a/lisp/net/rcirc.el
+++ b/lisp/net/rcirc.el
@@ -2080,8 +2080,13 @@ rcirc-print
                              (point)))
               (when (rcirc-buffer-process)
                 (save-excursion (rcirc-markup-timestamp sender response))
-                (dolist (fn rcirc-markup-text-functions)
-                  (save-excursion (funcall fn sender response)))
+                (save-restriction
+                  (when-let ((prop-match (text-property-search-forward 'rcirc-text)))
+                    (goto-char (prop-match-beginning prop-match))
+                    (narrow-to-region (prop-match-beginning prop-match)
+                                      (prop-match-end prop-match))
+                    (dolist (fn rcirc-markup-text-functions)
+                      (save-excursion (funcall fn sender response)))))
                 (when rcirc-fill-flag
                   (save-excursion (rcirc-markup-fill sender response))))
 
-- 
2.37.4

bug#59810: [PATCH] Fix rcirc-markup-text-functions applying to the whole message

[Philip Kaludercic]

Thuna <thuna.cing@gmail.com> writes:

> Since `rcirc-markup-text-functions' is meant to apply only to the text,
> the changed code, which applied it to the entirety of the message was
> bugged.  The simplest example is `\me NICK' with `rcirc-color', where
> NICK is supposed to be colored but isn't because the message reads
> `[... NICK]' and the `]' is recognized as a part of the nick.

I can not reproduce this, can you give a different example?

> The patch should fix the issue, although it is possible that some
> already existing functions put to `rcirc-markup-text-functions' will
> break due to them expecting the previous behavior.  I still propose the
> patch is applied as the issue is a bug on those functions' end and not
> rcirc.
>
>>From ae0334e422f84941218acdd0a953b6f1f2d34fac Mon Sep 17 00:00:00 2001
> From: Thuna <thuna.cing@gmail.com>
> Date: Sun, 4 Dec 2022 00:42:37 +0100
> Subject: [PATCH] Fix `rcirc-markup-text-functions' applying to the whole
>  message
>
> * lisp/net/rcirc.el (rcirc-print): Apply `rcirc-markup-text-functions'
> only to the text part of the message.
> ---
>  lisp/net/rcirc.el | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/lisp/net/rcirc.el b/lisp/net/rcirc.el
> index 1fdf41a35ee..96109dcd5c9 100644
> --- a/lisp/net/rcirc.el
> +++ b/lisp/net/rcirc.el
> @@ -2080,8 +2080,13 @@ rcirc-print
>                               (point)))
>                (when (rcirc-buffer-process)
>                  (save-excursion (rcirc-markup-timestamp sender response))
> -                (dolist (fn rcirc-markup-text-functions)
> -                  (save-excursion (funcall fn sender response)))
> +                (save-restriction
> +                  (when-let ((prop-match (text-property-search-forward 'rcirc-text)))
> +                    (goto-char (prop-match-beginning prop-match))
> +                    (narrow-to-region (prop-match-beginning prop-match)
> +                                      (prop-match-end prop-match))

Why is this necessary, if the only function that would have modified the
buffer gap is `rcirc-markup-timestamp'?

> +                    (dolist (fn rcirc-markup-text-functions)
> +                      (save-excursion (funcall fn sender response)))))
>                  (when rcirc-fill-flag
>                    (save-excursion (rcirc-markup-fill sender response))))

bug#59810: [PATCH] Fix rcirc-markup-text-functions applying to the whole message

[Thuna]

>
>> Since `rcirc-markup-text-functions' is meant to apply only to the text,
>> the changed code, which applied it to the entirety of the message was
>> bugged.  The simplest example is `\me NICK' with `rcirc-color', where
>> NICK is supposed to be colored but isn't because the message reads
>> `[... NICK]' and the `]' is recognized as a part of the nick.
>
> I can not reproduce this, can you give a different example?

Nothing else immediately comes to mind, unfortunately.  I am able to
reproduce it, however:
- emacs -Q 
- Load rcirc-color (version 0.4.5 as of now)
- M-x rcirc
- /query fsbot (assuming you're in libera.chat)
- Type a message so that fsbot responds and its color is registered
- "/me fsbot" --- You will see that fsbot is not colored
- "/me fsbot " --- You will see that fsbot is colored

My emacs version is GNU Emacs 30.0.50 (build 1, x86_64-pc-linux-gnu,
GTK+ Version 3.24.34) of 2023-01-09, in case that matters.

>> The patch should fix the issue, although it is possible that some
>> already existing functions put to `rcirc-markup-text-functions' will
>> break due to them expecting the previous behavior.  I still propose the
>> patch is applied as the issue is a bug on those functions' end and not
>> rcirc.

> Why is this necessary, if the only function that would have modified the
> buffer gap is `rcirc-markup-timestamp'?

I do not understand what you are referring to here.