bug#73129: 31.0.50; buffer-text-pixel-size vs. newlines

bug#73129: 31.0.50; buffer-text-pixel-size vs. newlines

[David Ponce via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Hello,

Its seems that the function `buffer-text-pixel-size' is not working
as described when there are newlines in the buffer.  Here is a quick
illustration (on my laptop (frame-char-width) returns 12 pixels):

;; As expected.
(with-temp-buffer
   (insert "abcdef")
   (car (buffer-text-pixel-size nil nil t)))
72

;; I suppose it is as expected because newline is not displayed.
(with-temp-buffer
   (insert "\n")
   (car (buffer-text-pixel-size nil nil t)))
0

;; ?
(with-temp-buffer
   (insert "abcd\nef")
   (car (buffer-text-pixel-size nil nil t)))
48

The doc string of `buffer-text-pixel-size' mentions:

"Return size of whole text of BUFFER-OR-NAME in WINDOW."

So ,I expect that the last example will return 72px (6 x 12px) + 0px
for the newline.  But the result is 48px, which means that all the
text after the first newline is not counted.

This also impacts "string-pixel-width" which is supposed to return
the pixel size of the passed string, not part of it:

(string-pixel-width "abcdef") => 72
(string-pixel-width "\n") => 0
(string-pixel-width "abcd\nef") => 48
(string-pixel-width "abcd\nef\ngh") => 48

At least this limitation should be mentioned in the doc string?

A possible fix for `string-pixel-width' could be to remove all
the newlines from the passed string before to call
`buffer-text-pixel-size'?  Something like this:

diff --git a/lisp/emacs-lisp/subr-x.el b/lisp/emacs-lisp/subr-x.el
index 4cdb065feeb..c68b3b37f9b 100644
--- a/lisp/emacs-lisp/subr-x.el
+++ b/lisp/emacs-lisp/subr-x.el
@@ -406,6 +406,11 @@ string-pixel-width
                          face-remapping-alist))
          (kill-local-variable 'face-remapping-alist))
        (insert string)
+      ;; Remove newlines.
+      (forward-line 0)
+      (while (not (bobp))
+        (delete-char -1)
+        (forward-line 0))
        ;; Prefer `remove-text-properties' to `propertize' to avoid
        ;; creating a new string on each call.
        (remove-text-properties


Thanks!

In GNU Emacs 31.0.50 (build 1, x86_64-pc-linux-gnu, GTK+ Version
  3.24.43, cairo version 1.18.0)
Repository revision: 2ce0d397b12cafcb3e1ac5630bc3fbca61bd6b87
Repository branch: master
Windowing system distributor 'The X.Org Foundation', version 11.0.12014000
System Description: Fedora Linux 40 (KDE Plasma)

Configured using:
  'configure --with-x-toolkit=gtk3 --with-cairo-xcb
  --with-native-compilation=no
  PKG_CONFIG_PATH=/usr/local/lib/pkgconfig:/usr/lib/pkgconfig'

Configured features:
ACL CAIRO DBUS FREETYPE GIF GLIB GMP GNUTLS GPM GSETTINGS HARFBUZZ JPEG
LCMS2 LIBOTF LIBSELINUX LIBSYSTEMD LIBXML2 M17N_FLT MODULES NOTIFY
INOTIFY PDUMPER PNG RSVG SECCOMP SOUND SQLITE3 THREADS TIFF
TOOLKIT_SCROLL_BARS TREE_SITTER WEBP X11 XDBE XIM XINPUT2 XPM GTK3 ZLIB

Important settings:
   value of $LC_TIME: fr_FR.utf8
   value of $LANG: fr_FR.UTF-8
   locale-coding-system: utf-8-unix

bug#73129: 31.0.50; buffer-text-pixel-size vs. newlines

[Eli Zaretskii]

> Date: Mon, 9 Sep 2024 00:22:11 +0200
> From:  David Ponce via "Bug reports for GNU Emacs,
>  the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>
> 
> Its seems that the function `buffer-text-pixel-size' is not working
> as described when there are newlines in the buffer.  Here is a quick
> illustration (on my laptop (frame-char-width) returns 12 pixels):
> 
> ;; As expected.
> (with-temp-buffer
>    (insert "abcdef")
>    (car (buffer-text-pixel-size nil nil t)))
> 72
> 
> ;; I suppose it is as expected because newline is not displayed.
> (with-temp-buffer
>    (insert "\n")
>    (car (buffer-text-pixel-size nil nil t)))
> 0
> 
> ;; ?
> (with-temp-buffer
>    (insert "abcd\nef")
>    (car (buffer-text-pixel-size nil nil t)))
> 48

The above is the intended behavior, assuming that the default width of
the frame's characters is 12 in your case.  The maximum width of the
text "abcd\nef" is 4 characters ("abcd"), which are 12*4 = 48 pixels.
The part after the newline is only 2 characters, so it is narrower,
and does not affect the result.

IOW, buffer-text-pixel-size measures the 2D _dimensions_ of the text,
not is total width.

> The doc string of `buffer-text-pixel-size' mentions:
> 
> "Return size of whole text of BUFFER-OR-NAME in WINDOW."

It also says:

  The return value is a cons of the maximum pixel-width of any text
  line and the pixel-height of all the text lines of the buffer
  specified by BUFFER-OR-NAME.

Does the above explain the behavior you see?

This is not a bug.

> So ,I expect that the last example will return 72px (6 x 12px) + 0px
> for the newline.  But the result is 48px, which means that all the
> text after the first newline is not counted.

Your expectations are incorrect, see above.

> This also impacts "string-pixel-width" which is supposed to return
> the pixel size of the passed string, not part of it:
> 
> (string-pixel-width "abcdef") => 72
> (string-pixel-width "\n") => 0
> (string-pixel-width "abcd\nef") => 48
> (string-pixel-width "abcd\nef\ngh") => 48

Yes.  If you want the _total_ width of a string, remove the newlines
from it.

I think I understand the source of your confusion, so I have now
clarified these subtle points in the doc strings of
window-text-pixel-size, buffer-text-pixel-size, and
string-pixel-width.

> At least this limitation should be mentioned in the doc string?

I've now done so (and they are not limitations, but conscious design
decisions which have good reasons).

> A possible fix for `string-pixel-width' could be to remove all
> the newlines from the passed string before to call
> `buffer-text-pixel-size'?  Something like this:

Thanks, but that would be the wrong thing.  If the caller wants a
total width, the caller should remover the newlines or measure each
substring separately and add the results.

May I ask where and for what purpose did you need to measure pixel
width of a string that included newlines?

bug#73129: 31.0.50; buffer-text-pixel-size vs. newlines

[David Ponce via Bug reports for GNU Emacs, the Swiss army knife of text editors]

On 09/09/2024 1:34 PM, Eli Zaretskii wrote:
>> Date: Mon, 9 Sep 2024 00:22:11 +0200
>> From:  David Ponce via "Bug reports for GNU Emacs,
>>   the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>
>>
>> Its seems that the function `buffer-text-pixel-size' is not working
>> as described when there are newlines in the buffer.  Here is a quick
>> illustration (on my laptop (frame-char-width) returns 12 pixels):
>>
>> ;; As expected.
>> (with-temp-buffer
>>     (insert "abcdef")
>>     (car (buffer-text-pixel-size nil nil t)))
>> 72
>>
>> ;; I suppose it is as expected because newline is not displayed.
>> (with-temp-buffer
>>     (insert "\n")
>>     (car (buffer-text-pixel-size nil nil t)))
>> 0
>>
>> ;; ?
>> (with-temp-buffer
>>     (insert "abcd\nef")
>>     (car (buffer-text-pixel-size nil nil t)))
>> 48
> 
> The above is the intended behavior, assuming that the default width of
> the frame's characters is 12 in your case.  The maximum width of the
> text "abcd\nef" is 4 characters ("abcd"), which are 12*4 = 48 pixels.
> The part after the newline is only 2 characters, so it is narrower,
> and does not affect the result.
> 
> IOW, buffer-text-pixel-size measures the 2D _dimensions_ of the text,
> not is total width.

Thanks for clarifying.

> 
>> The doc string of `buffer-text-pixel-size' mentions:
>>
>> "Return size of whole text of BUFFER-OR-NAME in WINDOW."
> 
> It also says:
> 
>    The return value is a cons of the maximum pixel-width of any text
>    line and the pixel-height of all the text lines of the buffer
>    specified by BUFFER-OR-NAME.
> 
> Does the above explain the behavior you see?

Yes

> 
> This is not a bug.
> 
>> So ,I expect that the last example will return 72px (6 x 12px) + 0px
>> for the newline.  But the result is 48px, which means that all the
>> text after the first newline is not counted.
> 
> Your expectations are incorrect, see above.
> 
>> This also impacts "string-pixel-width" which is supposed to return
>> the pixel size of the passed string, not part of it:
>>
>> (string-pixel-width "abcdef") => 72
>> (string-pixel-width "\n") => 0
>> (string-pixel-width "abcd\nef") => 48
>> (string-pixel-width "abcd\nef\ngh") => 48
> 
> Yes.  If you want the _total_ width of a string, remove the newlines
> from it.
> 
> I think I understand the source of your confusion, so I have now
> clarified these subtle points in the doc strings of
> window-text-pixel-size, buffer-text-pixel-size, and
> string-pixel-width.

Thanks you very much!

> 
>> At least this limitation should be mentioned in the doc string?
> 
> I've now done so (and they are not limitations, but conscious design
> decisions which have good reasons).
> 
>> A possible fix for `string-pixel-width' could be to remove all
>> the newlines from the passed string before to call
>> `buffer-text-pixel-size'?  Something like this:
> 
> Thanks, but that would be the wrong thing.  If the caller wants a
> total width, the caller should remover the newlines or measure each
> substring separately and add the results.

I agree now that you clarified the behavior of `string-pixel-width' when
it is passed a string with embedded newlines.

> May I ask where and for what purpose did you need to measure pixel
> width of a string that included newlines?

Actually, the current behavior doesn't really impact my current work. I
just noticed it while experimenting, having passed to `string-pixel-width'
a buffer substring that spanned multiple lines.  And I thought it would be
good to, at least, clarify the behavior I observed ;-)

You can close this bug.

Many thanks again for your time and assistance!

bug#73129: 31.0.50; buffer-text-pixel-size vs. newlines

[Eli Zaretskii]

> Date: Mon, 9 Sep 2024 14:56:22 +0200
> Cc: 73129@debbugs.gnu.org
> From: David Ponce <da_vid@orange.fr>
> 
> > May I ask where and for what purpose did you need to measure pixel
> > width of a string that included newlines?
> 
> Actually, the current behavior doesn't really impact my current work. I
> just noticed it while experimenting, having passed to `string-pixel-width'
> a buffer substring that spanned multiple lines.  And I thought it would be
> good to, at least, clarify the behavior I observed ;-)
> 
> You can close this bug.

Thanks, closing.