bug#33648: 25.2; feature request: configurable position of strike-through line

[Stefan Kangas <stefan@marxist.se>]

[-- Attachment #1: Type: text/plain, Size: 1155 bytes --]

tags 33648 + confirmed
thanks

Robert Samal <samal@iuuk.mff.cuni.cz> writes:

> When I use set-face-attribute to give some text a strike-through
> attribute, I see that emacs behaves differently from other pieces of
> software: Apparently, the line is drawn at the level of half the glyph's
> height -- which causes the line to be just touching letters like "a",
> "c", "e", ... and it isn't perceived as striking the text out (at least
> not by me).
>
> In comparison, web browsers interpreting html strike tag make the
> line at the level of half of the "small lowercase letters" -- basically
> at the level of the line in lowercase "e". The same is done by Libre
> Office ... and others.
>
> My suggestion is to make this configurable in some way. I'm not sure,
> if it would be better to have a global option for all files, or
> if "(set-face-attribute ... :strike-through" should accept more choices
> to make each line configurable. (I personally would change all of my
> striking lines, if possible.)
>
> Thanks in advance for considering this, for some uses (in particular
> org-mode) this would help a lot.

This is what I see on current master:

[-- Attachment #2: Type: text/plain, Size: 49 bytes --]


This is the same text rendered in LibreOffice:


[-- Attachment #3: bug-33648-emacs-orig.png --]
[-- Type: image/png, Size: 3871 bytes --]

[-- Attachment #4: bug-33648-libreoffice.png --]
[-- Type: image/png, Size: 4903 bytes --]

[-- Attachment #5: Type: text/plain, Size: 637 bytes --]


I see two problems here:

1. The strike-through in our case is placed at half the height of an
   upper-case glyph.  LibreOffice seems to have a smarter positioning.

2. The strike-through line is hard coded to 1 pixel in our case, whereas
   LibreOffice scales it up as the font size increases.

I'm have no idea how to fix it; the font code is pretty opaque to me.

I mucked around a bit and came up with the below diff.  It is probably
not the correct fix, but does make it slightly better for me.  Someone
who understands this better would need to investigate this properly.

FWIW, here is what it looks like with the below patch:


[-- Attachment #6: bug-33648-with-patch.png --]
[-- Type: image/png, Size: 2288 bytes --]

[-- Attachment #7: Type: text/plain, Size: 513 bytes --]


diff --git a/src/xterm.c b/src/xterm.c
index 2e0407aff4..5792574e05 100644
--- a/src/xterm.c
+++ b/src/xterm.c
@@ -3932,7 +3932,7 @@ x_draw_glyph_string (struct glyph_string *s)
 	  /* Strike-through width and offset from the glyph string's
 	     top edge.  */
           unsigned long h = 1;
-          unsigned long dy = (glyph_height - h) / 2;
+          unsigned long dy = (glyph_height - h) / 3;
 
 	  if (s->face->strike_through_color_defaulted_p)
 	    x_fill_rectangle (s->f, s->gc, s->x, glyph_y + dy,