DPS font backend uses the wrong font

DPS font backend uses the wrong font

[Po Lu]

As of 2028df78, font backends are required to use s->font instead of
s->face->font, even in the NS port, because s->face is now set to mouse
face by redisplay independently of terminal specific code.

Every font backend except for nsfont already does that, so it would make
sense to change nsfont to do it correctly as well.  But is there
something I'm missing here?

IOW, I'm suggesting to change the following snippet of nsfont_draw:

  font = (struct nsfont_info *)s->face->font;
  if (font == NULL)
    font = (struct nsfont_info *)FRAME_FONT (s->f);

to

  font = (struct nsfont_info *)s->font;
  if (font == NULL)
    font = (struct nsfont_info *)FRAME_FONT (s->f);

Thanks.

Re: DPS font backend uses the wrong font

[Po Lu]

Po Lu <luangruo@yahoo.com> writes:

> As of 2028df78, font backends are required to use s->font instead of
> s->face->font, even in the NS port, because s->face is now set to mouse
> face by redisplay independently of terminal specific code.

BTW, this also means a great deal of code in the NS port that looks
like:

      if (s->hl == DRAW_MOUSE_FACE)
        {
          face = FACE_FROM_ID_OR_NULL (s->f,
                                       MOUSE_HL_INFO (s->f)->mouse_face_face_id);
          if (!face)
            face = FACE_FROM_ID (s->f, MOUSE_FACE_ID);
        }
      else
        face = FACE_FROM_ID (s->f, s->first_glyph->face_id);

Can be removed now, because it's no longer necessary for
terminal-specific code to worry about the mouse face like this.

Re: DPS font backend uses the wrong font

[Alan Third]

On Sun, Oct 17, 2021 at 09:20:43PM +0800, Po Lu wrote:
> 
> As of 2028df78, font backends are required to use s->font instead of
> s->face->font, even in the NS port, because s->face is now set to mouse
> face by redisplay independently of terminal specific code.
> 
> Every font backend except for nsfont already does that, so it would make
> sense to change nsfont to do it correctly as well.  But is there
> something I'm missing here?
> 
> IOW, I'm suggesting to change the following snippet of nsfont_draw:
> 
>   font = (struct nsfont_info *)s->face->font;
>   if (font == NULL)
>     font = (struct nsfont_info *)FRAME_FONT (s->f);
> 
> to
> 
>   font = (struct nsfont_info *)s->font;
>   if (font == NULL)
>     font = (struct nsfont_info *)FRAME_FONT (s->f);

Feel free to do whatever is required in nsfont.m, it's been basically
unmaintained for a decade or so afaict. My changes were simply to fix
crashes and remove macOS code that's no longer used (macOS builds use
macfont.m).

We really need someone to update nsfont to actually support things
like bold and italics and ideally use harfbuzz, but I've really
struggled with it, and I don't think anyone actually uses GNUstep
Emacs.
-- 
Alan Third

Re: DPS font backend uses the wrong font

[Alan Third]

On Sun, Oct 17, 2021 at 09:35:21PM +0800, Po Lu wrote:
> Po Lu <luangruo@yahoo.com> writes:
> 
> > As of 2028df78, font backends are required to use s->font instead of
> > s->face->font, even in the NS port, because s->face is now set to mouse
> > face by redisplay independently of terminal specific code.
> 
> BTW, this also means a great deal of code in the NS port that looks
> like:
> 
>       if (s->hl == DRAW_MOUSE_FACE)
>         {
>           face = FACE_FROM_ID_OR_NULL (s->f,
>                                        MOUSE_HL_INFO (s->f)->mouse_face_face_id);
>           if (!face)
>             face = FACE_FROM_ID (s->f, MOUSE_FACE_ID);
>         }
>       else
>         face = FACE_FROM_ID (s->f, s->first_glyph->face_id);
> 
> Can be removed now, because it's no longer necessary for
> terminal-specific code to worry about the mouse face like this.

Hmm, I wonder if this means mouse face will be honoured in SVG images
now...?

That would be nice because I couldn't work out how to find the mouse
face where we construct the image struct.
-- 
Alan Third

Re: DPS font backend uses the wrong font

[Po Lu]

Alan Third <alan@idiocy.org> writes:

> Hmm, I wonder if this means mouse face will be honoured in SVG images
> now...?

No, because it's applied in fill_XXX_glyph_string, which is after any
work image.c does, almost right before the glyph string is actually
drawn.