Re: (try 2) gdi32: if a font substitution is given then only match tothat substitution not to the original family name.

[Dmitry Timoshkov]

"Aric Stewart" <aric(a)codeweavers.com> wrote:
> diff --git a/dlls/gdi32/freetype.c b/dlls/gdi32/freetype.c
> index 145be9b..c427694 100644
> --- a/dlls/gdi32/freetype.c
> +++ b/dlls/gdi32/freetype.c
> @@ -3408,8 +3408,8 @@ GdiFont *WineEngCreateFontInstance(DC *dc, HFONT hfont)
>  */
>         LIST_FOR_EACH(family_elem_ptr, &font_list) {
>             family = LIST_ENTRY(family_elem_ptr, Family, entry);
> -            if (!strcmpiW(family->FamilyName, FaceName) ||
> -                (psub && !strcmpiW(family->FamilyName, psub->to.name)))
> +            if (( psub && !strcmpiW(family->FamilyName, psub->to.name)) ||
> +                (!psub && !strcmpiW(family->FamilyName, FaceName)))
>             {
>                 LIST_FOR_EACH(face_elem_ptr, &family->faces) { 
>                     face = LIST_ENTRY(face_elem_ptr, Face, entry);

After loking at this change once more I think it's not correct. It will
lead to failures to find a font if there is a substitution but a face
specified by substitution doesn't exist. I'd suggest to test what Windows
does in that case, and if Windows allows to substitute existing faces
at all. Do you have a test cases which passes with your patch but doesn't
without it?

-- 
Dmitry.

(try 2) gdi32: if a font substitution is given then only match tothat substitution not to the original family name.

[Aric Stewart]

You are mostly correct. It looks like windows totally ignores the font 
substitute if it finds the correct font. So this patch is incorrect.

Is there a good mechanic to get wine to force a substitution for a 
broken font? If the user has FontX  however their FontX does not render 
correctly with freetype so we want to replace it with FontY...

-aric

Dmitry Timoshkov wrote:
> "Aric Stewart" <aric(a)codeweavers.com> wrote:
> > diff --git a/dlls/gdi32/freetype.c b/dlls/gdi32/freetype.c
> > index 145be9b..c427694 100644
> > --- a/dlls/gdi32/freetype.c
> > +++ b/dlls/gdi32/freetype.c
> > @@ -3408,8 +3408,8 @@ GdiFont *WineEngCreateFontInstance(DC *dc, HFONT 
> > hfont)
> >  */
> >         LIST_FOR_EACH(family_elem_ptr, &font_list) {
> >             family = LIST_ENTRY(family_elem_ptr, Family, entry);
> > -            if (!strcmpiW(family->FamilyName, FaceName) ||
> > -                (psub && !strcmpiW(family->FamilyName, psub->to.name)))
> > +            if (( psub && !strcmpiW(family->FamilyName, 
> > psub->to.name)) ||
> > +                (!psub && !strcmpiW(family->FamilyName, FaceName)))
> >             {
> >                 LIST_FOR_EACH(face_elem_ptr, &family->faces) { 
> >                     face = LIST_ENTRY(face_elem_ptr, Face, entry);
>
> After loking at this change once more I think it's not correct. It will
> lead to failures to find a font if there is a substitution but a face
> specified by substitution doesn't exist. I'd suggest to test what Windows
> does in that case, and if Windows allows to substitute existing faces
> at all. Do you have a test cases which passes with your patch but doesn't
> without it?

(try 2) gdi32: if a font substitution is given then only match tothat substitution not to the original family name.

[Dmitry Timoshkov]

"Aric Stewart" <aric(a)codeweavers.com> wrote:
> You are mostly correct. It looks like windows totally ignores the font 
> substitute if it finds the correct font. So this patch is incorrect.
>
> Is there a good mechanic to get wine to force a substitution for a 
> broken font? If the user has FontX  however their FontX does not render 
> correctly with freetype so we want to replace it with FontY...

Does font render correctly in other programs beside Wine? If no, there is
no point IMO to add any workarounds, just ask the user to move the font out
of the way. Is there a way to detect if the font is broken? What if either
freetype or the font get fixed later?

-- 
Dmitry.