Re: gdiplus: fix font test when liberation fonts are installed

[Dmitry Timoshkov]

Austin English  wrote:
> - ok(!lstrcmp(lf.lfFaceName, "Arial") ||
> -
>   !lstrcmp(lf.lfFaceName, "Liberation Sans"), "wrong face name %s\n", lf.lfFaceName);

The tests are supposed to reflect Windows behaviour, what Windows version
does return "Liberation Sans" in this test?

gdiplus: fix font test when liberation fonts are installed

[Austin English]

On Wed, Oct 17, 2012 at 7:54 PM, Dmitry Timoshkov  wrote:
> Austin English  wrote:
> > - ok(!lstrcmp(lf.lfFaceName, "Arial") ||
> > -
> >   !lstrcmp(lf.lfFaceName, "Liberation Sans"), "wrong face name %s\n", lf.lfFaceName);
>
> The tests are supposed to reflect Windows behaviour, what Windows version
> does return "Liberation Sans" in this test?

Obviously none, though it seems to a reasonable adjustment, since the
Liberation team can't use the Arial name. We already do something
similar in dlls/gdi32/freetype.c for liberation mono/sans/serif fonts.

gdiplus: fix font test when liberation fonts are installed

[Dmitry Timoshkov]

Austin English  wrote:
> > > - ok(!lstrcmp(lf.lfFaceName, "Arial") ||
> > > -
> > >   !lstrcmp(lf.lfFaceName, "Liberation Sans"), "wrong face name %s\n", lf.lfFaceName);
> >
> > The tests are supposed to reflect Windows behaviour, what Windows version
> > does return "Liberation Sans" in this test?
>
> Obviously none, though it seems to a reasonable adjustment, since the
> Liberation team can't use the Arial name. We already do something
> similar in dlls/gdi32/freetype.c for liberation mono/sans/serif fonts.

Just consider a failing test as a sign that something is broken in
the tested invironment, and some real world application will fail as
well (otherwise the test doesn't make sense). The environment should be
fixed instead of adjusting the test.