gdi32: Improve detection of symbol charset for old truetype fonts.

[Huw Davies]

On Wed, Nov 18, 2015 at 05:53:20PM +0800, Dmitry Timoshkov wrote:
> Huw Davies <huw at codeweavers.com> wrote:
> 
> > On Wed, Nov 18, 2015 at 02:16:36PM +0800, Dmitry Timoshkov wrote:
> > > This patch uses same check for symbol charset that get_outline_text_metrics()
> > > already does, and allows the font attached to the bug 33117 be correctly
> > > recognized as a symbol ttf (like Windows does).
> > 
> > I think this needs to wait until after 1.8.  It's probably correct,
> > but there's enough doubt to not risk it (and yes, I know it's tricky
> > to write a convincing test).  My concern is that although it works for
> > the font mentioned in the bug, there may well be other fonts out there
> > that are affected in undesirable ways.
> 
> If even so simple one-liners are considered risky, I can't think of any
> reasonable fix that could be included without a risk to break something
> during code freeze, not mentioning the regressions that sit in the bugzilla
> for ages without anyone even looking at them. Probably it's not worth spending
> time on trying to fix something for 1.8 with such an attitude.

Not at all.  The point is that this particular one-liner is most
certainly not simple, for reasons that I tried to outline.

Huw.