patch for error handling after driver callback calls.

[Alexandre Julliard]

Dave_Belanger at cimmetry.com writes:

> I understand your idea and it makes perfect sense. Maybe the description of my
> patch wasn't clear enough. And maybe I shouldn't have changed SetTextColor and
> SetBkColor (that were actually the only callbacks that were well implemented on
> both the driver side and the the GDI). But if you look at most of the other
> callback implementations in the EMF and WMF drivers (SetTextAlign for example),
> you'll see that they only return TRUE or FALSE (the return value from
> WriteRecord) and not what the corresponding API function should return. So, the
> code that handles the callback return value doesn't interpret the return code
> correctly.

Actually it's the driver that isn't returning the right thing, AFAICS
the caller is doing the right thing, at least for SetTextAlign.
Though it seems in some cases, like SetTextColor, it expects the
driver to return the new color instead of the old one; that should
probably be fixed too. The driver interface should really follow the
API as closely as possible.

-- 
Alexandre Julliard
julliard at winehq.com