http://bugs.winehq.org/show_bug.cgi?id=33378
--- Comment #19 from Rafał Mużyło galtgendo@o2.pl 2013-08-04 09:50:28 CDT --- FFS, this is getting off track fast.
My point is: awhile ago, a regression has been made it's been bisected to a particular commit a minimal revert has been proposed the revert has been later refined (here and on wine-devel) into a hypothesis of the origin of the problem based on the hypothesis an even smaller workaround has been offered (and sent to wine-patches)
yet the party that caused the regression still complains, not offering a better solution (and not offering any real help).
Well, 1.6 freeze muddled the situation a bit too, but that's a completely different matter.
https://mail.gnome.org/archives/gtk-devel-list/2013-April/msg00131.html (the above was between two gtk+ *devs*, but some of the sentiment holds)
I'm not saying original solution (pre-regression) wasn't a hack, cause it sort of was, just that the code was changed without taking into account why that hack was necessary.
If Microsoft pointed everybody to C++ classes, Point must have a very strict definition in terms of RectF and Width=Height=0 seems a natural one.