> Hi Joost:
> Fixed as of revision 12475. Thanks very much for spotting this.
> One peculiar issue remains. (For background see
> http://en.wikipedia.org/wiki/Even–odd_rule and
> http://en.wikipedia.org/wiki/Nonzero-rule.)
> software@raven> examples/c/x27c -dev wxwidgets -eofill -drvopt backend=0
> uses the even-odd filling rule as expected. But if I drop the -eofill
> option, that example still shows the even-odd filling rule is being
> used for the basic wxwidgets backend rather than the correct nonzero
> fill rule.
> In contrast,
> software@raven> examples/c/x27c -dev wxwidgets -eofill -drvopt backend=2
> software@raven> examples/c/x27c -dev wxwidgets -drvopt backend=2
> shows the very different even-odd fill versus nonzero fill results (as expected) for the wxGC
> wxwidgets backend.
> Does the even-odd fill rule work properly for you with the basic
> wxwidgets backend or do you also confirm that the even-odd fill rule
> is being used for the basic backend regardless?
> Alan
> __________________________

When I execute those examples, I get them both to plot with the different rules, no matter if backend=0 or backend=2. I did find out though, that the internal wxDC::DrawPolygon function has WXUNUSED_IN_WINCE(fillstyle) defined. Thus, if you're having a system where WXUNUSED_IN_WINCE is defined as WXUNUSED (for WinCE builds: #ifdef __WXWINCE__), then the fillstyle is completely ignored (thus I think it uses the only fillstyle available in WinCE).

Did you mean wxDC instead of wxGC? Because in the wxwidgets source I can't find any wxGC, only wxDC (the only thing I find is :"#define wxGCDC wxDC")