Re: [ft] Rendering a filled contour inside an unfilled one

[Werner LEMBERG]

> If the filled/unfilled nature is determined by the contour
> orientation, then how is it valid to provide the fill-rules, since
> they would subvert the determination of the filling by the
> orientation?

This is a good question, and I must admit that I don't understand the
code.  Apparently, the filling rules are connected.  Have a look at
the function `gray_hline' in src/smooth/ftgrays.c, which adjusts the
`coverage' dependent on the fill rule.

> I seem to have read that the FF code isn't robust in case of
> tangencies etc -- is that wrong?  I'm personally looking at a more
> formalized polished solution based on:
> http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Boolean_set_operations_2/Chapter_main.html#Subsection_19.4.3

Using polygons only, tangents aren't an issue, AFAIK, and if you use
the FontForge's `--enable-real=double' configure option (which is now
the default).  The problem is rather grazing intersections of third
order Bézier curves which are very hard to tackle.  Have a look at
this web page:

  http://www.truetex.com/bezint.htm


    Werner