CGAL users discussion list

[atsui <>]

This is not a solution but a continuation of investigating this bug.

I've tested the two-curve input on CGAL 4.4 and it fails at a different
point from what you had described. For me, it is trying to perform the sweep
line algorithm (see the _sweep method in Basic_sweep_line_2_impl.h:154) and
dies before finishing. Here's a trace that I generated by compiling with
CGAL_SL_VERBOSE defined: http://pastebin.com/xt0UQWGd

Note that the first curve that gets inserted is split as you said, into
(2.15 1.475 --> ~2.175 ~1.45) and (~2.175 ~1.45 --> 2.175 1.45). So the
expected result of inserting the input curves is 3 monotone curves:

A (2.15 1.475 --> ~2.175 ~1.45)
C (~2.175 ~1.45 --> 2.175 1.45)
B (2.175 1.45 --> 2.15 1.425)

The trouble starts on line 121 of the trace, where when setting up the
events, it tries to intersect B with A and "successfully" finds an
intersection point. In fact, A and B should be disjoint by the x-monotone
decomposition above. So that causes strange behavior downstream.

I actually also had no luck switching the order of the two input curves.



--
View this message in context: 
http://cgal-discuss.949826.n4.nabble.com/Precondition-violation-in-Bezier-x-monoton-2-h-using-Bezier-arrangement-tp4659201p4659495.html
Sent from the cgal-discuss mailing list archive at Nabble.com.