CGAL users discussion list

Hi David,

You have encountered a real bug that hasn't been reported yet. I was  
able to reproduce it. Thanks for the simple test case!
We'll look into it, and we'll let you know. Stay tuned...

Quoting David Keller 
<>:

> Hi,
>
> sorry for insisting on similar topics here: I came across a failure
> caused by overlaying two arrangements with partially overlapping curves.
>
> You find a test scenario containing two test cases attached to this
> mail. Both test cases fail by throwing an exception, actually a
> precondition failure in CGAL. The program gives the following output:
>
> ====== BEGIN OUTPUT ======
> Running 2 test cases...
> CGAL error: assertion violation!
> Expr: !pt.is_red_object_null() || !pt.is_blue_object_null()
> File:
> ../../../../libs/external/cgal/CGAL-3.3.1/include/CGAL/Arr_overlay_2/Overlay_visitor.h
> Line: 594
> Explanation:
> arr_overlay_with_orverlapping_curves.cpp(59): error in
> "simple_overlay_test": exception thrown by CGAL::overlay( arr1, arr2,
> ovl, ot )
> CGAL error: assertion violation!
> Expr: !pt.is_blue_object_null()
> File:
> ../../../../libs/external/cgal/CGAL-3.3.1/include/CGAL/Arr_overlay_2/Overlay_visitor.h
> Line: 271
> Explanation:
> arr_overlay_with_orverlapping_curves.cpp(83): error in "overlay_test":
> exception thrown by CGAL::overlay( arr1, arr2, ovl, ot )
>
> *** 2 failures detected in test suite "Master Test Suite"
>
> EXIT STATUS: 201
> ====== END OUTPUT ======
>
> When trying to debug these I found Overlay_visitor being called with
> event points which miss a correct point_info, i.e. their relations to
> objects in the "blue" and "red" arrangement is not set.
> Overlay_meta_traits actually gets the correct information on
> overlapping subcurves and associates these curves with "purple"
> information. However, the end points are apparently not correctly
> flagged with "purple" information and thus violate the preconditions in
> Overlay_visitor.
>
> I have actually no idea how to further debug these things, at least
> from my point of view the code in Overlay_visitor and
> Overlay_meta_traits is not easily readable. Hence, I would be glad to
> get further hints on where to look at or even on how to fix this issue.
> Maybe the problem is even known and there is a fix for it?
>
> Thanks and regards
> David Keller
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss