CGAL users discussion list

[thann <>]

Hi, 

I would like to use CGAL for my Master's Thesis dealing with geometric
computation. For this, I want to setup a Delauney Triangulation of a set of
points in 3D space. 

Unfortunately, I run into a weird problem when constructing the DT using the
EPICK kernel. At some point, a CGAL::Uncertain_conversion_exception is
thrown. Now I understand that this is used for guiding the internal control
flow of CGAL and the user is not supposed to be concerned with those.
However, unlike the FAQ suggests, the exception is NOT caught by the correct
catch handler in Filtered_predicate.h:166, nor by a sourrounding catch(…) in
my calling main() method. Instead, the exception remains uncaught, resulting
in a call to terminate():

>terminate called after throwing an instance of
'CGAL::Uncertain_conversion_exception'
>  what():  Undecidable conversion of CGAL::Uncertain<T>

I use the latest CGAL 4.1 compiled with CGG 4.7.2, using Boost 1.5.2, GMP
5.1.0, MPFR 3.1.1 on a MacBook Pro 5,4 running OS-X 10.8.2. 

The CGAL examples (those that I tried) all worked successfully with the same
build configuration.

A small code example, the make call, a full gdb stack trace and the CMake
build log for CGAL can be found in the attachments:

cgalmain.cc - the code example
<http://cgal-discuss.949826.n4.nabble.com/file/n4656528/cgalmain.cc>  
cgal_cmake_log.txt
<http://cgal-discuss.949826.n4.nabble.com/file/n4656528/cgal_cmake_log.txt>  
stacktrace.txt
<http://cgal-discuss.949826.n4.nabble.com/file/n4656528/stacktrace.txt>  
make_call.txt
<http://cgal-discuss.949826.n4.nabble.com/file/n4656528/make_call.txt>  

I would appreciate any input on this matter. 

Thanks, 
Tobias Hann



--
View this message in context: 
http://cgal-discuss.949826.n4.nabble.com/Uncaught-CGAL-Uncertain-conversion-exception-tp4656528.html
Sent from the cgal-discuss mailing list archive at Nabble.com.