http://quantlib.414.s1.nabble.com/calendar-problem-when-calling-quantlib-lib-from-java-tp5595p5604.html
Yes, thanks for your help about the cpp file ( ld.so.conf not well configured ).
The reason why we are doing this way with JNI is mainly historical: the program
has been done 5 years ago... we wanted to update the QL version in a smooth way.
Of course, if we cannot solve this problem, we'll use the SWIG java interfaces .
>
> Pierre,
>
> The problem is almost certainly your installation/environment. I was able
> to build and run the c++ test case you posted on feb 8th without issue (and
> had sent you an email affirming as much on the 11th). I suggest you rebuild
> your environment and verify that you can run the provided examples. I'm
> pretty sure that once you have a clean environment you will see that your
> test case works fine.
>
> As for the java aspect of it, I'm not quite clear on why you'd write your own
> jni interfaces when the provided swig interfaces work well. You might want
> to try them when you have a clean install so you can determine if you want to
> proceed with the roll-your-own jni approach or use the swig interfaces going
> forward.
>
> Regards,
>
> Tito.
>
>
> moloko <
[hidden email]> wrote:
> All,
>
> I've isolated my code from my system and I can send sources files to
> generate the crash.
> If anybody could have a look at it, please tell me, and I will send the
> files.
>
> to remember the problem :
> I'm building a shared lib which is a mix of JMI / QuantLib
> I call this lib in a java process
> I've got a crash when I call
> Calendar calendar = TARGET();
> but not
> Calendar calendar = NullCalendar();
> or
> DayCounter dayCounter = Thirty360(Thirty360::European);
> but not
> DayCounter dayCounter = Actual360();
>
> I have also made a post about building a "pure" QuantLib shared library
>
> Regards,
>
> Pierre
> --
> View this message in context:
>
Defy all challenges. Microsoft(R) Visual Studio 2008.