adutoit wrote:
[ a lot of trouble using boost and Intel C++ ]
Since I had no time to "configure" Intel's compiler and since I was so fed up with getting it up and running, I did 2 things with Intel-7.0 on linux which work for me: I use -ansi and -DBOOST_NO_CONFIG. Though the latter might be a bad idea in general (since no special platform-dependent optimization can be used then) this solved a lot of trouble, especially since some parts of boost make (made?) assumptions about Intel's behaviour that no longer hold.
From my own experience, if it compiles with gcc-3.2 -ansi it should never fail with Intel's compiler except when the code is not completely compliant. At present I use gcc only for the much more intelligent error messages for templates.
If You could provide a small code snippet which shows this error I can try to compile it under Linux and see what is happening there. Markus -- Compile time analytic differentiation? Yes, at http://daixtrose.sourceforge.net/