Hi Gustavo, On Wed, 4 Apr 2001, Gustavo Guerra wrote:
I've used GiNaC for some time with g++, but now I need to use it in Borland C++. Are there any plans on porting GiNaC to other platforms?
No, there are no plans to port CLN to other compilers. It is not only a question of static initializations, it is also a question of efficiency. Very few C++ compilers can generate good code for the particular programming style in which CLN is written. (A style which you will not find in the books: heavy use of smart pointers with nontrivial destructors.) BTW, why do you "need to use it in Borland C++". The Borland Compiler is rather poor at C++ compared to GCC, IIRC.
It seems pretty straightforward, but the problem is CLN. Is it possible to make GiNaC work on top of other similar library like Shoup's NTL? Can anyone help me do it?
NTL is no help here, it does not have the features we need for GiNaC. Still, you are welcome to give it a try. I had once successfully ported it to the Portland Group compiler and that worked just fine. Of course, you should try it with a static library first, not a dynamic one. Please refer to <http://www.ginac.de/FAQ.html#portability>. Regards -richy. -- Richard Kreckel <Richard.Kreckel@Uni-Mainz.DE> <http://wwwthep.physik.uni-mainz.de/~kreckel/>