Hi, I'm porting code from Win32 to Linux, a math science app, the thing is that the app have full dependency of de win32 function _fpclass().

Today I was looking all day that function on Linux but it doesn't exist.

I know that there are some functions that can make the job mixing them in a other one like in this link below, more or less.

http://www.macaulayenterprises.net/f...3.0-README.txt

Well the app use very high virtual memory, Gigas of it(around 500), the idea is to use Linux as Calculation Server using a very large swap space.

If we can't find some way to fix this trouble, we can change the Linux for some open *nix like Open Solaris, because those have fp_class_l().

Thank you very much and sorry for my bad english, I'm not native.

From quick WEB search I understand that the function is used to classify long doubles.

First of all, does your Linux box support long doubles ? If it does, can you yourself write the needed function ?

Hi -

Look in "math.h": fpclassify() and friends:

http://www.linuxquestions.org/questi...4/#post3927053

Unfortunately, you will have to do a bit of porting work. Nothing complicated, but a hassle nonetheless...

You may want to have a look at what the wine chaps have done. A quick google led me to this page You might be able to reuse what they have done.

Thank all of you , I think that I will adapt the version of _fpclass() of wine, is the better way for me.