[Insight-users] Building ITK with Borland C++ broken?

[Bill Hoffman]

For a free compiler on win32, you can use cygwin gcc which is supported by ITK.

-Bill

At 02:44 PM 5/8/2002 +0100, John Biddiscombe wrote:


>> ITK does not currently support the borland compiler.   However,
>> we have only tried 5.5, 5.6 might work, but I doubt it.
>> However, if you want to try, you should generate borland makefiles with
>cmake, and
>> not nmake makefiles.
>
>No. Not 5.6. I tried it the other day. It fails in (pauses whilst running
>make again)...
>
>Borland C++ 5.6 for Win32 Copyright (c) 1993, 2002 Borland
>D:\Insight\Code\Numerics\itkVersorTransformOptimizer.cxx:
>Error E2315 D:\Insight\Code\Numerics\vxl\vnl/vnl_complex.h 35: 'real' is not
>a m
>ember of 'std::complex<float>', because the type is not yet defined in
>function
>vnl_math_isnan(const std::complex<float> &)
>Error E2315 D:\Insight\Code\Numerics\vxl\vnl/vnl_complex.h 35: 'imag' is not
>a m
>ember of 'std::complex<float>', because the type is not yet defined in
>function
>vnl_math_isnan(const std::complex<float> &)
>etc etc...lots more of the same...
>
>I had a quick look at the vnl_complex file and saw several #ifdefs for each
>compiler, added a Borland clause (to the win32 option), but it didn't work.
>Other than that I've not looked at it recently. However, I've been following
>some of the edge detection threads in this list and want very much to play
>with ITK so I'd be glad for help getting Borland working. (Cheap copy of
>MSVC anyone?)
>
>JB
>
>
>
>_______________________________________________
>Insight-users mailing list
>Insight-users@public.kitware.com
>http://public.kitware.com/mailman/listinfo/insight-users