[Insight-users] Re: [Cmake] CMake and Borland linker

William A. Hoffman bill.hoffman@kitware.com
Tue, 08 Oct 2002 09:42:44 -0400


I would not use the cvs cmake for building ITK right now.
cmake cvs in unstable right now as many parts are being re-worked.

However, I think the change was made in ITK and cmake.   Until recently there
were no cmakelist files in the gcc-295 directory, and the new cmake 
organization
may check more directory paths for -.   We will fix this before the next cmake
release, but for now try the current release of cmake.

-Bill

At 10:14 AM 10/8/2002 +0100, John Biddiscombe wrote:
> > I'll try building itk with BCB6 when I get home next week
>
>As usual there are shed loads of errors. trivial stuff like not finding 
>atof()
>
>diff -r1.6 metaUtils.cxx
>0a1
> > #include <stdlib.h>
>
>complete failure of complex - use _USE_OLD_RW_STL or whatever it is fixes 
>this (but is a bad solution),
>
>trouble with vcl_cmath
>diff -r1.2 vcl_cmath.h
>6a7
> > #if (__BORLANDC__<0x0561)
>9a11
> > #endif
>13c15
>< #define vcl_abs ::abs
>---
> > #define vcl_abs std::abs
>
>but the best one...
>
>CMake Error: The Borland command line tools do not support path names that 
>have
>- in them.  Please re-name your output directory and use _ instead of -.
>Your path currently is: D:/cmakebuild/Insight/Code/Numerics/vxl/vcl/gcc-295
>Fatal: Unable to execute command: D:\cmakebuild\cmake\Source\Debug\cmake.exe
>
>But this is strange because yesterday it compiled numerics fine and died 
>on the FEM stuff with some error about not returning functions as 
>params...(Aha...of course, I updated cmake...)
>
>hmmm
>
>JB
>
>
>
>
>
>_______________________________________________
>Cmake mailing list
>Cmake@public.kitware.com
>http://public.kitware.com/mailman/listinfo/cmake