RE: [AD] Header splitup

[ Thread Index | Date Index | More lists.liballeg.org/allegro-developers Archives ]



On Mon, 5 Nov 2001, Vincent Penquerc'h wrote:

> > can you mail both the complete and 'incomplete' diff and the
> > tarball? i
>
> I'll send them to you just now.
>
> > >- fixsqrt -> fsqrt, it was not compiling in C++ mode (might
> > not work in
> > >  Irix, as I believe this was done for Irix?)
> >
> > IIRC this too? or does this change revert a recent name-clash fix?
>
> I think this change was done because fsqrt was defined by SGI headers,
> but it did not compile on my Linux box. Maybe the fix was only half
> committed ? Don't know, but I had to fix that to get it to compile.

Code below is from include/allegro/alucfg.h. It was done by Michael Bukin
and later extended by me with ALLEGRO_INCLUDE_MATH_H kludge also done by
me. It was invented because of SGI IRIX.

 38 #ifdef __sgi__
 39 # define fhypot sgi_fhypot
 40 #endif
 41 #include <math.h>
 42 #ifdef __sgi__
 43 /* SGI defines these macros in <math.h>.  */
 44 # undef fsin
 45 # undef fcos
 46 # undef ftan
 47 # undef fasin
 48 # undef facos
 49 # undef fatan
 50 # undef fatan2
 51 # undef fsqrt
 52 # undef fhypot
 53 # undef fceil
 54 #endif
 55 #undef ALLEGRO_INCLUDE_MATH_H

The flow of the inclusion is: allegro.h includes allegro/alconfig.h which
includes alucfg.h which includes math.h and undefs all the bad defines or
functions. Rest of the allegro is processed after this magic. There is no
fixsqrt I know of.

Have a nice day.

Stepan Roh



Mail converted by MHonArc 2.6.19+ http://listengine.tuxfamily.org/