Compilation report

Werner Koch wk at isil.d.shuttle.de
Wed Jan 21 18:32:23 CET 1998


Anand Kumria <wildfire at progsoc.uts.edu.au> writes:

> checking size of unsigned short... 0
> checking size of unsigned int... 0
> checking size of unsigned long... 0

Hmmm, configure should test this with this program

#include <stdio.h>
main()
{
  FILE *f=fopen("conftestval", "w");
  if (!f) exit(1);
  fprintf(f, "%d\n", sizeof($1));
  exit(0);
}

can you please look into config.log to see why this program failed.
(zlib does not appear here).  

> Two other suggestions: maybe you can do the test for zlib latter (it isn't

done.

> a critical library (or is it), and perhaps you could abort the configure
> process if you detect sizes of 0.

Will do so.

> SunOS doesn't have RAND_MAX defined, you should probably check for rand()
> and if found check for RAND_MAX and '#define RAND_MAX 32767' if RAND_MAX
> wasn't found.

Using rand() is just a kludge to let it compile until we have a RNG; but
I can check for it.

> * configure assumes everything is in '.', doesn't support multiple
> architectures. (VPATH) - the attached patch fixes configure, mpi/config.links
> needs some work though.

I'm new to autoconf; I used another make scheme in the past, which worked
quite different.   Hmmm, I don't see a attachment ... 

-- 
Werner Koch, Duesseldorf  -   werner.koch at guug.de   -  PGP keyID: 0C9857A5





More information about the Gnupg-devel mailing list