Re: could not setup config headers machinery

[Dr. David Kirkby]

On 10/14/10 09:02 PM, Ralf Wildenhues wrote:
> Hello David,
>
> * Dr. David Kirkby wrote on Tue, Oct 12, 2010 at 09:53:46PM CEST:
> > I'm trying to sort out why the latest version of the GNU Scientific Library 
> > (GSL)
>
> > does not build properly on some AIX systems (the developers are aware of it).
> >
> > I need to check if 'fprnd_t' is defined in float.h or not.
> >
> > Is there any particular trick needed to check for things in float.h?
>
> > The *only* change I've made is this single line in configure.ac
> >
> > AC_CHECK_DECLS([fprnd_t],[],[],[[#include<float.h>]])
> >
> > But after running will always cause the generated configure script
> > to fail to execute properly - see below. Notice it does not find
> > 'fprnd_t', (on the second line) but then later generates an error.
> [...]
>
> Can you send 'diff -u' of the two configure files (without above line
> and with)?  Of even of the one with float.h and the one with math.h?


Sure. First here's the diff between the configure.ac files

address@hidden:~/gsl-1.14$ diff -u configure.ac.orig configure.ac.float
--- configure.ac.orig   Wed Mar 10 10:57:12 2010
+++ configure.ac.float  Sat Oct 23 16:54:20 2010
@@ -197,6 +197,7 @@
 AC_CHECK_DECLS(atanh,,,[#include <math.h>])
 AC_CHECK_DECLS(ldexp,,,[#include <math.h>])
 AC_CHECK_DECLS(frexp,,,[#include <math.h>])
+AC_CHECK_DECLS([fprnd_t],[],[],[[#include <float.h>]])
 AC_CHECK_DECLS(isinf,,,[#include <math.h>])
 AC_CHECK_DECLS(isfinite,,,[#include <math.h>])
 AC_CHECK_DECLS(finite,,,[#include <math.h>

And here's the resulting diff in the generated configure scripts.



address@hidden:~/gsl-1.14$ diff -u configure.orig configure.float
--- configure.orig      Sat Oct 23 16:50:30 2010
+++ configure.float     Sat Oct 23 16:55:16 2010
@@ -11933,6 +11933,18 @@
 #define HAVE_DECL_FREXP $ac_have_decl
 _ACEOF

+ac_fn_c_check_decl "$LINENO" "fprnd_t" "ac_cv_have_decl_fprnd_t" "#include 
<float.h>
+"
+if test "x$ac_cv_have_decl_fprnd_t" = x""yes; then :
+  ac_have_decl=1
+else
+  ac_have_decl=0
+fi
+
+cat >>confdefs.h <<_ACEOF
+#define HAVE_DECL_FPRND_T $ac_have_decl
+_ACEOF
+
 ac_fn_c_check_decl "$LINENO" "isinf" "ac_cv_have_decl_isinf" "#include <math.h>
 "
 if test "x$ac_cv_have_decl_isinf" = x""yes; then :


> > Conversely, if I change it to
> >
> > AC_CHECK_DECLS([fprnd_t],[],[],[[#include<math.h>]])
> >
> > then the configure script works. But checking in math.h is not a lot
> > of use when I want to check in float.h
>
> If that line really is the cause of the issues you're seeing (rather
> than some other circumstance), then the diff should provide a strong
> hint.

What is also interesting is the difference between that which looks in the 
math.h file and that which looks in the float.h file:

i.e.

address@hidden:~/gsl-1.14$ diff -u configure.ac.math configure.ac.float
--- configure.ac.math   Sat Oct 23 16:55:44 2010
+++ configure.ac.float  Sat Oct 23 16:54:20 2010
@@ -197,7 +197,7 @@
 AC_CHECK_DECLS(atanh,,,[#include <math.h>])
 AC_CHECK_DECLS(ldexp,,,[#include <math.h>])
 AC_CHECK_DECLS(frexp,,,[#include <math.h>])
-AC_CHECK_DECLS([fprnd_t],[],[],[[#include <math.h>]])
+AC_CHECK_DECLS([fprnd_t],[],[],[[#include <float.h>]])
 AC_CHECK_DECLS(isinf,,,[#include <math.h>])
 AC_CHECK_DECLS(isfinite,,,[#include <math.h>])
 AC_CHECK_DECLS(finite,,,[#include <math.h>
address@hidden:~/gsl-1.14$


gives the expected change in the output file.

address@hidden:~/gsl-1.14$ diff -u configure.math configure.float
--- configure.math      Sat Oct 23 16:55:58 2010
+++ configure.float     Sat Oct 23 16:55:16 2010
@@ -11933,7 +11933,7 @@
 #define HAVE_DECL_FREXP $ac_have_decl
 _ACEOF

-ac_fn_c_check_decl "$LINENO" "fprnd_t" "ac_cv_have_decl_fprnd_t" "#include 
<math.h>
+ac_fn_c_check_decl "$LINENO" "fprnd_t" "ac_cv_have_decl_fprnd_t" "#include 
<float.h>
 "
 if test "x$ac_cv_have_decl_fprnd_t" = x""yes; then :
   ac_have_decl=1


Could this be a gcc bug, which is actually causing the configure script to fail?

The reason I'm thinking this is because it seems to be a gcc bug that the 
fprnd_t is not being defined on the AIX system. If fprnd_t  was defineded 
properly, I would not need to be testing for this at all. It seems a bug thats 
it not defined, since it is clearly defined in /usr/include/float.h


-bash-4.1$ grep fprnd_t  /usr/include/float.h
typedef unsigned short fprnd_t;
fprnd_t fp_read_rnd();
fprnd_t fp_swap_rnd();
fprnd_t fp_read_rnd(void);
fprnd_t fp_swap_rnd(fprnd_t rnd);


but it does not appear in the header file gcc is using.

-bash-4.1$ find /opt/pware -name float.h
/opt/pware/include/c++/4.2.4/tr1/float.h
/opt/pware/lib/gcc/powerpc-ibm-aix5.3.0.0/4.2.4/include/float.h
/opt/pware/lib/gcc/powerpc-ibm-aix5.3.0.0/4.2.4/install-tools/include/float.h
-bash-4.1$ grep fprnd_t  /opt/pware/include/c++/4.2.4/tr1/float.h
-bash-4.1$ grep fprnd_t 
/opt/pware/lib/gcc/powerpc-ibm-aix5.3.0.0/4.2.4/include/float.h
-bash-4.1$  grep fprnd_t 
/opt/pware/lib/gcc/powerpc-ibm-aix5.3.0.0/4.2.4/install-tools/include/float.h
-bash-4.1$

It looks to me like this is a gcc bug, but I'm puzzled why it can result in a 
problem with the autoconf generated script.

> Thanks,
> Ralf

Dave