Re: cvs autoconf testsuite failure under solaris 2.5.1

[Akim Demaille]

>>>>> "Nicolas" == Nicolas Joly <address@hidden> writes:

Nicolas> I noticed some testsuite failure for CVS autoconf under
Nicolas> solaris-2.5.1; it seems that most of them have the same
Nicolas> problem :

Jim> This system has a relatively recent kernel (2.2.17), but the libc
Jim> and associated header files come from RH5.2, and it doesn't have
Jim> inttypes.h.  So, `make check' gives this:

Thanks, I'm checking this in.



Index: ChangeLog
from  Akim Demaille  <address@hidden>
        * tests/atspecific.m4 (AT_CHECK_DEFINES): Discar
        STDLIB|INTTYPES|MEMORY|STRING|UNISTD.
        * tests/semantics.at (AC_CHECK_SIZEOF): Don't check the presence
        of default headers, as it's machine dependent.
        Reported by Jim Meyering and Nicolas Joly.

Index: tests/atspecific.m4
===================================================================
RCS file: /cvs/autoconf/tests/atspecific.m4,v
retrieving revision 1.37
diff -u -u -r1.37 atspecific.m4
--- tests/atspecific.m4 2001/01/17 15:26:43 1.37
+++ tests/atspecific.m4 2001/01/18 09:58:34
@@ -134,4 +134,5 @@
 # but those of automatically checked features (STDC_HEADERS etc.).
 # AT_CHECK_HEADER is a better name, but too close from AC_CHECK_HEADER.
 m4_define([AT_CHECK_DEFINES],
-[AT_CHECK([[fgrep '#' config.h | grep -v 'STDC_HEADERS']],, [$1])])
+[AT_CHECK([[fgrep '#' config.h |
+   egrep -v 'STDC_HEADERS|STDLIB|INTTYPES|MEMORY|STRING|UNISTD']],, [$1])])
Index: tests/semantics.at
===================================================================
RCS file: /cvs/autoconf/tests/semantics.at,v
retrieving revision 1.17
diff -u -u -r1.17 semantics.at
--- tests/semantics.at 2001/01/18 09:30:10 1.17
+++ tests/semantics.at 2001/01/18 09:58:34
@@ -103,13 +103,7 @@
 typedef unsigned char ucharchar[2];]])
 AC_CHECK_SIZEOF(ucharcharchar)]],
 [AT_CHECK_DEFINES(
-[#define HAVE_INTTYPES_H 1
-#define HAVE_MEMORY_H 1
-#define HAVE_STDLIB_H 1
-#define HAVE_STRINGS_H 1
-#define HAVE_STRING_H 1
-#define HAVE_UNISTD_H 1
-#define SIZEOF_CHAR 1
+[#define SIZEOF_CHAR 1
 #define SIZEOF_CHARCHAR 2
 #define SIZEOF_CHARCHARCHAR 0
 #define SIZEOF_UCHARCHAR 2





Nicolas, I must say I don't understand too well your trailing file
problem, but it might be related to a bug in the test suite framework
I recently realized: if you use -d it does not remove the trailing
files, which is what you want for the *last* check, but certainly not
for the others.  Maybe it's related to this.  Still, does the problem
still show up?

Thanks!

        Akim