This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Patch: libstdc++-v3/aclocal.m4 for i586-sco-sysv5uw7.1.0
- To: libstdc++ at sources dot redhat dot com
- Subject: Patch: libstdc++-v3/aclocal.m4 for i586-sco-sysv5uw7.1.0
- From: RDBrown at mira dot net
- Date: Fri, 16 Feb 2001 19:49:32 +1100 (EST)
- CC: gcc-patches at gcc dot gnu dot org
- Reply-To: RDBrown at mira dot net,RodneyBrown at mynd dot com
With a separate fixincludes patch, the following aclocal.m4 patch lets
fpclass satisify INFINITE_P on i586-sco-sysv5uw7.1.0.
fpclass is declared in <ieeefp.h> on UnixWare 2 & 7.
mkcheck still has all run or build failures, but the g++ testsuite
is now down to 52 unexpected failures, so the tests needing libstdc++
are working.
=== g++ Summary for unix ===
# of expected passes 6254
# of unexpected failures 52
# of unexpected successes 1
# of expected failures 119
# of untested testcases 15
2001-02-15 Rodney Brown <RodneyBrown@mynd.com>
* aclocal.m4 (GLIBCPP_CHECK_MATH_DECL_AND_LINKAGE_1):
Include ieeefp.h for fpclass on UnixWare{2,7}.
* mkcheck.in (size_command): Do without GNU-grep when using
size from binutils.
--- libstdc++-v3/aclocal.m4.orig Tue Feb 6 13:28:02 2001
+++ libstdc++-v3/aclocal.m4 Mon Feb 12 17:27:13 2001
@@ -338,17 +338,20 @@
dnl
dnl GLIBCPP_CHECK_MATH_DECL_AND_LINKAGE_1
AC_DEFUN(GLIBCPP_CHECK_MATH_DECL_AND_LINKAGE_1, [
AC_MSG_CHECKING([for $1 declaration])
if test x${glibcpp_cv_func_$1_use+set} != xset; then
AC_CACHE_VAL(glibcpp_cv_func_$1_use, [
AC_LANG_SAVE
AC_LANG_CPLUSPLUS
- AC_TRY_COMPILE([#include <math.h>],
+ AC_TRY_COMPILE([#include <math.h>
+#ifdef HAVE_IEEEFP_H
+#include <ieeefp.h>
+#endif],
[ $1(0);],
[glibcpp_cv_func_$1_use=yes], [glibcpp_cv_func_$1_use=no])
AC_LANG_RESTORE
])
fi
AC_MSG_RESULT($glibcpp_cv_func_$1_use)
if test x$glibcpp_cv_func_$1_use = x"yes"; then
AC_CHECK_FUNCS($1)
--- libstdc++-v3/mkcheck.in.orig Tue Feb 6 13:28:07 2001
+++ libstdc++-v3/mkcheck.in Wed Feb 14 11:07:37 2001
@@ -190,17 +190,17 @@
{
if size --version 2> /dev/null | grep -c GNU > /dev/null;
then # Then we're using a GNU size(1) built for this platform.
# We lose .rodata and .data1 and who knows what else... kludge.
function size_command()
{
case $1 in
TEXT) TEXT=$(size -A $EXENAME | grep ^.text | awk '{print $2}') ;;
- DATA) DATA=$(size -A $EXENAME | grep -w ^.data | awk '{print $2}') ;;
+ DATA) DATA=$(size -A $EXENAME | awk '/^\.data[ ]/{print $2}') ;;
SIZE) SIZE=$(size -A $EXENAME | grep otal | awk '{print $2}') ;;
esac
}
else
# Not using GNU size; check for platform. These numbers seem to match
# up to text/data/total, although their meanings seem to be different.
# THIS TABLE IS SORTED. KEEP IT THAT WAY.
case @host_os@ in