isalist was less trustable than I thought (or rather, one can trust it to
authorRichard Levitte <levitte@openssl.org>
Fri, 16 Aug 2002 09:41:14 +0000 (09:41 +0000)
committerRichard Levitte <levitte@openssl.org>
Fri, 16 Aug 2002 09:41:14 +0000 (09:41 +0000)
come up with all kinds of names we don't have in our targets).
Besides, our sparcv9 targets currently generate sparcv8 code, I'm told.

config

diff --git a/config b/config
index 7cfc4f93306154536d03ee3813e8bd686fcefd2d..3e9af7680a772d9351ec66d139cf0c073c5d6645 100755 (executable)
--- a/config
+++ b/config
@@ -595,10 +595,8 @@ EOF
         fi ;;
   *-*-linux1) OUT="linux-aout" ;;
   sun4u*-*-solaris2)
-       ISA=`(isalist) 2>/dev/null`
-       ISA64=`echo $ISA | grep sparcv9`
-       ISA=`set $ISA; echo $1`
-       OUT="solaris-$ISA-$CC" ;;
+       OUT="solaris-sparcv9-$CC"
+       ISA64=`(isalist) 2>/dev/null | grep sparcv9`
        if [ "$ISA64" != "" ]; then
            if [ "$CC" = "cc" -a $CCVER -ge 50 ]; then
                echo "WARNING! If you wish to build 64-bit library, then you have to"