Hello Mikael:
All,
Please review this backport of the fix for 8031709 to
jdk8u/jdk8u-dev. The fix is the very same as the one Erik did in jdk9:
http://mail.openjdk.java.net/pipermail/build-dev/2014-January/011534.html
http://hg.openjdk.java.net/jdk9/dev/rev/690bb11e1df2
Patch below (new generated-configure.sh excluded).
Cheers,
Mikael
diff -r 6247dd67fde7 common/autoconf/jdk-options.m4
--- a/common/autoconf/jdk-options.m4 Mon Jul 07 18:02:20 2014 -0700
+++ b/common/autoconf/jdk-options.m4 Thu Jul 10 00:47:52 2014 -0700
@@ -134,7 +134,7 @@
fi
# Replace the commas with AND for use in the build directory name.
- ANDED_JVM_VARIANTS=`$ECHO "$JVM_VARIANTS" | $SED -e 's/^,//' -e
's/,$//' -e 's/,/AND/'`
+ ANDED_JVM_VARIANTS=`$ECHO "$JVM_VARIANTS" | $SED -e 's/^,//' -e
's/,$//' -e 's/,/AND/g'`
COUNT_VARIANTS=`$ECHO "$JVM_VARIANTS" | $SED -e 's/server,/1/' -e
's/client,/1/' -e 's/minimal1,/1/' -e 's/kernel,/1/' -e 's/zero,/1/'
-e 's/zeroshark,/1/' -e 's/core,/1/'`
if test "x$COUNT_VARIANTS" != "x,1"; then
BUILDING_MULTIPLE_JVM_VARIANTS=yes
It took me a while, but this non-reviewer spotted it eventually :-)
The single char change looks most sensible.
Ah... let us all raise our glasses in a toast to the non-linear nature
of software....
Agreed. This looks good - approved.
Tim
