See the attached log. The jar seems to have built fine though.
Created attachment 103495 [details] merge log
Looks like my patch for bug 146807 wasn't perfect, this should die. I've fixed the eclass in CVS, can you check it dies? Will investigate why it happens tommorow.
It doesn't happen on x86 and the DEPEND string looks fine. So /usr/bin/depend-java-query is somehow broken on fbsd.
Okay, yes it dies. Why? No clue.
OK so this is caused by two things a) missing jdk-defaults.conf on fbsd b) some bug in the failback code (when no default/user setting exists) in depend-java-query a) is now fixed with updated (without revbump) java-config in CVS, will look to b) later
*** Bug 172926 has been marked as a duplicate of this bug. ***
So the problem was when lowest of the possible DEPENDs (1.3) was a VM that was not installed, because only the lowest was tried. Fixed in SVN by trying to match all specified versions.
Released in java-config-2.0.33-r1
ollect2: ld returned 1 exit status make[1]: *** [build/dns-sd] Ошибка 1 make[1]: Leaving directory `/var/tmp/portage/net-misc/mDNSResponder-107.6-r5/work/mDNSResponder-107.6/Clients' make: *** [../Clients/build/dns-sd] Ошибка 2