On 2014/01/01 10:11, Scott Vanderbilt wrote: > On 1/1/2014 10:00 AM, Ian Darwin wrote: > > >Given that Java 6 has been EOL'd by Oracle for almost a year (Feb 2013, > >as announced in Feb 2011), and it's on an operating system they > >don't even know about, your bug report will most likely be ignored. > >Sorry you wasted time there. > > > >Given the situation with Java 6 being broken, it would probably > >make sense to work on correcting/updating apps to work on Java 7. > > Thanks for that, Ian. > > I should have made clear that I'm not actually running JDK 1.6, but rather > 1.7.0_21 (from ports current). And my bug report to Sun indicated that as > well. > > I know nothing about JDK internals, so I can't even begin to speculate. All > I know is the bug occurred while running 1.7, yet 1.6 is the port now marked > as broken. I do know that JDK 1.6 is a dependency for building 1.7, so it > probably has something to do with that linkage. >
The problem is that JDK 1.6 itself quite often builds, but produces a package which does not work. Previously they used to fail intermittently and in many cases after 6 or 7 attempts to rebuild things, sometimes also needing a reboot onto an SP kernel, I've been able to get a full bulk build done eventually. But this is far too time-consuming to do regularly. It has now changed such that I am unable to get any working builds so marking jdk broken saves many hours dealing with failures in bulk builds. Your current options are: - move to amd64. - move back to code from a month or two ago. If anyone is able to help track down the timeframe this got broken (or even pin it to a particular commit in base), that would be extremely helpful...
