Posted by mikro2nd
on June 15, 2005 at 6:40 AM PDT
Actually, its probably not so much developers who want Bug 4885936 to get fixed as much as frustrated implementors, integrators and sysadmins.
The bug is a Hotspot crash and has been much-reported against JDK versions from 1.4.2 all the way through 1.5.0_03, but is marked "Closed" because the droid who evaluated it could not reproduce it.
Now I know its a hard-to-reproduce bug, and I certainly do not know how to produce a test-case (which is what Sun's bug-evaluators want). I mean, for G**'s sake, its an INTERMITTENT server crash! Its largely been seen in production environments. That means that people can't just dump a stack-trace or convince their boss to allow them to spend days or weeks chasing the bug; the priority is to get a prodcution server back online!
I've seen (many times!) this bug happen almost immediately during Tomcat initialisation, and I've had the same environment perform beautifuly for days before suddenly dying.
But just to close the bug and then ignore all subsequent pleas for somebody in Sun to pay attention?
Hopefully now, with the more open development process, somebody in the community with time on their hands can take a look at this thing.