Summary: | NPE setting output directory of a source folder | ||
---|---|---|---|
Product: | [Eclipse Project] JDT | Reporter: | Jared Burns <jared_burns> |
Component: | Core | Assignee: | JDT-Core-Inbox <jdt-core-inbox> |
Status: | RESOLVED WORKSFORME | QA Contact: | |
Severity: | major | ||
Priority: | P3 | CC: | jeffmcaffer, kent_johnson |
Version: | 2.1 | ||
Target Milestone: | 2.1 RC3 | ||
Hardware: | Other | ||
OS: | Linux-GTK | ||
Whiteboard: |
Description
Jared Burns
2003-03-05 15:10:45 EST
I tried to reproduce but it worked fine for me. Jared: Can you try this again with the jit disabled please? Disabling the JIT via -classic got rid of the problem. Wow? Kent, we have a situation where we want to use a different output directory for certain source folders. We want to do this for 2.1. The fact that you thought to disable JIT on this one makes me think that you've seen this problem before? I've noticed that I don't get this problem if I run Eclipse on JDK 1.4.1. Is this the result of a VM bug? If we can't use different output directories, we have to use black classloader magic. So what I really want to know is whether or not this problem is fixable. Can you advise? As a point of interest, what VM are you using? The timestamp looks like SR1 but I don't have all the dates in my head. Please try on SR2 and report. That is the officially approved IBM VM for Eclipse so it would be interesting to know what the behaviour is. Multiple output folders are legite in 2.1. The only problem here is a VM problem, which has nothing to do with JDT/Core. We cannot workaround a JIT bug throwing NPE at random points... Jared - can you see if you reproduce the problem on SR2 ? If so, you should ping Hursley with this testcase. I don't get the problem on the IBM 1.3.1 build from 20021102. For IBM 1.3.1 VM the VM timestamps are: linux SR2 = 20020622 SR3 = 20021102 Windows SR2 = 20020710 SR3 = 20021102 SR2 is the one with which we should be testing. Please try your case on that VM and report. Closing as a VM bug. |