This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
If I run the ant target several times (5-20x) I get OutOfMemoryError without an available stack trace. I'm able to reproduce it reliably only with jdk1.3.1 but not at all with 1.4. In order to reproduce the bug I'm using projects/test/built.xml/runtests target. No matter the target is launched from an explorer or a toolbar shortcut. To get sources look at http://projects.netbeans.org/proposals/prototype/index.html, stripped version, prj40_prototype branch. In most cases the ide freezes during or after a xsl transformation, I guess (according to the output tab).
Target milestone was changed from '3.4' to TBD.
I played with this problem and I can confirm that there is a _problem_. Unfortunately I don't have idea how to continue, how to track the bug. If I execute the JUnit tests from the IDE about 5x-6x times the IDE (Build 200207100858 on JDK 1.3.1_02) suddenly does following: * the running script is stopped at random place (from what I can see in the output window) * execution view portion of the output window contains "No Processes Running" * the IDE is frozen - I can minimize/maximize it, but no repaint is done * thread dump is _empty_ - bunch of threads, but nothing important there After a while IDE simply dies - not message box, no record in log file, nothing. Sometimes I had to kill the process. This behaviour is reproducible. I was executing the Ant scripts with DEBUG verbosity level and IDE crashes always at the different place: a few times if was at the end of fifth execution (just when the XSL processing is running), but once it happened early after the sixth execution. I was tracking the memory (via Memory View) and it does not shown any problems: the memory is all the time around 25-46MB. Has anybody some hints what to check, try, do??
For now may just have to document that users of XTest should beware of in-IDE execution, until we figure out what is wrong. David I suggest the IDE in this state be run through e.g. OptimizeIt! (we have a site license, let's use it) to find the memory leaks.
added PERFORMANCE keyword
I am afraid targets running xtests is not the only case. Building the ide or just modules alone results in the same issue after several uses. Possible workaround seems to be a migration to jdk 1.4.x
Other observations: * I tried to build project prototype from the IDE (see Jan's url for more details about projects prototype - it is openide, core, projects, java, editor and a few other modules). The build contained also first start of the IDE. The IDE died after 10th execution of build script. * I tried the same as above just without first start. The IDE died after about 13th execution. But this time I got also: # # HotSpot Virtual Machine Error, Internal Error # Please report this error at # http://java.sun.com/cgi-bin/bugreport.cgi # # Error ID: 4349254E560E43505000F4 # # Problematic Thread: prio=5 tid=0x983cc8 nid=0x548 runnable # That means that the problem is not related to XTest/JUnit scripts only, but to all Ant scripts in general. I think it is JVM bug. As Honza writes, on JDK 1.4 this issue does not happen. Therefore I tend to close it as WONTFIX. Before I do so I would like to give a try OptimizeIt.
Proposed readme text: "Description: If you run an Ant target in the IDE multiple times in rapid succession, an out of memory error may occur. This error occurs on JDK 1.3.1. Workaround: Run the IDE on JDK 1.4." Does this sound OK?
The relnotes text sounds OK to me.
Patrick you have to add yourself as a CC when adding comments to a bug or you won't see replies...
I tried OptimizeIt, but did not learn much. The Ant module classes looked OK. Closing as WONTIX.
To David: Does it mean we do not support to use the ant module for execution with jdk1.3.1 in future? To Patrick: It does not matter how rapid the succession is. Otherwise OK.
Adding pnejedly & jlahoda on the cc:. Jan has a bit similar problem. Jan, could you explain your problem?
jpokorsky told me that after running IDE on JDK 1.4 for longer time (>1day) the problem appeared there too.
*** Issue 27625 has been marked as a duplicate of this issue. ***
Attaching information I got from originator of issue 27625: I know that in my previous job, we saw this problem when we were creating several instances of javac to compile repeatedly. So it leaks some kind of handles like file handles and in the AWT thread there is a silent quit without any notification on that kind of leak. If you see the thread dump carefully, you will see there in NO AWT-Event-Queue thread. I think this bug is similar 4517726 on bug parade: <http://developer.java.sun.com/developer/bugParade/bugs/4517726.html> So my guess is that ANT is creating too many instances of javac (or something like that) and that is causing this bug to trigger. Do we run ANT inprocess? If so, making it out of process (optionally) can possibly help. We might also want to look at how many instances of ANT/Javac we are creating and try to minimize those.
I seem to recall that there have been major cleanups in Javac code for recent JDK releases (sorry, forget details) - basically meaning that it should now be really reentrant and clean up after itself properly. Yes, NB currently only runs Ant in-VM.
I believe that I am seeing this problem. I've recently started using XDoclet and the problem occurs frequently, i.e. after about an hours constant use. I understand that if I go to 1.4 the problem would mitigated somewhat, but unfortunately I'm stuck with 1.3 for a couple of reasons. First, our target platform is OS390/Linux where 1.4 is not yet supported. Secondly, when I run 1.4 on WinNt I get critical errors (application aborts) having to do with a spurious interaction with Microsoft's Direct Draw. Would you consider re-opening this problem? Plus, it seems like a P1 to me.
Could PERFORMANCE team look at this? I do not have time to work on this issue at the moment.
Chuck, re. 1.4, in case it is helpful for you - (1) remember you can run NB under 1.4 and still develop for 1.3 (I guess you know this); (2) I think there is some way to turn off DirectDraw support in AWT in case your driver is buggy or something. Don't remember offhand how.
removing RELNOTE keyword as jdk 1.3 is no longer supported
closed
what kind of forum is this? not giving any kind of solution for our problem???? we are working on the same problem since a week and still not getting the problem.. we get the error of "outofmemory"... the first worst thing about netbeans is that it takes eons to startup its modules and then works too slow giving errors of out of memory... we have a 2GB RAM with 360GB hard disk and still the same problem... do we have to buy a super computer and make netbeans work on it?? we are plannin to change the IDE if the same problem persists... netbeans is really the worst IDE... sorry, but before changing the IDE would like to have a solution to memory problem if you can suggest
Spuriously reopened.