Community
Participate
Working Groups
This bug was discovered on Eclipse 3.1 M5a (I20050219-1500): When debugging any Java application, if the variables view is open and visible, doing stepping operations in rapid succession, such as by pressing and holding the F6 key, would sometimes cause an error pop-up: Exceptions occurred attempting to step over the frame org.eclipse.jdi.TimeoutException: Timeout occurred while waiting for packet 2225 occurred creating step request. After the error pops up for the first time, all stepping operations will no longer work (it will give a similar org.eclipse.jdi.TimeoutException error each time), the variables view will display the values of all variables as <unknown value>, and sometimes Eclipse would freeze completely. Below is the full error log entry: !SESSION 2005-03-04 09:56:51.937 ----------------------------------------------- eclipse.buildId=I20050219-1500 java.version=1.4.2_05 java.vendor=Sun Microsystems Inc. BootLoader constants: OS=win32, ARCH=x86, WS=win32, NL=en_US Command-line arguments: -os win32 -ws win32 -arch x86 !ENTRY org.eclipse.debug.ui 4 5012 2005-03-04 09:56:51.937 !MESSAGE Exceptions occurred attempting to step over the frame !SUBENTRY 1 org.eclipse.jdt.debug 4 5010 2005-03-04 09:56:51.937 !MESSAGE org.eclipse.jdi.TimeoutException: Timeout occurred while waiting for packet 2225 occurred creating step request. !STACK 0 org.eclipse.jdi.TimeoutException: Timeout occurred while waiting for packet 2225 at org.eclipse.jdi.internal.connect.PacketReceiveManager.getReply(PacketReceiveManager.java:163) at org.eclipse.jdi.internal.connect.PacketReceiveManager.getReply(PacketReceiveManager.java:172) at org.eclipse.jdi.internal.MirrorImpl.requestVM(MirrorImpl.java:174) at org.eclipse.jdi.internal.MirrorImpl.requestVM(MirrorImpl.java:192) at org.eclipse.jdi.internal.MirrorImpl.requestVM(MirrorImpl.java:207) at org.eclipse.jdi.internal.ObjectReferenceImpl.isCollected(ObjectReferenceImpl.java:388) at org.eclipse.jdi.internal.request.EventRequestImpl.addStepFilter(EventRequestImpl.java:409) at org.eclipse.jdi.internal.request.EventRequestManagerImpl.createStepRequest(EventRequestManagerImpl.java:195) at org.eclipse.jdt.internal.debug.core.model.JDIThread$StepHandler.createStepRequest(JDIThread.java:1724) at org.eclipse.jdt.internal.debug.core.model.JDIThread$StepHandler.step(JDIThread.java:1675) at org.eclipse.jdt.internal.debug.core.model.JDIThread.stepOver(JDIThread.java:1215) at org.eclipse.jdt.internal.debug.core.model.JDIStackFrame.stepOver(JDIStackFrame.java:415) at org.eclipse.debug.internal.ui.actions.StepOverActionDelegate.stepAction(StepOverActionDelegate.java:30) at org.eclipse.debug.internal.ui.actions.StepActionDelegate.doAction(StepActionDelegate.java:25) at org.eclipse.debug.internal.ui.actions.AbstractDebugActionDelegate$DebugRequestJob.run(AbstractDebugActionDelegate.java:97) at org.eclipse.core.internal.jobs.Worker.run(Worker.java:67)
Is your target program also running on Sun's 1.4.2_05 VM?
Yes, the target program is running on Sun's j2re1.4.2_05
Does reselecting the stack frame in the debug view allow you to step once again? Does this always happen using "Step Over" or do all the step actions have the same effect?
Reselecting the stack frame does not allow me to step once again. All 3 stepping operations (Step Into, Step Over, and Step Return) can cause this problem. In order to reproduce this bug, there has to be several variables of sufficient complexity displayed in the variables view. Below is the code that I used and the steps to reproduce the bug: import java.io.*; public class Hello { private static int x=0; private static void f() { BufferedReader in = new BufferedReader(new InputStreamReader(System.in)); int i=0; int j=0; x++; if(x<100) f(); while(i<100) { j+=i; i++; } } public static void main(String[] args) { f(); } } - Set a breakpoint at the while loop and remove it after the debug session starts. - Expand the BufferedReader in the variables view so that at least some of its member variables are displayed - Try the various stepping operations in quick succession by holding down onto their keyboard shortcuts After causing this error for the first time. If I terminate the debug session, start a new one, and try stepping again, the following message gets printed to standard error: Transport error, error code = -1 After that the debug session terminates because it got disconnected. The stack in the Debug View displays: <disconnected> Hello at localhost:1304 There is no error log entry for the above error.
I am unable to reproduce the problem as shown. What version of Windows XP are you running? What sort of hardware are you using?
My machine is an IA32 running Windows XP with Service Pack 2.
released change to PacketReceiveManager. Slightly changed the synchronization of getReply(), and made error message more detailed when a Timeout occurs.
Created attachment 18544 [details] Replacement jdt.debug plugin Kevin (L), Is there any chance that you could run your test with today's I-Build and this jdt.debug plugin? It may give us some more useful info in the error messages.
Created attachment 18550 [details] Error log for StepReturn in Eclipse I20050308-0835 This is the complete error log as I started Eclipse I20050308-0835, created the test project, and then ran it. I tested the Step Return operation this time and held down the F9 key.
What we really need is a trace from today's build with the attached jdt.debug plugin. To get this you'll need to unzip the plugin attached and copy/paste the unzipped plugin into the <eclipse_home>/ plugins directory on your machine. After doing this it's best to start eclipse with the -clean option to be sure that the new plugin is loaded properly (eclipse.exe -clean). Today's changes are not included in any build yet. Sorry for the confusion.
Sorry, I forgot to mention that I overwrote the jdt.debug plugin in today's build with the one you supplied before I ran the tests. So the error log I attached should be correct.
Created attachment 18555 [details] new plugin ugh. I forgot to release changes to the properties file. Working on too many machines at once. Sorry, please try this new plugin.
Created attachment 18558 [details] New Error log for step return on Eclipse I20050308-0835 Here's the new error log, using the new plugin. Hope it'll be useful.
These look like real timeouts. You can increase the timeout on the Preferences/Java/Debug page. 3 seconds should be long enough though. These timeouts are occuring in the UI thread. Should move host-target communication out of UI thread.
Kevin, Are you running ZoneLabs (or another firewall)? Could you disable it and run your test again. I'm wondering if there is something causing packet loss between the host and target.
You're right. Turning off the firewall seems to solve the problem. I notice there's firewall activity whenever I perform stepping operations. However, increasing the timeout setting seems to solve the problem too. Seems like the higher I set it, the lesser the chance the exception occurs... Maybe it's a performance problem, and has nothing to do with packet loss. Perhaps I'm just stressing out Eclipse too much by stepping so fast?
Good questions! How fast is your machine? My WinXP box is a P4 2.66 GHz with 768MB of RAM running ZoneLabs firewall. I don't get Timeouts when holding the F6 on it. I'm also on SP1 still. I'll try upgrading my machine to SP2 Does your firewall have a log file? Is there anything in it?
My machine is a P4 3.2 GHz with 1.5 GB Ram running Windows XP with SP2 I'm also running Zone Labs, and I don't notice anything in the log file. Actually the error seems to be occuring a lot less often now... Maybe I was just running too many programs last time, causing Eclipse to run slowly and generate those timeout exceptions.
*** This bug has been marked as a duplicate of 83937 ***
