Community
Participate
Working Groups
There always seems to be odds that when I attempt to open a file, Eclipse hangs. I'm running RedHat 8.0. JDK 1.4.1. Eclipse M3 and M4 both exibit this behavior. I open a file by double clicking on the left hand tree. Usually from the java perspective. Odds seem to increase with larger files. Sometimes I can run a day before it hangs. Sometimes only the second or third file opened will hang it.
*** This bug has been marked as a duplicate of 28228 ***
I don't think this is a duplicate of bug 28228. Bug 28228 deals with save/build synchronization problems. In this bug, the java process hangs if I open a file. No save is occuring. No build operation should be taking place, as the file should already be built. Happens even on my sql source, which isn't java related. Further info: CPU is not at 100% according to the ps command. But the task bar is slow (10 sec) to respond to my click so I can bring up a console and kill off the task. The java screen does not update (except for background color) at all during the hang.
Assigning to owner of Editor management for comment.
Don't know if it makes a difference, but the box this hangs on has 2 processors.
The same thing is happening to me on my laptop (single cpu). No log gets generated. Also running RedHat 8.0 with JDK 1.4.1_01. It also happens when creating new file just before it gets opened. It happens somewhere between disabling all views and displaying the file in editor. If I see the file title in the editor I know it's going to work.
Please could you try to get the the stack trace next time it happens. It will be hard for us to fix it without being able to reproduce it. Maybe a Ctrl+Break or "kill -(something)" would force a java core file. If we knew where it is hanging maybe we could fix it.
Hey again. Do you have any ideas on how to get the stack trace?
Please try running eclipse with java instead of javaw (ex: eclipse -vm /jdk/jre/bin/java). Then try "kill -3 processnumber" when it hangs.
Created attachment 3194 [details] java stack trace when eclipse hangs
This stack trace was taken when eclipse >=M4 hangs as described above. I'm running Red Hat 8.0, gtk2-2.0.6-8 and didn't have a problem prior to M4.
Moving to SWT for comment. It seems we called the native OS.gtk_main_iteration () and never came back.
Looks similar to 28553 but I don't want to mark it a duplicate just yet. Asigning to Vikki and SN. Paul, do the windows still redraw when you click on them and layout when you resize or do they just show blank space or garbage?
The windows do not redraw the controls. They simply have a default gray or white background (depending on the pane). I have not been able to get a stack trace by doing a kill -3. Sometimes the JVM terminates without a chance to do a kill on it.
I am putting HANG in the title. From your most recent comments, you say that sometimes the JVM terminates. Is this really a HANG (like you say) or is the JVM dumping core? Is there any way you can make it happen every time? Do you see a message such as: Xlib: unexpected async reply (sequence 0xa59f94)!
I have never seen that message. When I wrote the bug, I had only experience an editor hang. Since then I have occasionally had the JVM terminate. Usually (9 times out of 10?) it hangs and does not terminate. The KDE command bar is very slow to respond when the app hangs. The CPU is still near idle. I'm not exactly sure what is happening under the covers, or how to find out. I can not reproduce the bug at will. It happens about two times a day. Some times not at all, sometimes more.
I checked the java stack trace and it shows "Xlib: unexpected async reply (sequence 0x43eb0b)!" Marking as a duplicate of 23472 *** This bug has been marked as a duplicate of 23472 ***
