Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 345138

Summary: Eclipse CDT hangs when bringing up autocompletion list
Product: [Tools] CDT Reporter: Paul Bombach <pmb>
Component: cdt-editorAssignee: Project Inbox <cdt-editor-inbox>
Status: CLOSED DUPLICATE QA Contact: Anton Leherbauer <aleherb+eclipse>
Severity: blocker    
Priority: P3 CC: cdtdoug
Version: 7.0.2   
Target Milestone: ---   
Hardware: Macintosh   
OS: Mac OS X - Carbon (unsup.)   
Whiteboard:
Attachments:
Description Flags
stack trace
none
jstack output none

Description Paul Bombach CLA 2011-05-09 09:30:25 EDT
Created attachment 195067 [details]
stack trace

Any time I try to bring up an auto completion list, CDT simply hangs. It has to be killed -9 in order to stop it. In fact, attaching to it via visualvm causes visualvm to hang when attempting to get a thread dump. I'm running the following version:

Version: Helios Service Release 2
Build id: 20110218-0911

It's being run on a Macbook Pro with the SandyBridge chipset

 Model Name:	MacBook Pro
  Model Identifier:	MacBookPro8,2
  Processor Name:	Intel Core i7
  Processor Speed:	2.2 GHz
  Number of Processors:	1
  Total Number of Cores:	4
  L2 Cache (per Core):	256 KB
  L3 Cache:	6 MB
  Memory:	8 GB
  Boot ROM Version:	MBP81.0047.B0E
  SMC Version (system):	1.69f1
  Serial Number (system):	C02FDAP6DF91
  Hardware UUID:	B2FB15E0-DC0C-5353-A25D-0C27E3F0BB37
  Sudden Motion Sensor:
  State:	Enabled

I've attached a thread dump obtained via the Activity Monitor program. Please let me know what other info might be necessary to get this resolved. Eclipse/CDT is a great tool and unfortunately, this is preventing me from using it.
Comment 1 Anton Leherbauer CLA 2011-05-12 05:21:23 EDT
We need a JVM thread dump.  The attached file is not useful.
Note that you can disable auto-activation of content assist in the preferences.
Comment 2 Paul Bombach CLA 2011-05-13 10:02:24 EDT
Created attachment 195593 [details]
jstack output

Since I can't get a java stack trace after the freeze happens, I ran the jstackSeries.sh command every .5 seconds. This output is the last output before it froze. I hope it can provide some clues.
Comment 3 Anton Leherbauer CLA 2011-05-16 03:32:39 EDT
(In reply to comment #2)
> Created attachment 195593 [details]
> jstack output
> 
> Since I can't get a java stack trace after the freeze happens, I ran the
> jstackSeries.sh command every .5 seconds. This output is the last output before
> it froze. I hope it can provide some clues.

The thread dump is only useful if taken _while_ the freeze occurs. Attached dump shows an idle event loop which does not help.
Did you try kill -QUIT <pid> on the command line?
Comment 4 Paul Bombach CLA 2011-05-16 10:51:19 EDT
(In reply to comment #3)
> (In reply to comment #2)
> > Created attachment 195593 [details] [details]
> > jstack output
> > 
> > Since I can't get a java stack trace after the freeze happens, I ran the
> > jstackSeries.sh command every .5 seconds. This output is the last output before
> > it froze. I hope it can provide some clues.
> 
> The thread dump is only useful if taken _while_ the freeze occurs. Attached
> dump shows an idle event loop which does not help.
> Did you try kill -QUIT <pid> on the command line?

Yes, I did. Unfortunately once it freezes it is completely unresponsive. Is there any debugging output that could be turned on that might provide a clue as to what is going on?
Comment 5 Anton Leherbauer CLA 2011-05-17 03:06:28 EDT
This looks similar to bug 345970.  Could you try with a 32 bit version of Eclipse and JVM?
Comment 6 Paul Bombach CLA 2011-05-18 08:48:06 EDT
(In reply to comment #5)
> This looks similar to bug 345970.  Could you try with a 32 bit version of
> Eclipse and JVM?

Yesterday, I used the 32 bit carbon version  for the entire day without any problems. Today, I'll try the 32 bit cocoa version.
Comment 7 Anton Leherbauer CLA 2011-05-23 04:23:14 EDT
Thanks for checking.  I think there is enough evidence that this is the same issue as bug 345970.

*** This bug has been marked as a duplicate of bug 345970 ***