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.

Bug 82376 - IDE hangs after drawing navigable aggregation link to 'java.lang.Class' Datatype
Summary: IDE hangs after drawing navigable aggregation link to 'java.lang.Class' Data...
Status: VERIFIED FIXED
Alias: None
Product: uml
Classification: Unclassified
Component: Synchronization (show other bugs)
Version: 5.x
Hardware: All All
: P1 blocker (vote)
Assignee: Kris Richards
URL:
Keywords: REGRESSION
Depends on:
Blocks:
 
Reported: 2006-08-10 12:30 UTC by Alexandr Scherbatiy
Modified: 2006-10-16 14:18 UTC (History)
0 users

See Also:
Issue Type: DEFECT
Exception Reporter:


Attachments
Hanging IDE snapshot (61.27 KB, image/png)
2006-08-10 12:32 UTC, Alexandr Scherbatiy
Details
Actions before hanging of IDE (Information Logging is enabled) (87.08 KB, text/plain)
2006-08-14 14:20 UTC, Alexandr Scherbatiy
Details
Actions after hanging of IDE (Information Logging is enabled) (117.79 KB, application/octet-stream)
2006-08-14 14:21 UTC, Alexandr Scherbatiy
Details

Note You need to log in before you can comment on or make changes to this bug.
Description Alexandr Scherbatiy 2006-08-10 12:30:58 UTC
Steps to reproduce:

- Create a UML Java-Platform Mopdel project and synchronize
  it with Java Application
- Create a Class digram
- Put an Class element on the diagram and name it 'MyClass'
- Go to source of the Class element
- insert 'Class cls = getClass();' attribute into source code
- Go to the diagram
- Drag and drop 'java.lang.Class' Datatype from Model tree to the
  diagram
- Draw 'Navigable Aggregation' link from the 'MyClass' Class element to the
  Datatype
   IDE hangs
Comment 1 Alexandr Scherbatiy 2006-08-10 12:32:17 UTC
Created attachment 32753 [details]
Hanging IDE snapshot
Comment 2 Craig Conover 2006-08-10 23:19:52 UTC
sunflower: I'm curious. Do you not see the Navigable Association link on the
diagram when you drag the java.lang.Class DataType onto the diagram? I see it.
And I even add an additional NavAggr link from MyClass to Class datatype and it
works fine. I noticed that there is a duplicate DataType element at the root
package level in the UML project. I dragged that to the diagram and added NavAgg
from MyClass to it and it still did not hang.

I am using WinXP. What OS are you testing on? Solaris I assume? I assigned this
issue to you. Just reassigned back to me (conover) when you reply. Thanks.
Comment 3 Peter Lam 2006-08-10 23:56:57 UTC
I can't reproduce either using jdk 1.6 (I assume that's what you use) on WinXP,
Linux FC3, Solaris 10 sparc, and MacOS ppc. I tried a few times on each platform
and I didn't see any ide hang at all.
Comment 4 Andrew Korostelev 2006-08-11 16:22:49 UTC
It seems like there is an issue with threads...

For me on milestone17 this bug is always reproducible on two computerns, rarely
reproducible on the third and 50/50 on the fourth.
And on the fourth I have different errors from time to time. Like UML attempt to
remove associated methods or NPE.

And on milestone m17a bug is not reproducible on the box where it was 100%
reproducible on m17.

Will retry with 17a and update comments.
Comment 5 Alexandr Scherbatiy 2006-08-14 14:18:58 UTC
Hi Conover,
I want to add the second aggregation link between class and datatype
to make one more attributes in my class. If I have some attributes with the
same types that means that they all will be connected to the same datatype.

The issue is still reproduced. Please, have a look at attached log_1.txt
and log_2.txt files. I enabled UML information logging. So these files
have all information what I did. My IDE was hanging  for 9 minutes according to 
the log files.



Comment 6 Alexandr Scherbatiy 2006-08-14 14:20:39 UTC
Created attachment 32883 [details]
Actions before hanging of IDE (Information Logging is enabled)
Comment 7 Alexandr Scherbatiy 2006-08-14 14:21:11 UTC
Created attachment 32884 [details]
Actions after hanging of IDE (Information Logging is enabled)
Comment 8 Kris Richards 2006-10-03 21:44:02 UTC
This was fixed as a result of the new reverse engineering code. (tested against jdk1.5)
Comment 9 Andrew Korostelev 2006-10-16 14:18:09 UTC
verified in all-nbms-griffin-061015.
To be sure have tried similar usecases with current RE/CodeGen.