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 82051

Summary: Bad performance displaying treetable data
Product: profiler Reporter: Jiri Sedlacek <jis>
Component: BaseAssignee: Jiri Sedlacek <jis>
Status: VERIFIED FIXED    
Severity: blocker Keywords: PERFORMANCE
Priority: P2    
Version: 6.x   
Hardware: All   
OS: All   
Issue Type: DEFECT Exception Reporter:
Bug Depends on:    
Bug Blocks: 82041    

Description Jiri Sedlacek 2006-08-07 11:57:40 UTC
JTreeTable displaying a lot of data has poor performance - most likely trying 
to render entire tree column instead of rendering just the visible area. Most 
significantly affecting heapwalker, but also problem in other Profiler views.
Comment 1 ehucka 2006-08-07 18:47:48 UTC
This completly 'disallows' using of heapwalker and rest of ide. I loaded cca
14MB of binary heapdump, went to instances and ide's ui was blocked.
Comment 2 Jiri Sedlacek 2006-08-07 20:07:28 UTC
The problem you are describing is caused also by the fact that too many
treenodes are created at once, rendering tree column is only one of causes. Note
that described JTreeTable performance problem is present from the beginning in
all Profiler views and never was a P1 problem.

Also, size of heapdump is not directly related to performance, I'm able to
browse 160MB dump without any problems. The problem occurs when displaying
arrays with thousands of items. This will be fixed in some of next heapwalker
updates (definitely not in M2) by merging every 500 items (generally nodes in
any heapwalker view) into container nodes:

   item 1
   item 2
    ...
   item 499
 + next 500 items
 + next 500 items

For now lowering priority back to P2 as fixing this issue won't fix performance
problems (blocking IDE) when displaying large arrays. You can file separate RFE
according to the above description if you want to have it tracked.
Comment 3 Jiri Sedlacek 2006-08-08 10:12:40 UTC
Fixed for M2.
Comment 4 Jiri Sedlacek 2007-04-03 14:36:18 UTC
Verified.