Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.
Bug 316514 - Debug context activation causes expressions view updates when hidden if selection is empty
Summary: Debug context activation causes expressions view updates when hidden if selec...
Status: RESOLVED FIXED
Alias: None
Product: Platform
Classification: Eclipse Project
Component: Debug (show other bugs)
Version: 3.4.2   Edit
Hardware: PC All
: P3 major (vote)
Target Milestone: 3.4.2+   Edit
Assignee: Platform-Debug-Inbox CLA
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-06-10 14:26 EDT by Natasha D'Silva CLA
Modified: 2010-06-18 17:09 EDT (History)
2 users (show)

See Also:

Attachments
patch (1.02 KB, patch)
2010-06-10 14:43 EDT, Darin Wright CLA 		no flags Details | Diff
patch for 3.4.x (1.39 KB, patch)
2010-06-10 14:44 EDT, Darin Wright CLA 		no flags Details | Diff
patch for 3.5.x (5.64 KB, patch)
2010-06-11 10:24 EDT, Darin Wright CLA 		no flags Details | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Natasha D'Silva CLA 2010-06-10 14:26:38 EDT 
Build Identifier: 

The expressions view overrides the variables view's contextActivated() method, and usually delegates these events to the parent.  However, if the selection is empty (such as when a suspended thread with its top stack frame selected is collapsed) the view does not check to see if it is visible, but goes ahead and sets the input.  This triggers the installation of the DefaultWatchExpressionModelProxy, (one per expression in the view), which, in turn, causes the expressions to be evaluated on suspend when the view is NOT visible. 

I'm not sure what the rationale behind this behaviour is, is there any specific reason why the view handles empty selections differently, or at all, when it is hidden?

Reproducible: Always
Comment 1 Darin Wright CLA 2010-06-10 14:43:59 EDT 
Created attachment 171663 [details]
patch
Comment 2 Darin Wright CLA 2010-06-10 14:44:19 EDT 
Created attachment 171664 [details]
patch for 3.4.x
Comment 3 Darin Wright CLA 2010-06-10 15:05:03 EDT 
Fixed in HEAD (3.7) and 3.6 maintenance branch. I have not released this to the maps yet as 3.6 is not yet complete.
Comment 4 Darin Wright CLA 2010-06-11 10:24:04 EDT 
Created attachment 171726 [details]
patch for 3.5.x
Comment 5 Darin Wright CLA 2010-06-11 10:31:38 EDT 
Released the fix to 3.4.2+, 3.5.2+. Also in 3.6 maintenance and HEAD, but not yet released to a build. I will open new bugs to track all the releases this fix is in.
Comment 6 Natasha D'Silva CLA 2010-06-18 15:57:13 EDT 
Just to clarify - this fix is not going to make it into the 3.6 GA?
Comment 7 Darin Wright CLA 2010-06-18 16:01:53 EDT 
(In reply to comment #6)
> Just to clarify - this fix is not going to make it into the 3.6 GA?

The fix is not in 3.6. The fix will be in 3.6.1.
Comment 8 Pawel Piech CLA 2010-06-18 17:09:05 EDT 
I don't think this bug applies to 3.6 anyway because we got rid of the direct call to setViewerInput().  The additional test that was added is unnecessary.