Bug 416902 - Regression from 4.3: Navigation History badly broken
Summary: Regression from 4.3: Navigation History badly broken
Status: VERIFIED FIXED
Alias: None
Product: Platform
Classification: Eclipse Project
Component: UI (show other bugs)
Version: 4.3.1   Edit
Hardware: All All
: P3 major (vote)
Target Milestone: 4.3.1   Edit
Assignee: Eric Moffatt CLA
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on: 417001
Blocks:
  Show dependency tree
 
Reported: 2013-09-10 07:14 EDT by Dani Megert CLA
Modified: 2013-09-12 05:44 EDT (History)
4 users (show)

See Also:
daniel_megert: pmc_approved+
pwebster: review+


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Dani Megert CLA 2013-09-10 07:14:48 EDT
M20130905-0705 a.k.a. 4.3 RC3. Works fine in R4.3.

1. start with new workspace
2. close Welcome
3. paste the following snippet into the 'Package Explorer':
public class A {
B b;
}
public class B {
C c;
}
public class C {
}
4. File > Close All
5. open A
6. in the editor Ctrl+click on B
7. in the editor for B Ctrl+click on C
8. hover over the Left arrow in the toolbar: it wrongly says: Back to A.java
9. click on the Left arrow: ==> goes to A instead of B
Comment 1 Dani Megert CLA 2013-09-10 07:15:31 EDT
This is a 4.3.1 candidate for me.
Comment 2 Eric Moffatt CLA 2013-09-10 10:13:09 EDT
The likely problem here may be some code I added to 'auto-activate' a view when it's brought to the top if the currently active part is in the same stack. I'll start there and take a look...
Comment 3 Eric Moffatt CLA 2013-09-10 16:20:47 EDT
Dani, I tracked this down (eventually) to code in setPartState that tried to activate an editor being opened with the initial state VISIBLE if it was in the same stack as the currently active part.

Unfortunately just calling 'activate' isn't enough...

Here's the Gerrit ref:

https://git.eclipse.org/r/16300

Now it appears to work and all tests are green...

While I'll initially put this fix into Luna (if it's approved) I'll then open another defect to force us to clean up this code; constructs like 'delegateBringToTop' should be unnecessary.
Comment 4 Dani Megert CLA 2013-09-11 07:17:42 EDT
(In reply to Eric Moffatt from comment #3)
> Dani, I tracked this down (eventually) to code in setPartState that tried to
> activate an editor being opened with the initial state VISIBLE if it was in
> the same stack as the currently active part.
> 
> Unfortunately just calling 'activate' isn't enough...
> 
> Here's the Gerrit ref:
> 
> https://git.eclipse.org/r/16300

Submitted with http://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=b26ac6b74678597edab1228bcb022842215fe924
 

> Now it appears to work and all tests are green...
> 
> While I'll initially put this fix into Luna (if it's approved)

I guess you meant Kepler SR1. I've just approved it.


> I'll then
> open another defect to force us to clean up this code; constructs like
> 'delegateBringToTop' should be unnecessary.

Filed bug 417001 for 'master.
Comment 5 Eric Moffatt CLA 2013-09-11 08:51:28 EDT
Thanks Dani !
Comment 6 David Williams CLA 2013-09-11 14:34:15 EDT
For what its worth, I "sanity check" the build input and found this change (plus one about changing 'api filters') were only two changes that went into RC4. Assuming all is as expected. 



629251e64d8bebd34caca0939ee2e4c460eee609 Build input for build M20130911-1000
Entering 'eclipse.jdt'
Entering 'eclipse.jdt.core'
Entering 'eclipse.jdt.core.binaries'
Entering 'eclipse.jdt.debug'
Entering 'eclipse.jdt.ui'
Entering 'eclipse.pde'
Entering 'eclipse.pde.build'
Entering 'eclipse.pde.ui'
Entering 'eclipse.platform'
Entering 'eclipse.platform.common'
Entering 'eclipse.platform.debug'
Entering 'eclipse.platform.releng'
Entering 'eclipse.platform.resources'
Entering 'eclipse.platform.runtime'
Entering 'eclipse.platform.swt'
Entering 'eclipse.platform.swt.binaries'
Entering 'eclipse.platform.team'
Entering 'eclipse.platform.text'
Entering 'eclipse.platform.ua'
Entering 'eclipse.platform.ui'
b26ac6b74678597edab1228bcb022842215fe924 Fix for Bug 416902 - Regression from 4.3: Navigation History badly broken
e6f79374eca103e2d3adc97a24d909be23cb8350 Bug 416843 - Lots of API errors in 4.3.1
Entering 'rt.equinox.binaries'
Entering 'rt.equinox.bundles'
Entering 'rt.equinox.framework'
Entering 'rt.equinox.p2'
Comment 7 Dani Megert CLA 2013-09-12 05:32:07 EDT
(In reply to David Williams from comment #6)
> For what its worth, I "sanity check" the build input and found this change
> (plus one about changing 'api filters') were only two changes that went into
> RC4. Assuming all is as expected. 
>
> e6f79374eca103e2d3adc97a24d909be23cb8350 Bug 416843 - Lots of API errors in 
> 4.3.1

Yes. That change only added API filters. Paul and I discussed this and decided to put it in for 4.3.1.
Comment 8 Dani Megert CLA 2013-09-12 05:44:18 EDT
Verified in M20130911-1000.