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

Bug 490945

Summary: Detached views get smaller each time the workbench is restarted (on GTK)
Product: [Eclipse Project] Platform Reporter: Stefan Xenos <sxenos>
Component: UIAssignee: Stefan Xenos <sxenos>
Status: RESOLVED WONTFIX QA Contact:
Severity: normal    
Priority: P3 CC: daniel_megert, Lars.Vogel, markus.kell.r
Version: 4.5   
Target Milestone: 4.6 M7   
Hardware: PC   
OS: Linux   
See Also: https://git.eclipse.org/r/69764
https://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=9492ed48a3660fd9da974d80f9e34660060d6865
https://git.eclipse.org/r/70043
https://git.eclipse.org/c/platform/eclipse.platform.ui.git/commit/?id=7def9e2031c4268c302bb7cfc3dfdc148bf0034e
https://git.eclipse.org/r/70116
Whiteboard:
Bug Depends on: 445900    
Bug Blocks: 491155    

Description Stefan Xenos CLA 2016-04-01 18:15:05 EDT
To reproduce:

1. Detach a view.
2. Measure its exact size (using a screenshot)
3. Shutdown and restart the workbench
4. Measure the size of the detached view again.

Observed:

The view is slightly smaller (by about 6 pixels) on each restart. Eventually the size shrinks to nothing.

Expected:

The view remains exactly the same size on each restart.


Notes:

The root cause in SWT is documented in bug 490944. However, I believe that a workaround is possible in the workbench - we can measure the size of a Shell before and after each call to setVisible(). If the size has changed, we can resize the shell back to its correct size. This bug will track the workaround.
Comment 1 Eclipse Genie CLA 2016-04-01 18:34:30 EDT
New Gerrit change created: https://git.eclipse.org/r/69764
Comment 3 Lars Vogel CLA 2016-04-06 09:50:34 EDT
Might have caused Bug 490945.
Comment 4 Dani Megert CLA 2016-04-06 12:27:52 EDT
(In reply to Lars Vogel from comment #3)
> Might have caused Bug 490945.

Confirmed. Stefan, please fix bug 491155 or revert the change. Thanks.
Comment 5 Stefan Xenos CLA 2016-04-06 12:43:33 EDT
Will do. I'll revert immediately and will re-push with the fix when it's ready.

FYI, if you discover such a regression there's no need to leave the buggy code in place out of politeness. If I couldn't be reached for some reason, I wouldn't want our builds to be unstable in the meantime.
Comment 6 Eclipse Genie CLA 2016-04-06 12:46:55 EDT
New Gerrit change created: https://git.eclipse.org/r/70043
Comment 7 Dani Megert CLA 2016-04-06 12:56:35 EDT
(In reply to Stefan Xenos from comment #5)
> FYI, if you discover such a regression there's no need to leave the buggy
> code in place out of politeness. If I couldn't be reached for some reason, I
> wouldn't want our builds to be unstable in the meantime.

I usually give a day to react unless it's really severe. But good to know you're not offended in such a case :-).
Comment 8 Lars Vogel CLA 2016-04-06 12:59:30 EDT
(In reply to Dani Megert from comment #7)
> (In reply to Stefan Xenos from comment #5)
> > FYI, if you discover such a regression there's no need to leave the buggy
> > code in place out of politeness. If I couldn't be reached for some reason, I
> > wouldn't want our builds to be unstable in the meantime.
> 
> I usually give a day to react unless it's really severe. But good to know
> you're not offended in such a case :-).

+1 for this approach
Comment 10 Eclipse Genie CLA 2016-04-07 14:04:21 EDT
New Gerrit change created: https://git.eclipse.org/r/70116
Comment 11 Stefan Xenos CLA 2016-04-08 11:34:26 EDT
From Markus' comments on the code review, it seems as though he's opposed to having any temporary workarounds for this issue.

I'm not a fan of this approach in general - I don't think users should suffer needlessly while we investigate fixes that may take awhile - provided the workarounds are well documented and easily removable; but this bug also isn't serious enough to fight about.

Marking as WONTFIX, since the root cause in SWT is being tracked elsewhere.