Summary: | SWTException: Graphic is disposed for PlaceholderImpl | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | [Eclipse Project] Platform | Reporter: | Marcel Bruch <marcel.bruch> | ||||||
Component: | UI | Assignee: | Bogdan Gheorghe <gheorghe> | ||||||
Status: | VERIFIED FIXED | QA Contact: | |||||||
Severity: | normal | ||||||||
Priority: | P3 | CC: | bsd, daniel_megert, emoffatt, gheorghe, john.arthorne, Mike_Wilson, ob1.eclipse, pwebster | ||||||
Version: | 4.2 | Flags: | gheorghe:
review+
|
||||||
Target Milestone: | 4.2 RC1 | ||||||||
Hardware: | PC | ||||||||
OS: | Mac OS X - Carbon (unsup.) | ||||||||
Whiteboard: | |||||||||
Attachments: |
|
Description
Marcel Bruch
2012-05-17 05:47:08 EDT
Eric, could this be effected by the fix for the part image disappearing after a DND operation? PW (In reply to comment #1) > Eric, could this be effected by the fix for the part image disappearing after a > DND operation? The patch for bug 378184 shouldn't have any impact here: it overrides the AbstractPartRenderer#getImage(MUIElement), and the committed code has a guard to prevent returning a disposed image. And the overridden image returned in that patch was already been in place until the DND (e.g., the test case of the Problems view: the custom icon is already seen when the part is first created). See also bug 379976 PW *** Bug 379976 has been marked as a duplicate of this bug. *** Turns out it's the interaction of bug 378184 with our org.eclipse.ui.internal.e4.compatibility.CompatibilityView.updateImages(MPart) code and the difference of how we treat an uninstantiated view in a non-active perspective when the view instance is closed. - In the short term, we will revert bug 378184. - We also need to open a new bug to fix how org.eclipse.ui.internal.e4.compatibility.CompatibilityPart.updateTabImages(MUIElement) works - Bogdan has agreed to put some protection CTabFolder/CTabRendering/CTabFolderRenderer to turn the exception into a no-op, as hitting this exception can destroy your perspective. PW I've reverted the changes made to fix Bug 378184 and re-opened it (for 4.2.1). Since this fixes this issue I'll mark this defect for RC1 and as FIXED... Created attachment 215865 [details]
Patch
Re-opening to fix an issue in the TrimStack... Created attachment 215870 [details]
Ensure that we never return a disposed 'override' image
commit 40ffef101aae6a9a1b7679c1edb682e2e9462b52 Ensures that we don't return a disposed 'override' image. I was testing essentially the same scenario using a minimized stack (since it's the other consumer of the image stored in the transient data) and encountered a similar issue... Verified in I20120518-2145. *** Bug 378359 has been marked as a duplicate of this bug. *** |