| Summary: | Some comboboxes in GIMP have wrong background | ||
|---|---|---|---|
| Product: | [Plasma] Oxygen | Reporter: | Ruslan Kabatsayev <b7.10110111> |
| Component: | gtk2-engine | Assignee: | Hugo Pereira Da Costa <hugo.pereira.da.costa> |
| Status: | CLOSED FIXED | ||
| Severity: | normal | CC: | b7.10110111, hugo.pereira.da.costa, web |
| Priority: | NOR | ||
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Platform: | Unlisted Binaries | ||
| OS: | Linux | ||
| Latest Commit: | Version Fixed In: | ||
| Sentry Crash Report: | |||
| Attachments: | alternative patch | ||
|
Description
Ruslan Kabatsayev
2011-11-18 16:09:08 UTC
interestingly, can't reproduce with gimp-2.6.11 See: http://wstaw.org/m/2011/11/18/plasma-desktoppN2811.jpg Checking our code, I wonder if there is not some "custom" widget involved on the gimp side: a GtkFrame, with a label, and with gtk_frame_get_shadow_type == GTK_SHADOW_OUT even though it is not rendered. (that's the only way a parent "groupbox" could be detected). Could that be ? Yeah, shadow_type is OUT, and the labels are present. But we can blacklist GimpFrame, which is the name for this widget (returned by G_OBJECT_TYPE_NAME()). I think I'll push this. Git commit 7283bcde2d8463caf16403063a8aad0e4b88d4fd by Ruslan Kabatsayev. Committed on 18/11/2011 at 19:17. Pushed by kabatsayev into branch 'master'. Don't consider GimpFrame to be group box CCBUG: 286933 M +2 -1 src/oxygengtkutils.h http://commits.kde.org/oxygen-gtk/7283bcde2d8463caf16403063a8aad0e4b88d4fd The above fixes it. ok. ... mmm. So I'm not so much in favor of yet another blacklist :( Maybe, we can actually better tag groupboxes, as frames that satisfy the above conditions, and that have (at least once) been actually rendered as groupboxes... Since I have gimp2.4 locally compiled here, and can reproduce the issue, I'll see if I can do that. (this could possibly fix other similar bugs). Created attachment 65828 [details]
alternative patch
this patch uses the existing (but unused so far) Oxygen::GroupBoxEngine to keep track of flagged groupboxes for which the frame is actually rendered at least once.
This is then used to check on renderGroupBackground.
Ruslan: could you test in place of your change ?
Double check for regressions in other places ?
So far I've found that
- it does fix the issue with gimp-2.4
- it does not introduce regressions
Since it is more generic than the previous patch, and can possibly fix other issues, I'd rather revert your change and commit + backport this change.
Comments ?
ok. In fact after double checking, this strategy is already used (succesfully) four groupbox *holes*. So if it works here, should work elsewhere. Will push. Yes, it works, and in fact is a nicer way to fix this. In fact, I wanted to suggest similar idea, but thought it'd take too much work for a single bug (didn't know/think of existing infrastructure). Git commit 6c93357f1c44ef76f9382465c214d87d496f0c4d by Hugo Pereira Da Costa. Committed on 18/11/2011 at 19:39. Pushed by hpereiradacosta into branch 'master'. check that parent groupbox is registered to engine before rendering groupbox background. CCBUG: 286933 M +6 -13 src/oxygenstyle.cpp M +1 -0 src/oxygenstylewrapper.cpp http://commits.kde.org/oxygen-gtk/6c93357f1c44ef76f9382465c214d87d496f0c4d Git commit b749ebef93ce808fb89376f22ff3bb1dda72c482 by Hugo Pereira Da Costa. Committed on 18/11/2011 at 19:39. Pushed by hpereiradacosta into branch '1.1'. check that parent groupbox is registered to engine before rendering groupbox background. CCBUG: 286933 M +6 -13 src/oxygenstyle.cpp M +1 -0 src/oxygenstylewrapper.cpp http://commits.kde.org/oxygen-gtk/b749ebef93ce808fb89376f22ff3bb1dda72c482 |
