Community
Participate
Working Groups
+++ This bug was initially created as a clone of Bug #460744 +++ The original fix for bug 421127 for CTabFolder rendering issues on GTK3 caused bug 460744 on all platforms which was a performance issue (adding an item to CTabFolder forces a relayout on ALL siblings of the CTabFolder). The performance problem has been fixed for all platforms including GTK+ now, but we should investigate if a fix that is not in CTabFolder code (that is common to all platforms) can be found, this bug will track the same.
A related CTabFolder bug 477798, which attempts to fix the root cause of CTabFolder initial size issue for all platforms via below gerrit patch: https://git.eclipse.org/r/#/c/56655/ Once above changes gets in, it seems we can also get rid of GTK only code which forces a re-layouting call on parent control of CTabFolder. Will share a gerrit for the same.
New Gerrit change created: https://git.eclipse.org/r/57479
Gerrit change https://git.eclipse.org/r/57479 was merged to [master]. Commit: http://git.eclipse.org/c/platform/eclipse.platform.swt.git/commit/?id=d8926d5b1b1025f39ea5c3556b6cea2310403642
Tested the patch on Ubuntu14 with GTK3.10, pushed the changes to master.
This bug has caused bug 487757. Should we revert it?
(In reply to Alexander Kurtakov from comment #5) > This bug has caused bug 487757. Should we revert it? I would say yes, since this is a regression for GTK3 MultiPageEditor's, but I'm not SWT committer to vote for it :-).
(In reply to Niraj Modi from comment #1) > A related CTabFolder bug 477798, which attempts to fix the root cause of > CTabFolder initial size issue for all platforms via below gerrit patch: > https://git.eclipse.org/r/#/c/56655/ > > Once above changes gets in, it seems we can also get rid of GTK only code > which forces a re-layouting call on parent control of CTabFolder. Hi Arun, The performance optimization that we did, to get rid of the GTK only code which forces a re-layouting call on parent control of CTabFolder.. has caused bug 487757 Going by observation by Andrey in comment 6, we should revert this GTK optimization CTabFolder in NeonRC2 ??
I noticed this re-opened bug was still targeted to Neon M3, but the last comment mentions Neon RC2, so I did not want it to "get lost". Please retarget if I have misunderstood your plans for this bug.
(In reply to Niraj Modi from comment #7) > (In reply to Niraj Modi from comment #1) > > A related CTabFolder bug 477798, which attempts to fix the root cause of > > CTabFolder initial size issue for all platforms via below gerrit patch: > > https://git.eclipse.org/r/#/c/56655/ > > > > Once above changes gets in, it seems we can also get rid of GTK only code > > which forces a re-layouting call on parent control of CTabFolder. > > Hi Arun, > The performance optimization that we did, to get rid of the GTK only code > which forces a re-layouting call on parent control of CTabFolder.. has > caused bug 487757 > > Going by observation by Andrey in comment 6, we should revert this GTK > optimization CTabFolder in NeonRC2 ?? We should ideally be fixing the actual root cause of bug 487757 instead of reverting this fix as any fix in CTabFolder for the GTK3 specific issue is only temporary and should not be necessary in the first place. Since Eric is attempting to fix the root cause via bug 487160, I think we should leave this one as it is and not revert any changes.
Moving back to RESOLVED state as bug 487160 has a patch that is targeted for RC2 and we may not need to make any changes regarding this bug.
(In reply to Arun Thondapu from comment #10) > Moving back to RESOLVED state as bug 487160 has a patch that is targeted for > RC2 and we may not need to make any changes regarding this bug. Reopening. Bug 487160 deals with a different issue and we don't have a fix for bug 487757 yet. Since bug 487757 affects usability with GTK3, its important to fix that one even if it means reverting the change for this bug and causing a performance impact for GTK3.
We should find a proper fix for bug 487757 in 4.6.1 which does not depend on this workaround in CTabFolder. Targeting this bug to 4.6.1 accordingly.
(In reply to Arun Thondapu from comment #12) > We should find a proper fix for bug 487757 in 4.6.1 which does not depend on > this workaround in CTabFolder. Targeting this bug to 4.6.1 accordingly. Hello Arun, You pointed it out well. The underlying bug is somewhere in ScrolledComposite, CTabFolder was just masking an underlying fault in SWT/Gtk3. I'll work on Bug 487757 now / this week to see if we can fix the underlying cause rather than revering CTabFolder fix. Please hang on.
Moving this bug back to RESOLVED state as reverting this fix doesn't actually seem to fix bug 487757 and therefore is not necessary.
Opened Bug 512270 to remove the now unsed IS_GTK flag.