Community
Participate
Working Groups
Deleting git branches in the History view, the UI became frozen. Running Photon Build id: I20180220-2000 org.eclipse.swt.gtk.linux.x86_64 (3.107.0.v20180220-1901) "main" #1 prio=6 os_prio=0 tid=0x00007f587000a800 nid=0x74df runnable [0x00007f5876e5e000] java.lang.Thread.State: RUNNABLE at org.eclipse.swt.internal.gtk.OS._g_main_context_iteration(Native Method) at org.eclipse.swt.internal.gtk.OS.g_main_context_iteration(OS.java:1559) at org.eclipse.swt.widgets.Shell.setVisible(Shell.java:2457) at org.eclipse.swt.widgets.Shell.open(Shell.java:1700) at org.eclipse.jface.window.Window.open(Window.java:790) at org.eclipse.jface.dialogs.ProgressMonitorDialog.open(ProgressMonitorDialog.java:646) at org.eclipse.jface.dialogs.ProgressMonitorDialog.aboutToRun(ProgressMonitorDialog.java:555) at org.eclipse.jface.dialogs.ProgressMonitorDialog.run(ProgressMonitorDialog.java:480) at org.eclipse.egit.ui.internal.history.command.DeleteBranchOnCommitHandler.execute(DeleteBranchOnCommitHandler.java:86) at org.eclipse.ui.internal.handlers.HandlerProxy.execute(HandlerProxy.java:291) at org.eclipse.ui.internal.handlers.E4HandlerProxy.execute(E4HandlerProxy.java:93) at sun.reflect.GeneratedMethodAccessor60.invoke(Unknown Source) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:498) at org.eclipse.e4.core.internal.di.MethodRequestor.execute(MethodRequestor.java:55) at org.eclipse.e4.core.internal.di.InjectorImpl.invokeUsingClass(InjectorImpl.java:317) at org.eclipse.e4.core.internal.di.InjectorImpl.invoke(InjectorImpl.java:251) at org.eclipse.e4.core.contexts.ContextInjectionFactory.invoke(ContextInjectionFactory.java:161) at org.eclipse.e4.core.commands.internal.HandlerServiceHandler.execute(HandlerServiceHandler.java:152) at org.eclipse.core.commands.Command.executeWithChecks(Command.java:494) at org.eclipse.core.commands.ParameterizedCommand.executeWithChecks(ParameterizedCommand.java:487) at org.eclipse.e4.core.commands.internal.HandlerServiceImpl.executeHandler(HandlerServiceImpl.java:210) at org.eclipse.e4.ui.workbench.renderers.swt.HandledContributionItem.executeItem(HandledContributionItem.java:431) at org.eclipse.e4.ui.workbench.renderers.swt.AbstractContributionItem.handleWidgetSelection(AbstractContributionItem.java:446) at org.eclipse.e4.ui.workbench.renderers.swt.AbstractContributionItem.lambda$2(AbstractContributionItem.java:472) at org.eclipse.e4.ui.workbench.renderers.swt.AbstractContributionItem$$Lambda$250/1124683276.handleEvent(Unknown Source) at org.eclipse.swt.widgets.EventTable.sendEvent(EventTable.java:86) at org.eclipse.swt.widgets.Display.sendEvent(Display.java:5632) at org.eclipse.swt.widgets.Widget.sendEvent(Widget.java:1348) at org.eclipse.swt.widgets.Display.runDeferredEvents(Display.java:4886) at org.eclipse.swt.widgets.Display.readAndDispatch(Display.java:4467) at org.eclipse.e4.ui.internal.workbench.swt.PartRenderingEngine$5.run(PartRenderingEngine.java:1150) at org.eclipse.core.databinding.observable.Realm.runWithDefault(Realm.java:336) at org.eclipse.e4.ui.internal.workbench.swt.PartRenderingEngine.run(PartRenderingEngine.java:1039) at org.eclipse.e4.ui.internal.workbench.E4Workbench.createAndRunUI(E4Workbench.java:153) at org.eclipse.ui.internal.Workbench.lambda$3(Workbench.java:681) at org.eclipse.ui.internal.Workbench$$Lambda$75/551797833.run(Unknown Source) at org.eclipse.core.databinding.observable.Realm.runWithDefault(Realm.java:336) at org.eclipse.ui.internal.Workbench.createAndRunWorkbench(Workbench.java:595) at org.eclipse.ui.PlatformUI.createAndRunWorkbench(PlatformUI.java:148) at org.eclipse.ui.internal.ide.application.IDEApplication.start(IDEApplication.java:152) at org.eclipse.equinox.internal.app.EclipseAppHandle.run(EclipseAppHandle.java:196) at org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.runApplication(EclipseAppLauncher.java:134) at org.eclipse.core.runtime.internal.adaptor.EclipseAppLauncher.start(EclipseAppLauncher.java:104) at org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:388) at org.eclipse.core.runtime.adaptor.EclipseStarter.run(EclipseStarter.java:243) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:498) at org.eclipse.equinox.launcher.Main.invokeFramework(Main.java:656) at org.eclipse.equinox.launcher.Main.basicRun(Main.java:592) at org.eclipse.equinox.launcher.Main.run(Main.java:1498) at org.eclipse.equinox.launcher.Main.main(Main.java:1471)
Doesn't occur on my system. Fedora 27, Gtk3.22, X11, SWT repo. Eclipse SDK Version: Photon (4.8) Build id: I20180228-2000 OS: Linux, v.4.15.6-300.fc27.x86_64, x86_64 / gtk 3.22.26, WebKit 2.18.6 I.e, create a branch. Select it and delete it. All works. What are your system info? Os/Gtk(3?) version/gtk3 theme. Which repo are you using.? is it a big one?
Hi Leo, org.eclipse.swt.internal.gtk.theme=Ambiance org.eclipse.swt.internal.gtk.version=3.10.8 org.eclipse.swt.internal.webkitgtk.version=2.4.10 os.arch=amd64 os.name=Linux os.version=3.13.0-141-generic I was deleting many local branches from org.eclipse.tracecompass repo, one at a time. The freeze happened after about the 50th delete. Is there any additional info that would be helpful to gather if it happens again?
(In reply to Patrick Tasse from comment #2) > Hi Leo, > > org.eclipse.swt.internal.gtk.theme=Ambiance > org.eclipse.swt.internal.gtk.version=3.10.8 > org.eclipse.swt.internal.webkitgtk.version=2.4.10 > os.arch=amd64 > os.name=Linux > os.version=3.13.0-141-generic > > I was deleting many local branches from org.eclipse.tracecompass repo, one > at a time. The freeze happened after about the 50th delete. > > Is there any additional info that would be helpful to gather if it happens > again? Well, as irony has it, the stack trace indicates that you're stuck in a workaround that was designed to prevent U.I dead locks that occured when interfacing with non-swt stuff (AWT etc..) :-/. It's waiting for a 'map-event' signal which never comes. Map-event occurs when a widget (in this case the shell) was invisible and then becomes visible. Soooo.. things on my mind: 1) Can we get rid of the workaround. 2) If not, is this a gtk3.10 specific issue or are all gtk affected (considering I haven't reproduced it on my system yet). (stable gtk3 is gtk3.22. Versions 3.0-3.20 were WIP). a) Do you have access to something that runs Gtk3.22, can you test on that? b) Can you test with "Adwaita" theme (if on Gtk3.22?). i) Btw, are you running Ubuntu here?
I'm running Ubuntu 14.04 with Gnome shell, but I should be upgrading to 16.04/GTK 3.18.9 in the coming weeks. Maybe we can close this for now and if it happens again on a more recent version I can re-open it. Or would 3.18.9 also be too old to bother?
(In reply to Patrick Tasse from comment #4) > I'm running Ubuntu 14.04 with Gnome shell, but I should be upgrading to > 16.04/GTK 3.18.9 in the coming weeks. > > Maybe we can close this for now and if it happens again on a more recent > version I can re-open it. Or would 3.18.9 also be too old to bother? Gtk3.18 had a memory leak :-/. just fyi... But still a step up from 3.10 I guess. Gtk3.20 introduced some significant (api breaking) theaming changes, so version 3.18 was (sort of) the last not-very-stable build. Gtk3.20 and Gtk3.22 are fairly stable. I have a mechanism by which I can run snippets/child eclipses in older versions of gtk. (all the way back to 3.4). If something is straight-forward enough to test, I can usually test if it's a gtk-version specific issue (theme issues aside). But this one seems a bit tricky to reproduce. Since I can't seem to reproduce this issue and you can't test things on gtk3.22(ish), we're a bit stuck :-|. If possible, I would try to consider upgrading all the way to gtk3.22 and if the issue still occurs then maybe try to find detailed steps to reproduce. If that's not possible, then at the moment known gtk3/Wayland bugs have more impact/priority than corner cases on older gtk versions (e.g TreeCheese issue) and yes we might have to delay until you have the chance to test things on gtk3.22ish. Thank you for bug-submission, they are always very useful.
FYI, I just got another freeze in Tracker.open() in a similar workaround, while doing a drag start of an editor tab. Same old GTK :) "main" #1 prio=6 os_prio=0 tid=0x00007f384000a800 nid=0x21f4 runnable [0x00007f3847422000] java.lang.Thread.State: RUNNABLE at org.eclipse.swt.internal.gtk.OS._g_main_context_iteration(Native Method) at org.eclipse.swt.internal.gtk.OS.g_main_context_iteration(OS.java:1559) at org.eclipse.swt.widgets.Tracker.open(Tracker.java:809) at org.eclipse.e4.ui.workbench.addons.dndaddon.DnDManager.startDrag(DnDManager.java:225) at org.eclipse.e4.ui.workbench.addons.dndaddon.DnDManager.lambda$0(DnDManager.java:81) at org.eclipse.e4.ui.workbench.addons.dndaddon.DnDManager$$Lambda$169/1040733616.dragDetected(Unknown Source) at org.eclipse.swt.widgets.TypedListener.handleEvent(TypedListener.java:128) at org.eclipse.swt.widgets.EventTable.sendEvent(EventTable.java:86) at org.eclipse.swt.widgets.Display.sendEvent(Display.java:5632) at org.eclipse.swt.widgets.Widget.sendEvent(Widget.java:1348) at org.eclipse.swt.widgets.Display.runDeferredEvents(Display.java:4886) at org.eclipse.swt.widgets.Display.readAndDispatch(Display.java:4467)
(In reply to Patrick Tasse from comment #6) > FYI, I just got another freeze in Tracker.open() in a similar workaround, > while doing a drag start of an editor tab. Same old GTK :) > > "main" #1 prio=6 os_prio=0 tid=0x00007f384000a800 nid=0x21f4 runnable > [0x00007f3847422000] > java.lang.Thread.State: RUNNABLE > at org.eclipse.swt.internal.gtk.OS._g_main_context_iteration(Native Method) > at org.eclipse.swt.internal.gtk.OS.g_main_context_iteration(OS.java:1559) > at org.eclipse.swt.widgets.Tracker.open(Tracker.java:809) > at > org.eclipse.e4.ui.workbench.addons.dndaddon.DnDManager.startDrag(DnDManager. > java:225) > at > org.eclipse.e4.ui.workbench.addons.dndaddon.DnDManager.lambda$0(DnDManager. > java:81) > at > org.eclipse.e4.ui.workbench.addons.dndaddon.DnDManager$$Lambda$169/ > 1040733616.dragDetected(Unknown Source) > at org.eclipse.swt.widgets.TypedListener.handleEvent(TypedListener.java:128) > at org.eclipse.swt.widgets.EventTable.sendEvent(EventTable.java:86) > at org.eclipse.swt.widgets.Display.sendEvent(Display.java:5632) > at org.eclipse.swt.widgets.Widget.sendEvent(Widget.java:1348) > at org.eclipse.swt.widgets.Display.runDeferredEvents(Display.java:4886) > at org.eclipse.swt.widgets.Display.readAndDispatch(Display.java:4467) I recall last time I tested on Ubuntu VM and older Gtk, I think I had similar issues. Works better on newer Gtk & gnome versions I think. Are you using Unity btw or did you hop onto Gnome yet?
I am seeing the same stack trace in a hang during our internal UI test runs on Ubuntu. Seems to be the same problem.
I'm running latest Ubuntu 16.04.
And running in Xvfb with metacity if that matters. GTK version: libgtk-3-0/xenial-updates,now 3.18.9-1ubuntu3.3 amd64 [installed,automatic]
Oh, and the stack trace I have is for the ProgressMonitorDialog open, not the drag and drop. I imagine 16.04, the latest LTS release is still widely in use and will be for a while. I get your focus on Red Hatty things, which is too bad we don't have more Ubuntu folk contributing.
(In reply to Doug Schaefer from comment #11) > Oh, and the stack trace I have is for the ProgressMonitorDialog open, not > the drag and drop. > > I imagine 16.04, the latest LTS release is still widely in use and will be > for a while. I get your focus on Red Hatty things, which is too bad we don't > have more Ubuntu folk contributing. I'd like for eclipse to be as smooth on Linux as possible on as many Linux distros as possible, but sometimes bugs occur due to customization of the gtk shipped with Ubuntu, which is not something that can be fixed on the SWT side (or at least not easily). Also some window managers don't get updated as gtk moves along, so window manager-specific bugs are kinda hard to nail down also. For example some themes are buggy (oxygen causes hard-crashes for instance). Due to so many different things that can break, Me/Eric/Alex tend to stick with stock Gtk3 & stock display manager (gnome3) and stock theme (Adwaita). I mean Ubuntu did move toward Gnome3 probably because of this very reason. Once Eclipse blockers/Tree/performance issues are fully resolved on stock systems, then we will probably start looking into platform-specific issues. E.g: 516839 – (Gtk3ImprovementTracker) [GTK3][Tracker] Improve support for newer versions of GTK+ 3 https://bugs.eclipse.org/bugs/show_bug.cgi?id=516839 I understand the importance of distribution-specific issues. If you can make a compelling case as to why it would be more important then Gtk3-blockers that affect all distors, then we would consider spending more time on them :-). Otherwise patches are always welcomed with open arms, we're always happy to review them. With that said, could you guys test with Ubuntu 17.10.1 & Gnome:? https://www.ubuntu.com/download/desktop & Stock theme. See if issue still occurs for you & post your findings.
Works fine on 17.10. 16.04 LTS continues to hang consistently, not always but almost.
Mine are also always with the ProgressMonitorDialog openning up.
Unfortunately, the only compelling argument I can make is that this will affect adoption in the commercial world. Almost every company I've worked with use Ubuntu and most are conservative and are slow to adopt newer versions. This will affect adoption of Eclipse Photon (I'm already working this with my management). Wish I had GTK experience to help and appreciate any help you can give.
(In reply to Doug Schaefer from comment #15) > Unfortunately, the only compelling argument I can make is that this will > affect adoption in the commercial world. Almost every company I've worked > with use Ubuntu and most are conservative and are slow to adopt newer > versions. This will affect adoption of Eclipse Photon (I'm already working > this with my management). Wish I had GTK experience to help and appreciate > any help you can give. Ubuntu 16.04 LTS uses a GTK3 stack that is almost 3 years old (GTK3.18). Ubuntu 18.04 LTS is set to be released on April 26th and uses the same GTK3 as 17.10. From SWT's POV we don't have the manpower to support versions of GTK3 that are already obsolete and do not even have stable API. GTK3.22 has been out for over a year and has been marked as the stable and final release of GTK3 -- with the limited development resources we have this is our best shot at stabilizing GTK3 ASAP.
I understand Eric. Wish we had better numbers on how many users will be affected by this. I'm just worried it will be very high. At the very least, it'll be most of my internal users here since I don't see our IT department upgrading the corporate image any time soon.
(In reply to Doug Schaefer from comment #17) > I understand Eric. Wish we had better numbers on how many users will be > affected by this. I'm just worried it will be very high. At the very least, > it'll be most of my internal users here since I don't see our IT department > upgrading the corporate image any time soon. I think the real-world solution may be to use eclipse with Gtk2 backend until people upgrade to something that use gtk3.20+ (gtk3.22 preferred). export SWT_GTK3=0 Although we may be in the process of removing gtk2 support, so the above solution won't work for that long before gtk2 support is removed. It would be nice to support users running older gtk versions, unfortunately we currently don't have the man power at the moment. Also sometimes if we want to, the issue is in gtk itself, so there isn't that much that we can fix at SWT level. We're mostly just wrapping around native gtk calls. So realistically, I think this bug is not going to be resolved at Gtk3 level. It could be worked around by using gtk2 on Ubuntu 16 or upgrading to newer Ubuntu with newer gtk3. Should we close as worksforme or is anyone interested in working on this?
No, please leave this open. I'm still hoping someone will come and save the day. And besides, that way you have a bug you can point the duplicates at as they come in. ;)
(In reply to Doug Schaefer from comment #19) > No, please leave this open. I'm still hoping someone will come and save the > day. And besides, that way you have a bug you can point the duplicates at as > they come in. ;) Well. It's gonna be one of those bugs stuck around for 5 years till someone re-visits and closes it due to old age. Setting appropriate priority & keywords.
(In reply to Leo Ufimtsev from comment #20) > (In reply to Doug Schaefer from comment #19) > > No, please leave this open. I'm still hoping someone will come and save the > > day. And besides, that way you have a bug you can point the duplicates at as > > they come in. ;) > > Well. It's gonna be one of those bugs stuck around for 5 years till someone > re-visits and closes it due to old age. > > Setting appropriate priority & keywords. Thanks. Again, and I could be wrong on this, but I don't think customers of commercial ISV's will let it sit that long.
(In reply to Doug Schaefer from comment #21) > Thanks. Again, and I could be wrong on this, but I don't think customers of > commercial ISV's will let it sit that long. I've set my self a reminder to check back on this bug in 5 years time :-). MfG.
Got one coming. https://bugs.eclipse.org/bugs/show_bug.cgi?id=535326 Though it hangs in Pango. But on Ubuntu 16.04. We should really come out with a statement that Photon will not work on 16.04 and make sure everyone knows.
Quick out of the loop question, coming from someone who will soon _UPGRADE_ to 16.04, how hard would it be to install gtk3.20 on the side and ld_preload? I didn't try yet.
(In reply to Matthew Khouzam from comment #24) > Quick out of the loop question, coming from someone who will soon _UPGRADE_ > to 16.04, how hard would it be to install gtk3.20 on the side and > ld_preload? I didn't try yet. Not sure, there was a user in another ticket that said he did it, but I can't recall if he upgraded to 18.04 or not. You might be able to find a PPA that will allow you to do it, but it might break other stuff on your system. (In reply to Doug Schaefer from comment #23) > Got one coming. https://bugs.eclipse.org/bugs/show_bug.cgi?id=535326 > > Though it hangs in Pango. But on Ubuntu 16.04. > > We should really come out with a statement that Photon will not work on > 16.04 and make sure everyone knows. I agree and we are looking into this.
(In reply to Eric Williams from comment #25) > (In reply to Matthew Khouzam from comment #24) > > Quick out of the loop question, coming from someone who will soon _UPGRADE_ > > to 16.04, how hard would it be to install gtk3.20 on the side and > > ld_preload? I didn't try yet. > > Not sure, there was a user in another ticket that said he did it, but I > can't recall if he upgraded to 18.04 or not. > > You might be able to find a PPA that will allow you to do it, but it might > break other stuff on your system. > Yes, I am hoping we can have the two gtks running in parallel and I like living dangerously like that. Remember, 16.04 is LTS so long term pain in the ankle. :( I will try this tonight: https://launchpad.net/~jonathonf/+archive/ubuntu/gtk3.24 Maybe a docker/flatpak may be the way to at least let users have a non-offputting experience with eclipse. Also, according to Linode, 48% of Linux installs are Ubuntu. Within those, most should now be LTS since if you just regularly upgrade, you'll end up being in LTS unless you explicitly opt out.
(In reply to Matthew Khouzam from comment #26) > (In reply to Eric Williams from comment #25) > > (In reply to Matthew Khouzam from comment #24) > > > Quick out of the loop question, coming from someone who will soon _UPGRADE_ > > > to 16.04, how hard would it be to install gtk3.20 on the side and > > > ld_preload? I didn't try yet. > > > > Not sure, there was a user in another ticket that said he did it, but I > > can't recall if he upgraded to 18.04 or not. > > > > You might be able to find a PPA that will allow you to do it, but it might > > break other stuff on your system. > > > > Yes, I am hoping we can have the two gtks running in parallel and I like > living dangerously like that. Remember, 16.04 is LTS so long term pain in > the ankle. :( > I will try this tonight: > https://launchpad.net/~jonathonf/+archive/ubuntu/gtk3.24 There is no GTK3.24 (at least not released) so this is either something else or the dev branch. You might have some moderate success with compiling GTK3.22, it doesn't take too long. You can then feed the .so files to Eclipse/SWT using LD_LIBRARY_PATH. > Maybe a docker/flatpak may be the way to at least let users have a > non-offputting experience with eclipse. There are efforts under way to provide this.
(In reply to Eric Williams from comment #27) > There is no GTK3.24 I stand corrected, apparently there is: https://blog.gtk.org/2018/06/23/a-gtk-3-update/
Does anyone still care about this issue? 18.04 is out for quite some time.
Well, the freeze still happened to me this week on Ubuntu 14.04 with gtk/3.10.8, but our customers are on RHEL 6.8 with gtk/3.16.0 installed (but running Eclipse with gtk/2.24.23). I'm sorry I still did not have time to upgrade my system to Ubuntu 16.04 with gtk/3.18.9 ;(
Note that in we are moving to Gtk 3.14 as min version so this one will probably have to be closed soon.
(In reply to Alexander Kurtakov from comment #31) > Note that in we are moving to Gtk 3.14 as min version so this one will > probably have to be closed soon. 16.04 has 3.18 so it will stay open just a little while longer. 14.04 support will definitely be gone though.
Swt requires GTK 3.20+ now (Bug 561047 ) . Closing.
