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

Bug 449444

Summary:	[GTK/Linux] syslog filled with window manager errors on Fedora
Product:	[Eclipse Project] Platform	Reporter:	Ferry Huberts <mailings>
Component:	SWT	Assignee:	Platform-SWT-Inbox <platform-swt-inbox>
Status:	RESOLVED WORKSFORME	QA Contact:
Severity:	normal
Priority:	P3	CC:	ericwill, lufimtse, pierre-charles.david
Version:	4.4.1	Keywords:	triaged
Target Milestone:	---
Hardware:	PC
OS:	Linux
Whiteboard:

Description Ferry Huberts

2014-10-31 05:37:04 EDT

Linux Fedora 20 x86_64
GNOME
Eclipse Luna SR1

My syslog is filled with the window manager reporting that Eclipse is doing things wrong:


Oct 31 10:29:47 stinkpad gnome-session: Window manager warning: Buggy client sent a _NET_ACTIVE_WINDOW message with a timestamp of 0 for 0x408e232 (Clean  )
Oct 31 10:29:47 stinkpad gnome-session: Window manager warning: meta_window_activate called by a pager with a 0 timestamp; the pager needs to be fixed.
Oct 31 10:29:47 stinkpad gnome-session: Window manager warning: Buggy client sent a _NET_ACTIVE_WINDOW message with a timestamp of 0 for 0x40005b1 (C/C++ - nb)
Oct 31 10:29:47 stinkpad gnome-session: Window manager warning: meta_window_activate called by a pager with a 0 timestamp; the pager needs to be fixed.

Comment 1 Paul Webster

2014-11-07 08:48:20 EST

Where does it reference eclipse?  Or those only happen when eclipse is open and disappear when eclipse is closed?

PW

Comment 2 Ferry Huberts

2014-11-07 08:56:00 EST

(In reply to Paul Webster from comment #1)
> Where does it reference eclipse?  Or those only happen when eclipse is open
> and disappear when eclipse is closed?
> 

Yes.
And they end up in the log as a direct result of me doing things with Eclipse.
This also seems to affect useability of Eclipse because sometimes windows/panes do not respond/repaint at all, I have to selected other windows/panes or scroll other windows/panes before the non-responding window/pane responds/repaints.

Rather annoying

Comment 3 Leo Ufimtsev

2014-11-07 10:44:23 EST

Hello Ferry, 

Based on this article:
http://www.linuxhomenetworking.com/wiki/index.php/Quick_HOWTO_:_Ch05_:_Troubleshooting_Linux_with_syslog#.VFzlGHVdWV4

I looked into my own /var/log/messages but could not find such messages.

I am running Eclipse 4.4.1 (Luna) on Fedora 20. 

Are the messages found in your '/var/log/messages' file? 

Do you have your Eclipse customized with plugins? Could you try to run a vanilla Eclipse downloaded from the site and see if the messages are still logged? 

Do you have any special themes applied to your Gnome (oxygen)?
If so, do the messages still occur under the 'Adwaita' theame? (themes can be changed with TweakTool).

Does this occur if you have 'only' Eclipse running with no other applications? (e.g Bluejeans kills Eclipse atm).  

I also noticed 'stinkpad' in your log. In my logs it is not included:
Nov  6 16:36:05 unused-10-15-18-183 gnome-session:
Do you know what that means by any chance?

Comment 4 Ferry Huberts

2014-11-07 10:54:21 EST

(In reply to Leo Ufimtsev from comment #3)
> Hello Ferry, 
> 
> Based on this article:
> http://www.linuxhomenetworking.com/wiki/index.php/Quick_HOWTO_:_Ch05_:
> _Troubleshooting_Linux_with_syslog#.VFzlGHVdWV4
> 
> I looked into my own /var/log/messages but could not find such messages.
> 
> I am running Eclipse 4.4.1 (Luna) on Fedora 20. 

I'm NOT running the Fedora package for Eclipse.
My Eclipse is downloaded from eclipse.org

> 
> Are the messages found in your '/var/log/messages' file? 

yes

> 
> Do you have your Eclipse customized with plugins? Could you try to run a
> vanilla Eclipse downloaded from the site and see if the messages are still
> logged? 
> 

will try later

> Do you have any special themes applied to your Gnome (oxygen)?
> If so, do the messages still occur under the 'Adwaita' theame? (themes can
> be changed with TweakTool).

no, entirely default

> 
> Does this occur if you have 'only' Eclipse running with no other
> applications? (e.g Bluejeans kills Eclipse atm).  
> 
> I also noticed 'stinkpad' in your log. In my logs it is not included:
> Nov  6 16:36:05 unused-10-15-18-183 gnome-session:
> Do you know what that means by any chance?

stinkpad is the name of my machine

Comment 5 Leo Ufimtsev

2014-11-26 10:13:31 EST

Hello Ferry, 

>> Do you have your Eclipse customized with plugins? Could you try to run a
>> vanilla Eclipse downloaded from the site and see if the messages are still
>> logged? 

>will try later

How are things :-)?

Comment 6 Ferry Huberts

2014-11-26 11:04:27 EST

(In reply to Leo Ufimtsev from comment #5)
> Hello Ferry, 
> 
> >> Do you have your Eclipse customized with plugins? Could you try to run a
> >> vanilla Eclipse downloaded from the site and see if the messages are still
> >> logged? 
> 
> >will try later
> 
> How are things :-)?

Thanks for pinging me, completely forgot.

So, downloaded clean Eclipse from the site, started it and immediately saw the same warnings.
Looks like it really is in the (SWT) core.

Comment 7 Leo Ufimtsev

2014-11-27 10:08:54 EST

Hello, 

I've been trying to reproduce it with my Fedora 21/Eclipse Luna, but I don't get these messages, at least not from Eclipse. I get some "net_active_window" messages from other applications on my machine.

Could you move/clear your log. Close down all applications. Then open eclipse for sometime until there are log entries. 

Then attach the full log here? (also please let me know how long eclipse ran for until entries poped up and ~things done that triggered the errors).

thnks.

Comment 8 Alexander Kurtakov

2017-04-20 18:36:52 EDT

Is this bug still reproducible with Neon or Oxygen?

Comment 9 Ferry Huberts

2017-04-21 03:39:55 EDT

(In reply to Alexander Kurtakov from comment #8)
> Is this bug still reproducible with Neon or Oxygen?

yes, but different error are logged.
I'm now on Fedora 25 x64

A short copy&paste from the log:

Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad gnome-terminal-[7107]: Allocating size to GtkScrollbar 0x56395b290310 without calling gtk_widget_get_preferred_width/height(). How does the code know the size to allocate?
Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:43 stinkpad java[16797]: Allocating size to GtkBox 0x7f008e520170 without calling gtk_widget_get_preferred_width/height(). How does the code know the size to allocate?
Apr 21 09:37:45 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:45 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:45 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:45 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:46 stinkpad java[16797]: Allocating size to GtkBox 0x7f008e520170 without calling gtk_widget_get_preferred_width/height(). How does the code know the size to allocate?
Apr 21 09:37:46 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:47 stinkpad java[16797]: Allocating size to GtkScrolledWindow 0x7f008d5e9a60 without calling gtk_widget_get_preferred_width/height(). How does the code know the size to allocate?
Apr 21 09:37:47 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:47 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:47 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:47 stinkpad java[16797]: Negative content height -3 (allocation 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:47 stinkpad java[16797]: Allocating size to GtkBox 0x7f008e520170 without calling gtk_widget_get_preferred_width/height(). How does the code know the size to allocate?
Apr 21 09:37:47 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:47 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:47 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
Apr 21 09:37:47 stinkpad java[16797]: Negative content width -1 (allocation 1, extents 1x1) while allocating gadget (node trough, owner GtkProgressBar)
Apr 21 09:37:47 stinkpad java[16797]: Negative content width -1 (allocation 1, extents 1x1) while allocating gadget (node trough, owner GtkProgressBar)
Apr 21 09:37:48 stinkpad java[16797]: Allocating size to GtkScrolledWindow 0x7f008d5e93e0 without calling gtk_widget_get_preferred_width/height(). How does the code know the size to allocate?
Apr 21 09:37:48 stinkpad java[16797]: Negative content width -6 (allocation 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)

Comment 10 Leo Ufimtsev

2017-04-21 09:05:43 EDT

It's probably because we have a custom container that doesn't get the size of all of it's children properly because we do manual layout in Java.

Fixing the warnings would probably be quite a bit of effort, but we should do this at some point. (As well as removing build warnings from build.sh).

At the moment there are a number of rather pressing issues (wayland,webkit2 ports), performance, functional and stability fixes that need to be done first, but at some point we should fix this guy also.

Comment 11 Pierre-Charles David

2017-07-27 04:26:24 EDT

(In reply to Ferry Huberts from comment #9)
> (In reply to Alexander Kurtakov from comment #8)
> > Is this bug still reproducible with Neon or Oxygen?
> 
> yes, but different error are logged.
> I'm now on Fedora 25 x64
> 
> A short copy&paste from the log:
> 
> Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation
> 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
> Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation
> 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)

I've just been hit by these, and it seems this is tracked by https://bugs.eclipse.org/bugs/show_bug.cgi?id=486068 and is specific to recent Gtk version (> 3.20, I'm on 3.22 myself). This is different from the _NET_ACTIVE_WINDOW messages.

Comment 12 Eric Williams

2017-07-27 10:05:32 EDT

(In reply to Pierre-Charles David from comment #11)
> (In reply to Ferry Huberts from comment #9)
> > (In reply to Alexander Kurtakov from comment #8)
> > > Is this bug still reproducible with Neon or Oxygen?
> > 
> > yes, but different error are logged.
> > I'm now on Fedora 25 x64
> > 
> > A short copy&paste from the log:
> > 
> > Apr 21 09:37:43 stinkpad java[16797]: Negative content height -3 (allocation
> > 1, extents 2x2) while allocating gadget (node toolbar, owner GtkToolbar)
> > Apr 21 09:37:43 stinkpad java[16797]: Negative content width -6 (allocation
> > 1, extents 4x3) while allocating gadget (node toolbar, owner GtkToolbar)
> 
> I've just been hit by these, and it seems this is tracked by
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=486068 and is specific to
> recent Gtk version (> 3.20, I'm on 3.22 myself). This is different from the
> _NET_ACTIVE_WINDOW messages.

If the NET_ACTIVE_WINDOW messages are no longer occurring then we should close this bug. The allocation warnings are a totally different beast and can be tracked by bug 486068.