Home » Eclipse Projects » GEF » I20040205 + M6 -> 100% cpu usage on gtk
| I20040205 + M6 -> 100% cpu usage on gtk [message #116069] |
Fri, 06 February 2004 13:04  |
Eclipse User |
|
|
|
Originally posted by: sylvain.pasche.switzerland.org
Hi,
I developed recently a Gef editor based on the logic samples and it
ran fine on M5.
But when I wanted to port it to M6 with Gef build I20040205, I experienced
some troubles with cpu hogging.
Sometime, when I resize the editor, suddenly eclipse consumes 100% cpu.
I found a reproductible way for this problem while interacting with the
rulers on the logic samples: Just unselect the View->Rulers menu option,
and it runs with 100% cpu.
I tested on win32, and it works well. So it might be gtk-related. On
motif, I got a "unable to create editor" when opening a logic sample.
I found some entries on bugzilla about gtk and 100% cpu, but they are
closed. I don't know if this is related, but it looks like to be a weird
to solve bug :-(
Has anyone some ideas about this problem ?
Thanks
Sylvain
|
|
|
| Re: I20040205 + M6 -> 100% cpu usage on gtk [message #116303 is a reply to message #116069] |
Mon, 09 February 2004 22:54  |
| Eclipse User |
|
|
|
Originally posted by: none.us.ibm.com
Is the CPU just busy, or is the UI unresponsive?
"Sylvain Pasche" <sylvain.pasche@switzerland.org> wrote in message
news:pan.2004.02.06.18.04.34.310462@switzerland.org...
> Hi,
>
> I developed recently a Gef editor based on the logic samples and it
> ran fine on M5.
> But when I wanted to port it to M6 with Gef build I20040205, I experienced
> some troubles with cpu hogging.
> Sometime, when I resize the editor, suddenly eclipse consumes 100% cpu.
>
> I found a reproductible way for this problem while interacting with the
> rulers on the logic samples: Just unselect the View->Rulers menu option,
> and it runs with 100% cpu.
>
> I tested on win32, and it works well. So it might be gtk-related. On
> motif, I got a "unable to create editor" when opening a logic sample.
>
> I found some entries on bugzilla about gtk and 100% cpu, but they are
> closed. I don't know if this is related, but it looks like to be a weird
> to solve bug :-(
>
> Has anyone some ideas about this problem ?
>
> Thanks
>
> Sylvain
|
|
|
| Re: I20040205 + M6 -> 100% cpu usage on gtk [message #116788 is a reply to message #116303] |
Tue, 10 February 2004 07:25 |
| Eclipse User |
|
|
|
Originally posted by: sylvain.pasche.switzerland.org
On Mon, 09 Feb 2004 22:54:58 -0500, Randy Hudson wrote :
> Is the CPU just busy, or is the UI unresponsive?
Well, the CPU is busy, but the UI continues to be responsive.
I just lauched the debugger to see where all the cpu is consumed:
Looking into swt.Display :
public boolean readAndDispatch () {
checkDevice ();
runPopups ();
int status = OS.gtk_events_pending ();
if (status != 0) {
OS.gtk_main_iteration ();
runDeferredEvents ();
return true;
}
return runAsyncMessages ();
}
OS.gtk_events_pending seems to be always returning true, the the following
backtrace is always happening on pause:
Thread [main] (Suspended)
OS.gtk_main_do_event(int) line: not available [native method]
Display.eventProc(int, int) line: 710
OS.gtk_main_iteration() line: not available [native method]
Display.readAndDispatch() line: 1624
Workbench.runEventLoop(Window$IExceptionHandler, Display) line: 1506
Workbench.runUI() line: 1482
Workbench.createAndRunWorkbench(Display, WorkbenchAdvisor) line: 246
PlatformUI.createAndRunWorkbench(Display, WorkbenchAdvisor) line: 139
IDEApplication.run(Object) line: 47
PlatformActivator$1.run() line: 226
EclipseStarter.run(String[], Runnable) line: 85
NativeMethodAccessorImpl.invoke0(Method, Object, Object[]) line: not available [native method]
NativeMethodAccessorImpl.invoke(Object, Object[]) line: 39
DelegatingMethodAccessorImpl.invoke(Object, Object[]) line: 25
Method.invoke(Object, Object[]) line: 324
Main.basicRun(String[]) line: 279
Main.run(String[]) line: 742
Main.main(String[]) line: 581
It would be useful to set breakpoint on the right place, to see why all
these gtk events are generated, but I don't know where to look ..
I hope it's a bit clearer.
Regards,
Sylvain
|
|
|
| Re: I20040205 + M6 -> 100% cpu usage on gtk [message #117111 is a reply to message #116788] |
Wed, 11 February 2004 15:15 |
| Eclipse User |
|
|
|
Originally posted by: none.us.ibm.com
We are working on setting up a new Redhat machine so we can look into this
further. Thanks.
"Sylvain Pasche" <sylvain.pasche@switzerland.org> wrote in message
news:c0aidl$2mi$1@eclipse.org...
> On Mon, 09 Feb 2004 22:54:58 -0500, Randy Hudson wrote :
>
> > Is the CPU just busy, or is the UI unresponsive?
>
> Well, the CPU is busy, but the UI continues to be responsive.
>
> I just lauched the debugger to see where all the cpu is consumed:
>
> Looking into swt.Display :
>
> public boolean readAndDispatch () {
> checkDevice ();
> runPopups ();
> int status = OS.gtk_events_pending ();
> if (status != 0) {
> OS.gtk_main_iteration ();
> runDeferredEvents ();
> return true;
> }
> return runAsyncMessages ();
> }
>
> OS.gtk_events_pending seems to be always returning true, the the following
> backtrace is always happening on pause:
>
>
> Thread [main] (Suspended)
> OS.gtk_main_do_event(int) line: not available [native method]
> Display.eventProc(int, int) line: 710
> OS.gtk_main_iteration() line: not available [native method]
> Display.readAndDispatch() line: 1624
> Workbench.runEventLoop(Window$IExceptionHandler, Display) line: 1506
> Workbench.runUI() line: 1482
> Workbench.createAndRunWorkbench(Display, WorkbenchAdvisor) line: 246
> PlatformUI.createAndRunWorkbench(Display, WorkbenchAdvisor) line: 139
> IDEApplication.run(Object) line: 47
> PlatformActivator$1.run() line: 226
> EclipseStarter.run(String[], Runnable) line: 85
> NativeMethodAccessorImpl.invoke0(Method, Object, Object[]) line: not
available [native method]
> NativeMethodAccessorImpl.invoke(Object, Object[]) line: 39
> DelegatingMethodAccessorImpl.invoke(Object, Object[]) line: 25
> Method.invoke(Object, Object[]) line: 324
> Main.basicRun(String[]) line: 279
> Main.run(String[]) line: 742
> Main.main(String[]) line: 581
>
> It would be useful to set breakpoint on the right place, to see why all
> these gtk events are generated, but I don't know where to look ..
>
> I hope it's a bit clearer.
>
> Regards,
>
> Sylvain
>
>
|
|
|
Goto Forum:
Current Time: Fri May 09 02:05:52 EDT 2025
Powered by FUDForum. Page generated in 0.08769 seconds |
Home