! Please note that this is a snapshot of our old Bugzilla server, which is read only since May 29, 2020. Please go to gitlab.xfce.org for our new server !
xfce switchs its desktop indefinitely without asking for it
Status:
RESOLVED: INVALID

Comments

Description polytan 2007-03-09 12:42:42 CET
User-Agent:       Mozilla/5.0 (X11; U; Linux i686; fr; rv:1.8.1.2) Gecko/20060601 Firefox/2.0.0.2 (Ubuntu-edgy)
Build Identifier: 

I had this bug a lot of time, on x86_64 gentoo linux and ubuntu linux (edgy) using xfce-4.4 (beta,rc and stable).

I use this powerfull feature (it's the best feature in xfce to me :) ) : switching virtual desktops using the mouse only. I can go to the first desktop going on the right of the last one. 

My problem : sometimes, when I only move the mouse (not adding windows, not moving them, just moving the mouse), xfce switchs desktop, from the first to the last one, really fastly and indefinetly.
I have to move my mouse again to stop to.

I think it when my cursor is near a border (especially a corner but I'm not sure, it's so fast that I can't really see).

Maybe it's because of the feature I've checked called "change desktop fonction of windows fitting".

This bug is really stressing because the screen highlights all the time.

Reproducible: Couldn't Reproduce

Steps to Reproduce:
1.I don't know, but i had this bug a lot of times (10-12 times)
2.
3.



Expected Results:  
doing nothing, not to change the desktop involuntarily

just ask me :)
Comment 1 Olivier Fourdan editbugs 2007-03-09 13:04:43 CET
Can you elaborate on what you call "change desktop fonction of windows fitting" because there is no such option in xfce.
Comment 2 polytan 2007-03-09 13:18:41 CET
(In reply to comment #1)
> Can you elaborate on what you call "change desktop fonction of windows fitting"
> because there is no such option in xfce.
> 

I've just seen this option checked for 2 hours :)

It's in french so I try to translate it :
Advance windows feature (paufinage des fenêtres)
  => Workspaces (espaces de travail)
    => Switch desktop function of the fittingof windows (Basculer d'espace en fonction de l'agencement du bureau)

Comment 3 Olivier Fourdan editbugs 2007-03-09 15:39:25 CET
When warping workspaces, the pointer is sent far from the side to avoid the kind of effects you describe. Where is located the mouse pointer when this occurs, does  it move or does it stay at the same place?

Another possible cause could be the mouse wheel. I noticed some mouse wheels may send continuous data in some case. You may try disabling the option "Switch workspaces using the mouse wheel" in the wm-tweaks settings.
Comment 4 Olivier Fourdan editbugs 2007-03-09 19:37:05 CET
A few more questions:

- Do you have multiple rows in the pager (like 2x2)?
- Do you have set the warp threshold to the min?
Comment 5 polytan 2007-03-09 21:39:21 CET
(In reply to comment #3)
> When warping workspaces, the pointer is sent far from the side to avoid the
> kind of effects you describe. Where is located the mouse pointer when this
> occurs, does  it move or does it stay at the same place?
I remember it's in the top of the screen or in the bottom
 
> Another possible cause could be the mouse wheel. I noticed some mouse wheels
> may send continuous data in some case.
I use both mouse and touchpads on which there is no wheel

>  You may try disabling the option "Switch
> workspaces using the mouse wheel" in the wm-tweaks settings.
I've done it. I've not seen the bug again, but I hope, to solve it I mean ... :) 

Comment 6 polytan 2007-03-09 21:41:10 CET
(In reply to comment #4)
> A few more questions:
> 
> - Do you have multiple rows in the pager (like 2x2)?
no, just 5x1
> - Do you have set the warp threshold to the min?
What that ? the force I have to apply to switch the desktop ? Yes, it's to the min.

Comment 7 Olivier Fourdan editbugs 2007-03-10 21:10:48 CET
(In reply to comment #5)
> (In reply to comment #3)
> > When warping workspaces, the pointer is sent far from the side to avoid the
> > kind of effects you describe. Where is located the mouse pointer when this
> > occurs, does  it move or does it stay at the same place?
> I remember it's in the top of the screen or in the bottom

Yes, but does the pointer moves? If you warp workspace using the mouse, you'll notice that the pointer is moved far from the edge automatically. Does that happen when the bug occurs?
Comment 8 polytan 2007-03-11 21:20:52 CET
(In reply to comment #7)
> (In reply to comment #5)
> > (In reply to comment #3)
> > > When warping workspaces, the pointer is sent far from the side to avoid the
> > > kind of effects you describe. Where is located the mouse pointer when this
> > > occurs, does  it move or does it stay at the same place?
> > I remember it's in the top of the screen or in the bottom
> 
> Yes, but does the pointer moves? If you warp workspace using the mouse, you'll
> notice that the pointer is moved far from the edge automatically. Does that
> happen when the bug occurs?
> 
hum no. I'm not sure (and I don't have seen this bug again) but I quite sure it's no.

Comment 9 polytan 2007-03-14 15:27:16 CET
Hi, lot of news ... bad news ... :)

The bug happened a lot of time today (maybe because I've done something different than going to the web...)

I've made a video of the bug, but It's been made with my mobile, so I'm not sure the quality is pretty good.

I was using a lot the function to switch from a desktop to another with the pointer. 

I've just added a bit of resistance of the edge of the screen (just a bit) since the first time I've adding this bug in the database. 
Comment 10 polytan 2007-03-14 19:20:43 CET
Created attachment 1032 
the bug in action

I'm sorry for th 3gp format, I hope you can read it :/
Comment 11 Olivier Fourdan editbugs 2007-03-14 20:48:28 CET
Really hard to tell from that screencast, it's really small, I cannot even see the pointer... 

Where is located the pointer?
Comment 12 polytan 2007-03-14 21:16:32 CET
(In reply to comment #11)
> Really hard to tell from that screencast, it's really small, I cannot even see
> the pointer... 
> 
> Where is located the pointer?
> 

In this video, it's in the bottom, going from the left to the right. You can "see" it moving itself.

The bug happened 12 times today. It made me craze ! :p Just when I used my computer for something else than game.

Today, it happened with the cursor on the top too.

I'm not sure (It's really hard to describe, it's really fast, so I'm really sorry for the poor quality of the bug report I do) but, it looks like the cursor is attracted by something when I'm moving it and then the bug starts.
Comment 13 Olivier Fourdan editbugs 2007-03-14 21:27:45 CET
(In reply to comment #12)
> In this video, it's in the bottom, going from the left to the right. You can
> "see" it moving itself.
> 
> The bug happened 12 times today. It made me craze ! :p Just when I used my
> computer for something else than game.
> 
> Today, it happened with the cursor on the top too.
> 
> I'm not sure (It's really hard to describe, it's really fast, so I'm really
> sorry for the poor quality of the bug report I do) but, it looks like the
> cursor is attracted by something when I'm moving it and then the bug starts.
> 

You mean the cursor moves all by itself? Really sounds odd, a bit like a bug in the mouse driver or something... Are you sure it doesn't happen with workspace switch disabled?
Comment 14 polytan 2007-03-14 21:34:44 CET
> 
> 
> You mean the cursor moves all by itself? Really sounds odd, a bit like a bug in
> the mouse driver or something... Are you sure it doesn't happen with workspace
> switch disabled?
> 

Actually, I've never use xfce without this feature...

Do you want me to do it ?
Comment 15 Olivier Fourdan editbugs 2007-03-14 21:45:49 CET
(In reply to comment #14)
> Actually, I've never use xfce without this feature...
> Do you want me to do it ?

Well, yes, it's worth it... If you disable the feature and see the pointer moving toward the edge all by itself, then you'll know it's a bug in the driver.

I'm asking that because I've never heard of such an issue and I cannot reproduce this bug on any system I have access. 

Comment 16 polytan 2007-03-15 07:47:36 CET
> (In reply to comment #14)
> Well, yes, it's worth it... If you disable the feature and see the pointer
> moving toward the edge all by itself, then you'll know it's a bug in the
> driver.
> 
> I'm asking that because I've never heard of such an issue and I cannot
> reproduce this bug on any system I have access. 
> 

So, I'm sad, but I'm using xfce without the function enable. No sign of the bug currently. I hope it's a bug in the mouse driver...
Comment 17 polytan 2007-03-17 12:19:05 CET
Re,

I'm using xfce on a computer without the feature (no problem) and on another with the feature enable (problem, today again).

But, yes, but, I prefer to think about a X.org mouse driver problem (or a mouse problem, but it is new (less than 3 weeks) ) because my cursor is trembling when I'm on a graphical session. 

It's horrible and I'm near an epilepsy crisis !

Did you heard about such bug in X.org currently ?
Comment 18 Olivier Fourdan editbugs 2007-03-18 19:49:16 CET
(In reply to comment #17)
> Re,
> 
> I'm using xfce on a computer without the feature (no problem) and on another
> with the feature enable (problem, today again).

Without the option "warp workspace when the pointer reaches a screen edge" enabled, it might be harder to notice when the bug occurs.

> But, yes, but, I prefer to think about a X.org mouse driver problem (or a mouse
> problem, but it is new (less than 3 weeks) ) because my cursor is trembling
> when I'm on a graphical session. 
> 
> It's horrible and I'm near an epilepsy crisis !

You said you have a mouse and a touchpad on the laptop? Did you try to unplug the mouse?

> Did you heard about such bug in X.org currently ?

Nope, but I'm not an Xorg developer so I'm not closely following Xorg bugs (it it's an Xorg bug, I don't know yet)

Comment 19 Nick Urbanik 2007-07-06 00:57:45 CEST
I am troubled by this bug (the workspaces rapidly and endlessly cycling, with no way to stop them other than restarting X), but it is triggered for me by a rapid pressing of Alt-Control-arrow_key.

This happens for me on two machines, both at home and at work.

I can trigger it quite easily.  This has also been a problem with earlier Xfce installations on FC6.

Currently using FC7, xfce 4.4.1.
Comment 20 polytan 2007-07-06 05:59:08 CEST
(In reply to comment #19)
> I am troubled by this bug (the workspaces rapidly and endlessly cycling, with
> no way to stop them other than restarting X), but it is triggered for me by a
> rapid pressing of Alt-Control-arrow_key.
> 
> This happens for me on two machines, both at home and at work.
> 
> I can trigger it quite easily.  This has also been a problem with earlier Xfce
> installations on FC6.
> 
> Currently using FC7, xfce 4.4.1.
> 

So I'm not alone.

In a way, I would have prefered :/
Comment 21 Olivier Fourdan editbugs 2007-07-10 20:17:47 CEST
(In reply to comment #20)
> (In reply to comment #19)
> > I am troubled by this bug (the workspaces rapidly and endlessly cycling, with
> > no way to stop them other than restarting X), but it is triggered for me by a
> > rapid pressing of Alt-Control-arrow_key.
[...]
> 
> So I'm not alone.
> 
> In a way, I would have prefered :/

Depends. Is that how it'triggered for you too? If not (as I understood it, this had to do with the mouse and not keyboard shortcut), then it's different things.
Comment 22 Olivier Fourdan editbugs 2007-07-10 20:33:37 CEST
Created attachment 1278 
Release the svent queue prior to grabbing keys in further keyboard shortcut processing

(In reply to comment #19)
> I am troubled by this bug (the workspaces rapidly and endlessly cycling, with
> no way to stop them other than restarting X), but it is triggered for me by a
> rapid pressing of Alt-Control-arrow_key.

Unfortunately, I cannot reproduce that problem. First of all, please make sure you have not enabled any accessibility option from the keyboard settings (just to make sure)

> This happens for me on two machines, both at home and at work.

Which is a good things, it means that you can do the testing for me if you will.

> [...]

Does the attached patch improve things? (this patch is against SVN trunk, let me know if you have any trouble applying the patch or building from sources)
Comment 23 Olivier Fourdan editbugs 2007-07-19 21:11:10 CEST
@Nick: I believe your problem is related to Fedora using tickless kernel. But as I'm using x86_64 tight now, I cannot test that (tickless is nto available on x86_64 yet)

So unless you test the proposed patch and report if that makes any difference, or even better, the xfwm-4.5 version from this link (currently xfwm4-4.5.0-r25926.tar.bz2 but I update the snapshot fairly often):

    http://www.foo-projects.org/~olivier/preview/

Then I cannot help any further...
Comment 24 Yves-Alexis Perez editbugs 2008-03-13 09:11:41 CET
Some Debian users may have the same problem. I reproduced it myself once.

This is only using keybindings. It seems to be triggered when switching from one desktop to another and switch again fastly, before the first switch ends.

Sometimes (and especially on slower or loaded computers), it enters some kind of loop where xfwm constantly switches desktop. It's not easily reproducible, but I observed it once.

This is on xfwm 4.4.2. If you have some patch against current 4.4.2 branch, I can ask those Debian users to test it. (at least one seems to reproduce it quite easily).

If you need more info, please ask.
Comment 25 Olivier Fourdan editbugs 2008-03-15 07:28:36 CET
I have seen it once too so I have no doubt that this problem is real, unfortunately as you said it is not easily reproducible (I could not reproduce it myself on any of my systems)
Comment 26 Yves-Alexis Perez editbugs 2008-03-15 13:22:02 CET
(In reply to comment #25)
> I have seen it once too so I have no doubt that this problem is real,
> unfortunately as you said it is not easily reproducible (I could not reproduce
> it myself on any of my systems)
> 

In fact, I just reproduced it quite easily.
In Keyboard settings, put the first slider (Delay) at extreme right (so “extra long”). Then, if you have “Remember and recall previous workspace…” activated in Window Manager tweaks, press Ctrl+Fx and keep pressed. The bug is triggered instantly here.

HTH.
Comment 27 Yves-Alexis Perez editbugs 2008-03-15 13:35:58 CET
(In reply to comment #26)
> (In reply to comment #25)
> > I have seen it once too so I have no doubt that this problem is real,
> > unfortunately as you said it is not easily reproducible (I could not reproduce
> > it myself on any of my systems)
> > 
> 
> In fact, I just reproduced it quite easily.
> In Keyboard settings, put the first slider (Delay) at extreme right (so
> “extra long”). Then, if you have “Remember and recall previous
> workspace…” activated in Window Manager tweaks, press Ctrl+Fx and keep
> pressed. The bug is triggered instantly here.

In fact, as soon as the slider not at the default position, the bug can be triggered quite easily. When the slider is at the left (“short delay”), the flicker is really fast when you let the finger on the Ctrl+Fx. If the slider is at “long”, the flicker happens after letting the finger for longer.

But in either case, the flicker eventually happens, while it's quite rare here if I don't touch anything on the default setting. (On other install, with a slower cpu and maybe slower video card, even default setting trigger the bug easily, but changing the default setting leads to triggering flick even more easily).

Cheers,
Comment 28 Olivier Fourdan editbugs 2008-03-15 13:46:04 CET
Not sure we are talking about the same thing, the problem you mention looks like keyboard repeat, because beside this I cannot reproduce any such bug here.
Comment 29 Yves-Alexis Perez editbugs 2008-03-15 19:23:21 CET
(In reply to comment #28)
> Not sure we are talking about the same thing, the problem you mention looks
> like keyboard repeat, because beside this I cannot reproduce any such bug here.

Well, even after stopping pressing the Ctrl+Fx keys, it's still impossible to interrupt the desktop switching. I guess it switches back and forth between the two same desktops, but maybe it tries all desktops. I'll test and report back, but there's definitely a bug there (not sure if it's in xfwm4 or in the keyboard code).
Comment 30 Olivier Fourdan editbugs 2008-03-20 18:56:27 CET
Can you try current svn code?
Comment 31 Yves-Alexis Perez editbugs 2008-03-20 19:05:04 CET
(In reply to comment #30)
> Can you try current svn code?

Is it buildable/usable with Xfce 4.4.2? If yes, I'll do that, but either, it may be tricky.
Comment 32 polytan 2008-03-20 22:19:54 CET
(In reply to comment #30)
> Can you try current svn code?
> 

I didn't have the problem for a quite long time now.

If you want, I can try the svn code, but for all programs or just xfwm4 ?
Comment 33 polytan 2008-03-29 10:28:05 CET
(In reply to comment #32)
> (In reply to comment #30)
> > Can you try current svn code?
> > 
> 
> I didn't have the problem for a quite long time now.
> 
> If you want, I can try the svn code, but for all programs or just xfwm4 ?
> 

The bug has run again today, just a few seconds ago. I don't know what happens and I don't know how to have more information... :(

My cursor was block in the bottom (fr: en bas) of the screen, as usually.

What svn package do you want me to compile and use ?
Comment 34 Olivier Fourdan editbugs 2008-03-29 20:27:50 CET
unfortunaly, this bug is a total mess, ppl posted different issues in the report

i believe your problem is different from the others as it is related to the mouse pointer and could be from az driver/mouse bug as i never heard of such an issue elsewhere.

Other ppl report an issue with kbd shortcuts...
Comment 35 Yves-Alexis Perez editbugs 2008-04-28 16:56:27 CEST
It seems that, for Debian at least, this was an Xorg bug.

See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=473165 for reference.

Cheers,
Comment 36 Skunnyk editbugs 2020-05-21 21:58:55 CEST
Closing old bugs.

Bug #2991

Reported by:
polytan
Reported on: 2007-03-09
Last modified on: 2020-05-21

People

Assignee:
Olivier Fourdan
CC List:
4 users

Version

Attachments

Additional information