cube usually rotates by 2 faces when using ctrlaltleft or ctrlaltright

Bug #157588 reported by Terry Browning
6
Affects Status Importance Assigned to Milestone
compiz (Ubuntu)
Fix Released
Undecided
Unassigned

Bug Description

Binary package hint: compiz

There are 4 faces to the cube, and I can see the face I want whizz by as I try to rotate by only 1 face.
Every other method for changing the faces works fine, it's just the keyboard that has the problem.
One workaround I have found is to rotate the cube a full rotation or so. This often gets a rotation by an odd number of faces. :-\
This may be complicated by the use of emerald, and may be the related to bug #155981, since 2 key presses causes a full rotation (usually).
Annoying.

Revision history for this message
Terry Browning (ubuntu9648) wrote :

A curious observation:
both ctrlaltleft and ctrlaltright are affected by the same bug, but differently.
Given the faces are numbered 0..3 (I'm using the gnome workspace switcher to track faces/virtual screens):
ctrlaltleft moves to an odd numbered face
ctrlaltright moves to an even numbered face
8-/
curiouser and curiouser.

except it's not the last key press, but the last-but-one key press that counts!

ctrlaltleft ctrlaltleft - gets an odd face
ctrlaltleft ctrlaltright - gets an odd face!
ctrlaltleft ctrlaltright ctrlaltright - gets an even face!!

A variable is being used before initialization (although initialization is occuring). Race condition?
Maybe the move is happening twice, once with uninitialized state, and once with initialized state. Doesn't explain the left>odd, right>even pattern.

I have noticed that this pattern breaks down when I press the buttons quickly, but I can't pin down the new pattern. More evidence of a race.

I tried 6 virtual screens, and the same pattern occurs. ctrlaltleft cycles 153153, and ctrlaltright cycles 024024.

I tried 5 virtual screens. ctrlaltleft cycles 420420420, ctrlaltright cycles 024024024.

I can still get the odd faces by fast-keying, but the 024/420 pattern re-establishes itself quickly when I return to a slow pace.
However, if I am at an odd face, I can move between them reliably unless I try to go around the end:
Starting at face 1, ctrlaltright gets me face 3
Starting at fare 3, ctrlaltleft gets me face 1
Starting at face 1, ctrlaltleft gets me face 4
Starting at face 3, ctrlaltright gets me face 0
The end-of-list code is affecting the odd/even thing. Also, I can't see the last-but-one effect. Could just be a side effect from the bug.

Revision history for this message
Terry Browning (ubuntu9648) wrote :

After more testing, I've a simpler and more complete pattern.

ctrlaltleft and ctrlaltright always move 2 faces

except:
when moving off the left (left of zero), the face is set to last face
when moving off the right of the last face, the face is set to face 0.

except if I press keys too quickly.

This fits all tested cases. Forget the bit about last-key-but-one.

Why does the cube rotate by 2 faces instead of 1?
Why does the cube rotate by odd numbers of faces when I press the keys too fast?

Revision history for this message
Travis Watkins (amaranth) wrote :

 Thank you for taking the time to report this bug and helping to make Ubuntu better. The issue that you reported is one that should be reproducible with the live environment of the Desktop CD of the development release - Hardy Heron. It would help us greatly if you could test with it so we can work on getting it fixed in the next release of Ubuntu. You can find out more about the development release at http://www.ubuntu.com/testing/ . Thanks again and we appreciate your help.

Changed in compiz:
status: New → Incomplete
Revision history for this message
Pedro Villavicencio (pedro) wrote :

We are closing this bug report because it lacks the information we need to investigate the problem, as described in the previous comments. Please reopen it if you can give us the missing information, and don't hesitate to submit bug reports in the future. To reopen the bug report you can click on the current status, under the Status column, and change the Status back to "New". Thanks again!.

Changed in compiz:
status: Incomplete → Invalid
Revision history for this message
Pietro Battiston (toobaz) wrote :

I'm testing up to date (updated yesterday) Hardy, bug is still here.

Unfortunately, I can't test it in the live CD, since my card (ATI Xpress200M) needs fgrlx driver for acceleration, and I can't (or just don't know how to) install it without rebooting. But I tested it after removing .gconf/compiz folder.

Please let me know if some more testing is needed. Notice that if I start gnome-compiz-preferences I get an error ("Wrong type: list expected, got string"), then all seems to work fine (apart from the reported bug).

Changed in compiz:
status: Invalid → New
Revision history for this message
Pietro Battiston (toobaz) wrote :

Changing status to "confirmed".

It is probably worth mentioning that:
- I'm using a 64 bit Hardy
- the bug occurs not only in the cube but also when I "unroll" the cube with CTRL+ALT+DOWN: if I move left/right, it still skips one screen
- the bug doesn't occurr if I directly click on a desktop in the panel destkops area: it moves to the right one

Changed in compiz:
status: New → Confirmed
Revision history for this message
Pietro Battiston (toobaz) wrote :

More testing showed that the bug also occurs if I switch to a window in another viewport.

E.g: I am on screen one, and on my docking bar I see the icon of a terminal that is in screen 2; if I click on it, instead that to screen 2, I get flipped to screen 3.

Similarly, if the window is in screen 3 (and I still am on 1), clicking on its icon makes me turn around all the cube and come back to screen 1.

Revision history for this message
Hew (hew) wrote :

I cannot reproduce this issue. Is the bug still present for you with Intrepid?

Changed in compiz:
status: Confirmed → Incomplete
Revision history for this message
Pietro Battiston (toobaz) wrote :

No, the bug really seems to be solved in Intrepid. Marking as "fix released", someone please reopen it they think it is still valid.

Changed in compiz:
status: Incomplete → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.