control+w performs undo function instead of cycling buffers

[Ebibbee-gmail]

I am using PyRoom on Ubuntu Hardy from the launchpad repositories. Everything works very well except for the control+w keybinding. When I press it, it performs the undo function instead of cycling through the buffers and then exiting if I am in the last buffer. Control+z, on the other hand, does nothing at all.

I am using the standard USA keyboard layout and have replicated this behavior on two different machines, both running Ubuntu Hardy.

Any suggestions/ideas?

[Florian Heinle (tiax)]

Hi,

sorry, I've been unable to reproduce it. I changed keyboard bindings on my dev machine (intrepid), another regular machine (hardy) and some freshly set up virtual machines (hardy, intrepid, jaunty) and/or installed with default USA keyboard layouts and the keys work as expected.

Did you change anything else, concerning keyboard layouts or shortcuts?

[Ebibbee-gmail (ebibbee-gmail)]

Hi.

The only things that I've changed in the keyboard layouts is to
1) add the French layout as an option
2) change the keybinding used to cycle from US to French and back (Windows key)

At first, I thought it to be a fluke, so I was especially surprised
when it happened on my office computer, as well.

I'll try to disable the additional keyboard setting and let you know
what happens - maybe that will help, though I can't imagine it would
make much difference.

Best,
Evan

On Sun, Mar 29, 2009 at 6:42 AM, Florian Heinle
<email address hidden> wrote:
> Hi,
>
> sorry, I've been unable to reproduce it. I changed keyboard bindings on
> my dev machine (intrepid), another regular machine (hardy) and some
> freshly set up virtual machines (hardy, intrepid, jaunty) and/or
> installed with default USA keyboard layouts and the keys work as
> expected.
>
> Did you change anything else, concerning keyboard layouts or shortcuts?
>
> ** Changed in: pyroom
>   Importance: Undecided => Medium
>       Status: New => Incomplete
>
> --
> control+w performs undo function instead of cycling buffers
> https://bugs.launchpad.net/bugs/349715
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Ebibbee-gmail (ebibbee-gmail)]

OK. This is very surprising, but once I remove the additional keyboard
setting from my preferences pane, CTRL+W works as it should. However,
it is strange that just having an international layout available
(though not in use) creates a keybinding problem in PyRoom.

Any thoughts?

- ejb

On Sun, Mar 29, 2009 at 6:42 AM, Florian Heinle
<email address hidden> wrote:
> Hi,
>
> sorry, I've been unable to reproduce it. I changed keyboard bindings on
> my dev machine (intrepid), another regular machine (hardy) and some
> freshly set up virtual machines (hardy, intrepid, jaunty) and/or
> installed with default USA keyboard layouts and the keys work as
> expected.
>
> Did you change anything else, concerning keyboard layouts or shortcuts?
>
> ** Changed in: pyroom
>   Importance: Undecided => Medium
>       Status: New => Incomplete
>
> --
> control+w performs undo function instead of cycling buffers
> https://bugs.launchpad.net/bugs/349715
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Ebibbee-gmail (ebibbee-gmail)]

I've just added the French layout as an available layout in my
preferences dialog and (without actually selecting that as my current
layout) reproduced the same problem.

Just let me know if/how I can be of more help.

--ejb

On Sun, Mar 29, 2009 at 6:42 AM, Florian Heinle
<email address hidden> wrote:
> Hi,
>
> sorry, I've been unable to reproduce it. I changed keyboard bindings on
> my dev machine (intrepid), another regular machine (hardy) and some
> freshly set up virtual machines (hardy, intrepid, jaunty) and/or
> installed with default USA keyboard layouts and the keys work as
> expected.
>
> Did you change anything else, concerning keyboard layouts or shortcuts?
>
> ** Changed in: pyroom
>   Importance: Undecided => Medium
>       Status: New => Incomplete
>
> --
> control+w performs undo function instead of cycling buffers
> https://bugs.launchpad.net/bugs/349715
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Florian Heinle (tiax)]

Ok, marking the bug as confirmed now, I've been able to reproduce it

In a true moment of WTF I found out it only seems to happen if the French layout is added and it doesn't matter if it's selected or not. I tested with random other keyboard layouts (German per default for me, US, Arabic, Albania, Yemen, etc) and (so far) could only reproduce it with the French layout.

I'll be treating this bug as a blocker for 0.4.1

Bruno, what's so different with French layouts? :)

[Florian Heinle (tiax)]

will be fixed in 0.4.1, blocks release

[Ebibbee-gmail (ebibbee-gmail)]

Great work, Florian and yes, that is a true WTF moment, indeed!

Best,
Evan

On Sun, Mar 29, 2009 at 12:47 PM, Florian Heinle
<email address hidden> wrote:
> ** Changed in: pyroom
>       Status: Confirmed => In Progress
>
> --
> control+w performs undo function instead of cycling buffers
> https://bugs.launchpad.net/bugs/349715
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Florian Heinle (tiax)]

tentatively fixed in Revision 176, still waiting for someone with a Russian, Chinese, ... keyboard with entirely different keyboard layout to test it, before I merge this into 0.4

explanation: the mechanism we used in the first place broke when users with layouts like that used PyRoom, they had to switch keyboard layouts to be able to use our keyboard layouts. So we switched to hardware keyboard event-detection which seemed to work fine until someone used different keyboard layouts. Apparently that behaviour works fine as long as there is only one keyboard layout available.

On the other hand, that method doesn't seem to be the right way anyway, I've switched PyRoom to gtk "accelerators" now. I'm still testing it before merging it into 0.4

[SkAZi (scsiboy)]

It works fine with Russian layout.

[Florian Heinle (tiax)]

fixed in release 0.4.1