[winswitch] capslock acting like shiftlock?

[Antoine Martin]

On 23/10/2020 18:33, Thomas Esposito wrote:
>> "Shift" and "Caps_Lock" are treated interchangeably by the key mapping
> code for non-native keymaps. (non-X11 clients for seamless servers)
> 
> I'm not sure what you mean by "non-native keymaps" in this context. 
Non-x11 clients, like MS Windows or MacOS.
For X11 clients connecting to seamless servers, things are easier to map
since the software is the same (X11) at both ends.

> FWIW, my xpra client is running on Windows 10 and I'm using the same
> xpra client version (3.0) that I was using when the xpra server version
> was 1.x running on Centos 6.6. I didn't have this problem before.
So this sounds like a regression, can you please file a ticket?
(with all the details for the keymap and keys pressed, etc)

> Also not sure what you mean when you say that "Shift" and "Caps_Lock"
> are treated interchangeably.
If either is pressed but not both, then the key is shifted and we use
the symbol found at level 1 of the keymap. (often the uppercase version
of the same keysym)

> Are you talking about the xpra client?
No, like I said: the client tries hard not to modify key events.

> It
> sounds like this explains my problem but that can't be because I haven't
> changed the client, only the server OS version and the xpra version
> running on it.
Yes, the key mapping happens all server side.

Cheers,
Antoine

> 
> 
> On Fri, Oct 23, 2020, 5:08 AM Antoine Martin via shifter-users
> <shifter-users at lists.devloop.org.uk
> <mailto:shifter-users at lists.devloop.org.uk>> wrote:
> 
>     On 23/10/2020 11:44, Thomas Esposito via shifter-users wrote:
>     > If I had to guess, I'd say the the xpra client is translating the
>     capslock
>     > to a shift before it ever gets to the server. Is that possible?
>     No, the client tries hard not not modify key events before sending them
>     to the server.
> 
>     > If so, how
>     > can I verify that this is happening and how would I fix it?
>     Run the server and client with "-d keyboard" to get details on events
>     and how they are processed.
>     (just be aware that this is hard to interpret as it is very verbose)
> 
>     "Shift" and "Caps_Lock" are treated interchangeably by the key mapping
>     code for non-native keymaps. (non-X11 clients for seamless servers)
> 
>     To fix it, please file a ticket as per:
>     https://xpra.org/trac/wiki/Keyboard#ReportingBugs
> 
>     Cheers,
>     Antoine
> 
> 
> 
>     >
>     > On Fri, Oct 23, 2020, 12:08 AM Thomas Esposito
>     <tmesposito00 at gmail.com <mailto:tmesposito00 at gmail.com>>
>     > wrote:
>     >
>     >> We finally upgraded from Centos 6.6 to 7.6 at work (it will be YEARS
>     >> before we see 8.x) and I took the opportunity to upgrade xpra to 3.x.
>     >>
>     >> So far, everything seems to be working well, except for some strange
>     >> keyboard behavior. My capslock key is acting as shiftlock (i.e.
>     effects ALL
>     >> keys, not just letters). I DON'T get this behavior in a VNC
>     session on the
>     >> same machine. I have compared xmodmap and setxkbmap settings
>     between my vnc
>     >> and xpra sessions and haven't found any differences that seem
>     like they
>     >> could explain this behavior.
>     >>
>     >> Any ideas?
>     >>
>     > _______________________________________________
>     > shifter-users mailing list
>     > shifter-users at lists.devloop.org.uk
>     <mailto:shifter-users at lists.devloop.org.uk>
>     > https://lists.devloop.org.uk/mailman/listinfo/shifter-users
>     >
> 
>     _______________________________________________
>     shifter-users mailing list
>     shifter-users at lists.devloop.org.uk
>     <mailto:shifter-users at lists.devloop.org.uk>
>     https://lists.devloop.org.uk/mailman/listinfo/shifter-users
>