Tested with the following against a Fedora 21 r9329 trunk and 0.15.0 server:
In the Windows HTML5 viewer, I can hit the NumLock? button and it will work for one character, then it goes back to disabled. This is problematic as Windows will also toggle its NumLock? which will cause a desync of NumLock? - this was fixed in #312
Some information that might be useful for fixing this:
xkbmap_mod_missingto tell the server not to try to sync the numlock state for those events
Grepping the source code for num*lock will turn up quite a few hits.
@maxmylyn: is this still a problem?
Done in r12797.
So numlock can very easily start / go out of sync with the real keyboard state.
Noted. Quick and dirty workaround is to disconnect, toggle numlock (so it'll match with the server), then reconnect. I'm not sure if there's going to be an elegant fix since browsers are so sandboxed.
Quick and dirty workaround is to disconnect, toggle numlock
As per comment:4, you can do the same thing without disconnecting: just hit numlock whilst the xpra browser window is NOT receiving the key events (focus some other window or minimize the browser temporarily). That's a lot easier than re-connecting.
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/858