It is just a little bit too late to get this in version 0.11 (though it could get backported if the fix is simple enough and does not cause any regression)
It looks like the value for is_native_keymap
can end up being True
even on client platforms that do not have native X11 keymap support. This shortcuts a number of codepaths which will need re-testing:
set_keymap
setting the modifiers_filter
to ignore lock
and the modifier for Num_Lock
- do we still need this?
get_keycode
uses a more strict key lookup, which may cause some unusual foreign keys not to be looked up?
I've ripped out the whole code because it seemed bogus: see r5303. I've kept the is_native_keymap
flag, which is still shown in xpra info
but no longer used.
Please test thoroughly to make sure this does not cause keyboard regressions with modifiers (Alt
, Shift
, etc).
We did some testing, and the keyboard functions properly in r5319 Windows and OSX, including the alt
, shift
, control
, and numpad/numlock
keys.
side note: we were using server r5319.
I'm not sure this is working properly, I had problems today with "|" with a Linux client, and numbers on a win32 vbox client..
Testing with windows native r5319 I can't find any keyboard function key not behaving as expected. With osx I even noticed that the clear
button of the numpad behaves like a numlock, allowing the osx numpad to be used as arrow, etc. keys (for anyone who likes that sort of thing).
The numbers on the numpad and keyboard itself are behaving as expected. The "|" key is interpreted correctly in xterms.
I can't think of any other modifier keys to test.
I am unsure if this applies to this bug (otherwise I will open a new one):
I have an Italian keyboard. Some keys (notably ',?', '{', '}', '#' and '@') are only available via 'Alt_Gr' escape. All these keys do not appear to work.
Setup:
comment:6 is not specifying on which version the problem is occurring so I assume it isn't trunk, and it isn't related to this ticket, which is about changes present in v0.12 only. @MCondarelli: please provide more details as per wiki/ReportingBugs and in particular wiki/Keyboard, and if appropriate, place those in a new ticket.
afarr: a bit of testing with Fedora 20 wouldn't hurt, but I haven't had the problem since, maybe it was just a glitch.
Not heard back, assuming this works OK. (see also #561)
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/483