Comment #7 on issue 3916 by
[hidden email]: Can't input some Russian
characters at text editors
http://code.google.com/p/pharo/issues/detail?id=3916Well, yes and no.
It is perfectly reasonable to assume arrow up to always be in the
single-byte ascii keyValue.
It is however also perfectly reasonable to assume keyValue will not contain
garbage values for Unicode input, which it does on Windows.
So either one could change uses of keyValue like proposed here, which would
require different patches for 1.2 and 1.3 based on some refactorings made
in TextEditor, or one could hack in consistency in the event processor
where KeyboardEvents are generated, no always return 0 as keyValue when
charCode is outside single-byte range.
There may still be trouble as the case of ctrl-character presses are
different for charCode and keyValue (at least on Windows), but that's
another matter.
Find modified method attached for testing, the relevant part is:
"Adjustments to provide consistent key value data for different VM's:
- charCode always contains unicode code point.
- keyValue contains 0 if input is outside legacy range"
"If there is no unicode data in the event, assume keyValue contains a
correct (<256) Unicode codepoint, and use that"
(charCode isNil
or: [charCode = 0])
ifTrue: [charCode := keyValue].
"If charCode is not single-byte, we definately have Unicode input. Nil
keyValue to avoid garbage values from som VMs."
charCode > 255 ifTrue: [keyValue := 0].
Attachments:
HandMorph-generateKeyboardEvent.st 2.4 KB
_______________________________________________
Pharo-bugtracker mailing list
[hidden email]
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-bugtracker
Locked
