The fix itself is good.
Could you, please, remove the author tag from the test header? We have
agreement to not use it anymore.
ok. thank you.
And the test frame need to be disposed after the test.
--Semyon
On 2/6/2017 4:32 PM, Mikhail Cherkasov wrote:
Hi all,
Bug: https://bugs.openjdk.java.net/browse/JDK-8173145
Webrev: http://cr.openjdk.java.net/~mcherkas/8173145/webrev.00/
The problem appears because AltProcessor doesn't reset altKeyPressed
flag when
not-alt key pressed, but event is consumed.
This situation is possible when all events consumed by key bindings.
A menubar is activated while
user expects only key binding actions.
In my test a button should be activated by alt+m mnemonic, but a menu
is activate, because
key pressed/released consumed by mnemonic combination and key typed
event is
consumed by DefaultEditorKit$DefaultDeyTypedAction that is installed
for JTextField,
as result instead of button, menu is activated.
To fix the issue I added resetting altKeyPressed flag if any non-alt
key pressed despite consuming state,
because obviously user what make some alt+[KEY] action instead of
menu activation.
Thanks,
Mikhail.
