Re: [cp-patches] FYI: fixed keyboard focus handling in java.awt.DefaultKeyboardFocusManager

[Thomas Fitzsimmons]

On Fri, 2005-05-20 at 21:26 +0200, Roman Kennke wrote:
> Hi,
> 
> the DefaultKeyboardFocusManager as it was implemented had a serious
> problem: it recognized the KEY_PRESSED and KEY_RELEASED event for
> focusTraversalKeys and consumed them properly. BUT it did not recognize
> the KEY_TYPED event for the same key and let it go through the
> component. This ended up in TAB characters appearing in text components
> when shifting focus with TAB for example. The problem here is, that
> KEY_TYPED events are different than KEY_PRESSED and KEY_RELEASED (the
> former carry the character for the key, the latter the key codes VK_*).
> So the mapping is not straightforward and probably not even always
> possible. What I implemented is that all KEY_TYPED events that follow a
> KEY_PRESSED event for a focusTraversalKey are swallowed until the
> corresponding KEY_RELEASED event is received. I think that this is ok
> so.
> 
> 2005-05-20  Roman Kennke  <address@hidden>
> 
>         * java/awt/DefaultKeyboardFocusManager.java
>         (processKeyEvent): Properly swallow KEY_TYPED events after
>         a KEY_PRESSED event for a focusTraversalKey has been received.
>         Otherwise the focusTraversalKey (e.g. a TAB) would end up
>         beeing delivered to the focused component.
> 
> /Roman
> 
> _______________________________________________
> Classpath-patches mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/classpath-patches

+   * This flag indicates for which focus traversal key release event we
+   * possibly wait, bevor letting any more KEY_TYPED events through.

Typo, "before".

+    // Here we check if we are currently waiting for a KEY_PRESSED and

Aren't you typically waiting for a KEY_RELEASED event here, and
generally waiting for the keystroke opposite to the one that initiated
the focus traversal?

Once these comments are fixed, this can go in.

Thanks,
Tom