10.3 Key Events

This section describes issues related to handling key events.

The following keyboard issues are currently unresolved:

• On some non-English keyboards certain accented keys are engraved on the keytop and therefore are primary layer characters. Nevertheless, they cannot be used for mnemonics because there is no corresponding Java keycode.

• Changing the default locale at runtime does not change the text that is displayed for the menu accelerator keys.

• On a standard 109-key Japanese keyboard, the yen key and the backslash key both generate a backslash, because they have the same character code for the WM_CHAR message. AWT should distinguish them. This will be fixed in a future release.

The following keyboard issues concern the Oracle Solaris 10 and Linux x86 systems.

• Keyboard input in these systems is usually based on the X keyboard extension (XKB) of the X Window System. Users can configure one keyboard layout (for instance, Danish: dk) or several layouts to switch between (for example, us and dk).

• With some keyboard layouts, for instance sk, hu, and cz, pressing the decimal separator on the numeric keypad not only enters a delimiter but also deletes the previous character. This is due to a native bug. A workaround is to use two layouts, for example, us and sk. In this case the numeric keypad works correctly in both layouts.

• On UNIX systems that support dynamic keyboard changes, a running Java application does not recognize such a change. For instance, changing the keyboard from US to German does not change the keyboard mapping. Although the X server detects the change and sends out a MappingNotify event to interested clients, AWT does not refresh its notion of the keycode-keysym mapping.