+2002-01-16 Bill Haneman <bill.haneman@sun.com>
+
+ * registryd/deviceeventcontroller.c:
+ Regression fix for keylistener de-registration; global keygrabs
+ were not being released when deregisterKeystrokeListener was
+ called in cspi, since
+ Accessibility_DeviceEventController_deregisterKeystrokeListener
+ was called with a zero-length keyset. That is because the cspi
+ method, SPI_deregisterKeystrokeListener, does not take a keyset
+ parameter but instead should remove all the key grabs held be a
+ keystroke listener.
+ The code in impl_deregister_keystroke_listener was changed to copy
+ the keylist from the listener instance previously registered with
+ the DeviceEventController before releasing the grabs.
+
+ * registryd/registry.c:
+ * registryd/deviceeventcontroller.c:
+ * libspi/spi-private.h:
+ * libspi/util.c:
+ Changed spelling of "re-enterant" to "re-entrant" globally.
+
+2002-01-16 Bill Haneman <bill.haneman@sun.com>
+
+ * test/test-simple.c:
+ Changed key listeners test - the test was slightly mis-using the
+ (admittedly poorly documented) SPI_generateKeyboardEvent API.
+ It now uses '=' as the key event listened to and generated,
+ relying on a keysym match rather than assuming that keycode 33
+ is always equal to keysym '!'.
+
+2002-01-15 Bill Haneman <bill.haneman@sun.com>
+
+ * test/simple-at.c:
+ Changed simple-at to use a specific keyset, rather than
+ SPI_KEYSET_ALL_KEYS - this helps minimize clashes with the window
+ manager, desktop, etc.
+
2002-01-11 Bill Haneman <bill.haneman@sun.com>
- * registryd/deviceeventcontroller.c:
+ * registryd/deviceeventcontroller.c:
Replaced standard X error handler with a special handler that
is non-fatal for failed keygrabs. This works around
problems with keygrab clashes sometimes observed on