summaryrefslogtreecommitdiffstats
path: root/tests/auto/atwrapper/kramer.ini
diff options
context:
space:
mode:
authorJason Barron <jbarron@trolltech.com>2009-09-28 15:52:22 (GMT)
committerJason Barron <jbarron@trolltech.com>2009-09-29 07:06:27 (GMT)
commit8aec69a5cae285174b41df6d268f007edcdec84b (patch)
tree588ec77100d8cff6381674de2d63de0787b6ad5e /tests/auto/atwrapper/kramer.ini
parentff6c8f9439630c2ea3bf78d5568a44521e915898 (diff)
downloadQt-8aec69a5cae285174b41df6d268f007edcdec84b.zip
Qt-8aec69a5cae285174b41df6d268f007edcdec84b.tar.gz
Qt-8aec69a5cae285174b41df6d268f007edcdec84b.tar.bz2
Add a more proper shutdown sequence for Symbian applications.
Previously both EEikCmdExit and EAknSoftkeyExit were hard coded to call exit() which promptly exited the event loop without the chance to do any UI operations. With this patch, we handle the two cases differently and allow people to essentially make the right soft key a hide instead of an exit simply by reimplementing closeEvent() and ignoring the event. EAknSoftkeyExit is now a soft exit since this does not seem to be a mandatory exit in S60. By sending a QCloseEvent to QApplication this has the effect that all windows are closed (sent another QCloseEvent) and if one of them chooses to ignore this event, the application and UI are both kept alive. This is similar to how Qt behaves on other platforms. If all windows accept the close event (default), then the application will quit() as normal. This command is sent from the right soft key. EEikCmdExit is mapped similarly as before, but calls quit() instead of exit() to be aligned with other platforms. This will cause the aboutToQuit signal to be emitted which maps very well to Qt because UI interaction is not permitted from this signal and wouldn't work in S60 anyway because in the case of EEikCmdExit, the UI has already been torn down by the UI framework. This command is sent by the End key. We could take this one step further and intercept the key event and prevent it from going to the UI framework and tearing down the UI, but this is a little too evil IMHO. If people really want to intercept that event, they can do it by reimplementing the platform specific event filter. This is equivalent to how it would be done in S60 (reimplementing HandleWsEvent). Reviewed-by: axis Reviewed-by: mread
Diffstat (limited to 'tests/auto/atwrapper/kramer.ini')
0 files changed, 0 insertions, 0 deletions