Re: patch for GtkDialog

[Havoc Pennington <hp redhat com>]

Hi,

The patch looks good to me except that we're still "eating" the Escape
key here; is that OK? Or should we be passing it through?

To pass it through, we'd need to add a new keybinding signal 
activate_cancel or the like, and return TRUE/FALSE from that according
to whether we have cancel. The new signal can be added without using a
slot in GtkDialogClass, Owen can post his hack for that maybe...

This raises a small question, should Escape result in
GTK_RESPONSE_CANCEL then, rather than GTK_RESPONSE_DELETE_EVENT?

Havoc