as discussed a while back[1], here's a way to remove the need for Glib::Object::Subclass to provide GET_PROPERTY/SET_PROPERTY. this is step one, providing a fallback; step two is adding the per-property setters and getters that would take precedence over [SG]ET_PROPERTY and this fallback. this fallback does the same as the default behavior provided by Subclass; provide property storage in the wrapper hash. this goes a step further for GETting, fetching a default value for the property if one is available. thus, the default value you give when creating properties actually has a purpose. :-) attached is the patch that adds this fallback behavior to Glib HEAD (or 1.050), and a simple driver program. the patch as provided includes some debugging output which i think is the cause for the warnings from the test suite. your comments would be greatly appreciated. [1] the thread starts here, with jan's suggestion: http://mail.gnome.org/archives/gtk-perl-list/2004-May/msg00037.html after discussion, this is roughly what i'm starting to implement: http://mail.gnome.org/archives/gtk-perl-list/2004-May/msg00048.html
Attachment:
param_test.pl
Description: Text Data
Attachment:
property-fallback.patch
Description: Text document