2001-08-09 jacob berkman <jacob@ximian.com> * configure.in: bumped up to 0.10.99.3 * gal/e-text/e-completion-view.c (e_completion_view_init): init matches array (e_completion_view_destroy): free matches array (e_completion_view_disconnect): disconnect from the clear and lost_match signals (e_completion_view_clear_choices): fixup for ptr array (e_completion_view_set_cursor_row): ditto (e_completion_view_select): ditto (e_completion_view_key_press_handler): ditto (clear_completion_cb): handle the clear signal (lost_completion_cb): remove a row from the model (table_row_count): fixup for ptr array (table_value_at): ditto (e_completion_view_construct): connect to clear and lost completion signals (e_completion_view_set_width): ptr array fixup again * gal/e-text/e-completion.c (e_completion_class_init): add new clean / lost completion signals (e_completion_init): initialize the matches array (e_completion_destroy): free the matches array (e_completion_add_match): fixup for matches being a GPtrArray (e_completion_clear_matches): ditto (e_completion_clear): new function to clear the list of matches (e_completion_foreach_match): fixup for ptrarray (e_completion_restart): ditto (e_completion_found_match): if we are at the limit return first, rather than adding and then unreffing. there very well could be some weird crashes this caused but it is kind of unlikely. (e_completion_lost_match): new function to forget about a match (e_completion_end_search): add a comment on why this is broken * gal/e-text/e-entry.c (e_entry_text_keypress): only start a delayed completion if our completion timeout is >= 0. this provides a way for the completion starting / stopping to be controlled by something other than the EEntry (ie, file-sel) svn path=/trunk/; revision=11843