Re: Open API bugs, 29 June

[Brian Cameron <Brian Cameron sun com>]

Bill & Others:

> >56551  GtkTreeView row_delete bug doesn't tell you enough info [?]
> This one is still live, there is an asymmetry between insert/delete 
> events if you collapse+expand a tree row, which means that n_old_rows - 
> n_deleted_rows != n_new_rows, but n_rows+n_added_rows == n_new_rows.  
> Seems broken because of the asymmetry.

This issue has been resolved.  Jonathan added a function to GtkTreeView 
gtk_tree_view_set_destroy_count_func() which returns the proper
number of visible rows that have been deleted.  The bug is marked
as RESOLVED, which I think is correct.

Jonathan and I are currently discussing whether this function should
return the number of visible rows that were deleted or the number
of visible child rows that were deleted.  Either the docs or the
function should be changed.  Either way, this doesn't really affect
the API.

There is a separate assymetry issue with insert/delete signals
when the user does drag & drop operations.  That is bug #56467

  http://bugzilla.gnome.org/show_bug.cgi?id=56467

This bug is still open and should probably be fixed.  I do not
think it is an API issue though.  Just the wrong number of 
GtkTreeModel "delete" signals being generated.

Hopefully I've cleared things up.

Brian