Implement result streaming for list and fetch jobs and use it in the

corresponding models. Feels much faster now since you see the first
results immediately.

svn path=/trunk/KDE/kdepim/akonadi/libakonadi/; revision=716143