Build some automatic structure for downloading of paged start urls

[redhog]

Many lists of documents are paged, and this could easily be abstracted by the docrepo base class, possibly even optimizing restarts of downloads by going directly to the last page used when a download has been interrupted by e.g. an error (or the user pressing ctrl-c).