Py-webdav will not allow us to have unicode in the mimetype. And that's
not in the protocol, anyway. So, ensure we are only sending a string (was
broken since the db query would convert to unicode).
bzr revid: p_christ@hol.gr-20100623140720-nzst9y9hmarfo7qv
This is a reshape of the code, since the porting. While the MultiHttpHandler
will strip the '/webdav' part of the path, the DAV code will try to
generate more paths without it. So, there must be consinstent functions
that will convert (at mk_prop_response and mk_propname_response) to the
full path or back. Also, only have one function that will strip the
db name from the path, so that we control the auth requirements (for the
purpose of database listings, which shouldn't require authentication).
bzr revid: p_christ@hol.gr-20100623105437-kmy0ccqihu5h9ydp
This is not a complete fix yet, but handles the case when one of the
scanned db's is not a proper OpenERP one (happens if a db has same user
as in the openerp conf, or if openerp db is not initialized properly).
bzr revid: p_christ@hol.gr-20100623104956-w5r3n035ozvz0h0v
P. Christeas