This bug introduced a regression whereby if any authorized value has no
branch limitations then the authorized value administration page would
fail to display at all.
A simple ternary corrects the issue, but I'm now wondering about the
logic of returning 'undef' from 'get_branch_limits'.. perhaps an empty
(or full) resultset might be a better option to prevent us from
requireing ternaries before calling chained methods on such objects.
Signed-off-by: Martin Renvoize <martin.renvoize@ptfs-europe.com>