Try refetching bridge descriptors each second, not each ten seconds
Currently we only call fetch_bridge_descriptors() from main.c inside the "time_to_try_getting_descriptors < now" clause, which happens every 10 or 60 seconds depending on whether we 'have enough dir info'.
I think that schedule is fine for normal descriptors (because we're trying to batch requests), but I think for bridge descriptors we could retry every second without much harm and with some benefit.
In particular, when we setconf a new bridge, we currently need to wait 10ish or 60ish seconds before we'll even start its descriptor download (see legacy/trac#2355 (moved)).
I'd originally thought to call fetch_bridge_descriptors() from inside options_act() as a special case, but now I think we should just call it every second. (It looks through a smartlist, does a download_status_is_ready() check on each entry, and returns.)