This fixes a race whereby:
 - User hits an endpoint.
 - No cached transaction so executes main code.
 - User hits same endpoint.
 - No cache transaction so executes main code.
 - Main code finishes executing and caches response and returns.
 - Main code finishes executing and caches response and returns.

 This race is common in the wild when Synapse is struggling under load.
 This commit fixes the race by:
  - User hits an endpoint.
  - Caches the promise to execute the main code and executes main code.
  - User hits same endpoint.
  - Yields on the same promise as the first request.
  - Main code finishes executing and returns, unblocking both requests.

Don't assume providers raise ConfigError's

This only makes a difference for versions of ldap3 before 1.0, but a)
its best to be explicit and b) there are distributions that package
ancient versions for ldap3 (e.g. debian).

Limit the number of prev_events of new events

Respect use_frozen_dicts option in workers

default config: blacklist more internal ips

Remove unused but buggy function

More updates to Promethese metrics exposition

Replace postgres GIN with GIST

This is because GIN can be slow to write too, especially when the table
gets large.

Don't error on non-ascii passwords

Set CORs headers on responses from the media repo

Remove sent_transactions table.