try not to drop context after federation requests

Add another missing yield on check_device_registered

Reduce some CPU work on DB threads

cursor_to_dict can be surprisinglh expensive for large result sets, so lets
only call it when we need to.

Fix token request for addition of phone numbers

Fix rejection of invites not reaching sync

preserve_context_over_fn uses a ContextPreservingDeferred, which only restores
context for the duration of its callbacks, which isn't really correct, and
means that subsequent operations in the same request can end up without their
logcontexts.

push federation retry limiter down to matrixfederationclient

Fix time_bound_deferred to throw the right exception

Always allow the user to see their own leave events, otherwise
they won't see the event if they reject an invite for a room whose
history visibility is set such that they cannot see events before
joining.

This prevents unnecessary construction of lists

JSONSchema Validation For Filters

Signed-off-by: pik <alexander.maznev@gmail.com>

 * use a valid filter in rest/client/v2_alpha test

Signed-off-by: pik <alexander.maznev@gmail.com>

Signed-off-by: pik <alexander.maznev@gmail.com>

 * add invalid filter tests

Signed-off-by: pik <alexander.maznev@gmail.com>

Add a param to the federation client which lets us ignore historical backoff
data for federation queries, and set it for a handful of operations.

Due to a failure to instantiate DeferredTimedOutError, time_bound_deferred
would throw a CancelledError when the deferred timed out, which was rather
confusing.

Update README: specify python2.7 in virtualenv

Add a missing yield in device key upload

Fix a couple of logcontext leaks

rather than having to instrument everywhere we make a federation call,
make the MatrixFederationHttpClient manage the retry limiter.

rename _create_request to _request, and push ascii-encoding of `destination`
and `path` down into it

Use preserve_fn to correctly manage the logcontexts around things we don't want
to yield on.

(this would only very very rarely actually be a useful thing, so the main
problem was the logcontext leak...)

fix up some key verif docstrings

Fix caching of remote servers' signature keys

The `@cached` decorator on `KeyStore._get_server_verify_key` was missing
its `num_args` parameter, which meant that it was returning the wrong key for
any server which had more than one recorded key.

By way of a fix, change the default for `num_args` to be *all* arguments. To
implement that, factor out a common base class for `CacheDescriptor` and `CacheListDescriptor`.

Don't send the full event json over replication

Add some debug to help diagnose weird federation issue