Make sure that we delete devices whenever a user is logged out due to any of
the following situations:

 * /logout
 * /logout_all
 * change password
 * deactivate account (by the user or by an admin)
 * invalidate access token from a dynamic module

Fixes #2672.

Non-functional refactoring to move set_password. This means that we'll be able
to properly deactivate devices and access tokens without introducing a
dependency loop.

Non-functional refactoring to move deactivate_account. This means that we'll be
able to properly deactivate devices and access tokens without introducing a
dependency loop.

Whenever an access token is invalidated, we should remove the associated
pushers.

Clear logcontext before starting fed txn queue runner

Fix 500 when joining matrix-dev

matrix-dev has an event (`$/6ANj/9QWQyd71N6DpRQPf+SDUu11+HVMeKSpMzBCwM:zemos.net`)
which has no `hashes` member.

Check for missing `hashes` element in events.

These processes take a long time compared to the request, so there is lots of
"Entering|Restoring dead context" in the logs. Let's try to shut it up a bit.

Allow guest access to group APIs for reading

Fix error on sqlite 3.7

Avoid retrying forever on IntegrityError

Fix error handling on dns lookup

Tiny code cleanups

Fix some logcontext leaks in replication resource

Fix OPTIONS on preview_url

Improve tracebacks on exceptions

Use failure.Failure to recover our failure, which will give us a useful
stacktrace, unlike the rethrown exception.

pass the right arguments to the errback handler

Fixes "TypeError('eb() takes exactly 2 arguments (1 given)',)"

This is never used; let's remove it to stop confusing things.

The @measure_func annotations rely on the wrapped function respecting the
logcontext rules. Add the necessary yields to make this work.

Fixes #2706

so that the right thing happens on workers.

Try to avoid having multiple PreviewUrlResource instances

... to stop us doing the cache cleanup jobs on the master.

This avoids the scenario where we have four different PreviewUrlResources
configured on a single app, each of which have their own caches and cache
clearing jobs.

* Improve documentation of workers

Fixes https://github.com/matrix-org/synapse/issues/2554

Fixes https://github.com/matrix-org/synapse/issues/2554

Clean up dependency list

Avoid locking account_data tables for upserts

remove those that aren't used at all, and replace the ones that don't have
builders with simple getters rather than dynamically-generated methods.

Create the url_cache index on local_media_repository as a background update, so
that we can detect whether we are on sqlite or not and create a partial or
complete index accordingly.

To avoid running the cleanup job before we have built the index, add a bailout
which will defer the cleanup if the bg updates are still running.

Fixes https://github.com/matrix-org/synapse/issues/2572.

Avoid locking for upsert on pushers tables