#8037 changed the default `autoescape` option when rendering Jinja2 templates from `False` to `True`. This caused some bugs, noticeably around redirect URLs being escaped in SAML2 auth confirmation templates, causing those URLs to break for users.

This change returns the previous behaviour as it stood. We may want to look at each template individually and see whether autoescaping is a good idea at some point, but for now lets just fix the breakage.

1.19.3

Synapse 1.19.3 (2020-09-18)
===========================

Bugfixes
--------

- Partially mitigate bug where newly joined servers couldn't get past
events in a room when there is a malformed event.
([\#8350](https://github.com/matrix-org/synapse/issues/8350))

This is a bit of a hack, as `_check_sigs_and_hash_and_fetch` is intended
for attempting to pull an event from the database/(re)pull it from the
server that originally sent the event if checking the signature of the
event fails.

During backfill we *know* that we won't have the event in our database,
however it is still useful to be able to query the original sending
server as the server we're backfilling from may be acting maliciously.

The main benefit and reason for this change however is that
`_check_sigs_and_hash_and_fetch` will drop an event during backfill if
it cannot be successfully validated, whereas the current code will
simply fail the backfill request - resulting in the client's /messages
request silently being dropped.

This is a quick patch to fix backfilling rooms that contain malformed
events. A better implementation in planned in future.

Instead of just using the most recent extremities let's pick the
ones that will give us results that the pagination request cares about,
i.e. pick extremities only if they have a smaller depth than the
pagination token.

This is useful when we fail to backfill an extremity, as we no longer
get stuck requesting that same extremity repeatedly.

c.f. #8295 for rationale

Dendrite's implementing MSC2753 over at https://github.com/matrix-org/dendrite/pull/1370 to prove the implementation for MSC purposes, and so sytest has sprouted tests for it over at https://github.com/matrix-org/sytest/pull/944. But we don't want them to run on synapse until synapse implements it.

This fixes #8319.

This takes about 10 seconds in the best case; often more.

Fix unread counts making sync fail if the value of the `unread_count`
column in `event_push_summary` is `None`.

I'm hoping this will provide some pointers for debugging
https://github.com/matrix-org/synapse/issues/7968.

All the callers want this info in the same place, so let's reduce the
duplication by doing it here.

* Fixup `ALTER TABLE` database queries

Make the new columns nullable, because doing otherwise can wedge a
server with a big database, as setting a default value rewrites the
table.

* Switch back to using the notifications count in the push badge

Clients are likely to be confused if we send a push but the badge count
is the unread messages one, and not the notifications one.

* Changelog