Fedora 35 is:

- `ansible_os_family = 'RedHat'`
- `ansible_distribution_major_version = '35'`

Our RedHat checks against v7/v8 are really for RHEL derivatives (CentOS, Rockylinux,
AlmaLinux), but the same checks (by coincidence) apply for Fedora 35.

The problem is that `'35' > '7'` (comparing these as strings) is
`false`.

This patch makes sure that we always cast
`ansible_distribution_major_version` to an integer.

Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/1610

The value of `off` was taken to be a boolean, but it shouldn't be.

Synapse expects a string (currently one of: `all`, `invite`, `off`).

This also removes the `matrix_synapse_version_arm64` variable we've
been dragging around for a long time.

Since https://github.com/matrix-org/synapse/pull/11810, a multiarch Synapse
container image (for AMD64 and ARM64) is released at the same time.

Not hardcoding 'CentOS' and using the OS family ('RedHat') instead,
we now behave better on Rockylinux and AlmaLinux, etc.

With that said, we may or may not fully support CentOS/Rockylinux/AlmaLinux v8 yet.
Certain things were improved in
https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/300.

v8 support is discussed here: https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/300
Certain things (firewalld?) may still be problematic. This patch does not try to address those.
If the remaining issues are confirmed to be fixed in the future, we can mark v8 as supported.

Fixes https://github.com/spantaleev/matrix-docker-ansible-deploy/issues/1605

Fixes a regression introduced in ccb85b31a4. The
`docker.io/devture/email2matrix:1.0.2` container image was faulty due to
this 49f966d6cc

Reverts b1b4ba501fdfaa, 90c9801c560b6, a3c84f78ca9c65a, ..

I haven't really traced it (yet), but on some servers, I'm observing
`ansible-playbook ... --tags=start` completing very slowly, waiting
to stop services. I can't reproduce this on all Matrix servers I manage.
I suspect that either the systemd version is to blame or that some
specific service is not responding well to some `docker kill/rm` command.

`ExecStop` seems to work great in all cases and it's what we've been
using for a very long time, so I'm reverting to that.

1.0.2 is the first container image tag that is available as a multi-arch image
with support for linux/amd64, linux/arm64/v8 (arm64) and linux/arm/v7 (arm32),
so self-building is no longer necessary on all these platforms.

4.95-r0-1 was problematic, because `/etc/exim/exim.conf` in the
container had the wrong permissions (writable by the `exim` user).

Fixed in 697f3cff7e
which is built as 4.95-r0-2

4.95-r0-1 is the first container image tag that is available as a multi-arch image
with support for linux/amd64, linux/arm64/v8 (arm64) and linux/arm/v7 (arm32),
so self-building is no longer necessary on all these platforms.

2.2.3 is the first container image tag that is available as a multi-arch image
with support for linux/amd64, linux/arm64/v8 (arm64) and linux/arm/v7 (arm32),
so self-building is no longer necessary on all these platforms.

Not tested but shouldn't break something

The OAuth credentials method seems to be the only viable way to
configure the mx-puppet-bridge now. Legacy tokens can no longer be
created, and the other methods (xoxs and xoxc tokens) come with warnings
about them being against Slack's terms of service.

(untested change, but unlikely to cause breakage)

Synapse v1.51.0 requires to be built with BuildKit since
https://github.com/matrix-org/synapse/pull/11691

The `docker_image` Ansible module does not support BuildKit
(https://github.com/ansible-collections/community.general/issues/514),
so we had to switch to a `docker build` call.

Fixes a issue with room version 1.

Merging has to wait until ARM images are built ofc.

using b33ea48

v1.50.0 was found to be buggy for people using a `webclient` listener.
This is fixed in v1.50.1.

We don't use such a listener, so we weren't affected anyway.