Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Related to:

- https://github.com/element-hq/element-web/pull/29471
- 16f9e7dd46

Newer Element Web versions allow for the nginx port to be
overriden, etc., and provide instructions for running in read-only mode.

This makes our custom `nginx.conf` patches unnecessary.

Passing the correct `ELEMENT_WEB_PORT` environment variable
also helps with future changes.

Another benefit of this (besides keeping closer to upstream
recommendations and the improved simplicity) is that:

- the container can run its entrypoint env-substitutions code now,
  without reporting errors

- IPv6 for nginx works, so `matrix-client-element:8080` is accessible
  via IPv6 on the container network now
  (this affects only for Traefik's communicaton with Element Web
  internally; public connectivity was handled by Traefik and IPv6 was
  available there even before)

Ref:

- 2052080d7d/docs/install.md (docker)
- https://github.com/element-hq/element-web/pull/28849
- https://github.com/element-hq/element-web/pull/28840

This reverts commit 178f7a9fff.

Element v1.11.95 doesn't seem to work. The container starts and is
unhealthy.

Until this is investigated and fixed, reverting.

Previously `https://example.com` and `https://example.com/` would be
redirected to `https://matrix.example.com`.

However, `https://example.com/?something` was not handled and was tried
to be served as a file by matrix-static-files for this base domain.

From now on, paths containing query strings will be handled as well
and the query string will be passed along.
Example: `https://example.com/?something` -> `https://matrix.example.com/?something`

Since nginx 1.27.3, we can make use of the `resolve` parameter for an `upstream`'s `server`,
to allow DNS resolution to happen continuously at runtime, not just once during startup.

Previously, this was not possible to do in an `upstream` block without
an nginx-plus subscription. Outside of an `upstream` block, we've used
and still use `set $backend ..` workarounds to get DNS resolution at
runtime, but now we can do it in `upstream` as well.

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

This commit adds the hint to fetch Ansible roles which have been replaced, as it is pretty easy to overlook necessity of doing so, especially if you had not enabled (and disabled thereafter) roles, whose variable names has been changed.

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>

Signed-off-by: Suguru Hirahara <acioustick@noreply.codeberg.org>