This patch removes the `initial` field from the `Response` for
MSC4186. There is no top-level `initial` value. This field exists for
`response::Room` though.
RFC 2046[1] is somewhat ambiguous on whether or not it's valid to omit the
preceding CRLF for the first boundary. The prose on page 19 suggests
that it is not:
> The boundary delimiter MUST occur at the beginning of a line, i.e.,
> following a CRLF, and the initial CRLF is considered to be attached
> to the boundary delimiter line rather than part of the preceding
> part. The boundary may be followed by zero or more characters of
> linear whitespace. It is then terminated by either another CRLF and
> the header fields for the next part, or by two CRLFs, in which case
> there are no header fields for the next part. If no Content-Type
> field is present it is assumed to be "message/rfc822" in a
> "multipart/digest" and "text/plain" otherwise.
>
> NOTE: The CRLF preceding the boundary delimiter line is conceptually
> attached to the boundary so that it is possible to have a part that
> does not end with a CRLF (line break). Body parts that must be
> considered to end with line breaks, therefore, must have two CRLFs
> preceding the boundary delimiter line, the first of which is part of
> the preceding body part, and the second of which is part of the
> encapsulation boundary.
But the BNF on page 22 suggests that it is, as long as there is no
preamble:
> dash-boundary := "--" boundary
> ; boundary taken from the value of
> ; boundary parameter of the
> ; Content-Type field.
>
> multipart-body := [preamble CRLF]
> dash-boundary transport-padding CRLF
> body-part *encapsulation
> close-delimiter transport-padding
> [CRLF epilogue]
Dendrite currently generates multipart responses without a preceding CRLF
for the first boundary[2], which were rejected by the previous ruma
parsing logic.
[1]: https://datatracker.ietf.org/doc/html/rfc2046
[2]: https://github.com/matrix-org/dendrite/issues/3414
Requires a complete change of the HTML tree, to have interior mutability
and be able to return borrowed data from a Handle.
It is now a Rc-based tree where each node has a weak reference to
its parent and strong references to its children.
This results ina few breaking changes to the public API.
Allow MSC3401-style call membership state events to have state keys that
are arbitrary strings, and not just pure MXIDs.
This allows for state keys that are suffixed with the sender's device
ID, as well as ones prefixed by an underscore to bypass auth rules over
state keys that start with a `@` but do not strictly equal an MXID.
