Open jplatte opened 4 years ago
This could probably be a comment on https://github.com/matrix-org/matrix-doc/issues/667
Also MSC1597, which appears to be merged, clarified this and said it can't be empty. Why hasn't it landed in the spec? https://github.com/matrix-org/matrix-doc/blob/rav/proposals/id_grammar/proposals/1597-id-grammar.md
MSC1597 is still sitting in the implementation-needed pile: https://github.com/matrix-org/matrix-doc/issues/1597
There are now both room IDs and room alias IDs with an empty localpart federated to matrix.org AFAIK. Looks like MSC1597 has to be closed (reverted?). I've allowed them in Ruma too.
@jplatte the state of matrix-org/matrix-spec-proposals#1597 has been changed to in review so I think it hasn't been agreed on. Also if something like that is accepted it would be like capitals in mxids, where a spec compliant implementation shouldn't make any more and the ones that exist have to continue to work.
The spec says that room alias and group identifier's localparts must not be empty, but for room ids (and old-style event ids) this is not stated. Is
!:matrix.org
a valid room ID?