You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Under certain conditions and with some implementations, a message (the one in the payload) might be missing either 'from' or 'to' attributes, because those were missing on the original message and were not sanitized by the server.
One example from prosody's carbon code, where a 'sent' carbon of a message-to-myself ends up without a 'to':
The exact semantics of filling-in the right 'from' and 'to' attributes are defined within the scope of a session, but not for messages.
I propose to change XEP-0280 to either require the server to set both 'from' and 'to' on the inner message (good), or to add a section to the Business Rules that describes how a client has to work around this, because, frankly, it's a complex matter and I'm sure many clients will be doing it wrong.
As I still have a pending PR on 0280, #434, I could imagine adding the change as part of that breaking change, as soon as a way forward is determined there.
The text was updated successfully, but these errors were encountered:
Under certain conditions and with some implementations, a message (the one in the payload) might be missing either 'from' or 'to' attributes, because those were missing on the original message and were not sanitized by the server.
One example from prosody's carbon code, where a 'sent' carbon of a message-to-myself ends up without a 'to':
The exact semantics of filling-in the right 'from' and 'to' attributes are defined within the scope of a session, but not for messages.
I propose to change XEP-0280 to either require the server to set both 'from' and 'to' on the inner message (good), or to add a section to the Business Rules that describes how a client has to work around this, because, frankly, it's a complex matter and I'm sure many clients will be doing it wrong.
As I still have a pending PR on 0280, #434, I could imagine adding the change as part of that breaking change, as soon as a way forward is determined there.
The text was updated successfully, but these errors were encountered: