On 4 December 2017 at 11:36, Jonas Wielicki <[email protected]> wrote: > On Montag, 4. Dezember 2017 09:53:59 CET Daniel Gultsch wrote: >> I was considering to just move it up by one level. Validation may or >> may not be an issue but I figured some libraries won't have a >> getExtension or getChildren method on the error class if they don't >> expect the class to have children. >> Are there any other examples of XEPs 'extending' the error element? > > FWIW, the semantics of "it can be retried later" are essentially encoded in > type="wait". > > Extending the error element in this way feels weird, because as you say it is > not a condition but an instruction, and implementations won’t expect this.
Agreed. This is exactly what the error types are for. Regards, Matthew _______________________________________________ Standards mailing list Info: https://mail.jabber.org/mailman/listinfo/standards Unsubscribe: [email protected] _______________________________________________
