Jonas, comparison can be case-insensitively and (probably also) normalization-insensitively without having to case-/normalization-convert both identifiers first. An example is strcasecmp() [1].
As JIDs are supposed to be case-preserving, I would expect several implementations do not downcase them first. I would suggest language similar to the following: "Before hashing, implementations should ensure that the JID has been converted to lower case and Unicode normalization form C (NFC)" -Marcel [1] http://man7.org/linux/man-pages/man3/strcasecmp.3.html On Fri, 2017-12-08 at 10:25 +0100, Jonas Wielicki wrote: > On Freitag, 8. Dezember 2017 10:00:33 CET Marcel Waldvogel wrote: > > I came upon one more ambiguity in the "same JID, different > > hash/color" > > domain > > First, what I already mentioned, is case: The JID must be mapped to > > a > > consistent case; I would suggest lower. > > Isn’t that mandated anyways by RFC 7622 for "comparing" JIDs (which > this is, I > think)? Surely I can add a reference that RFC 7622 normalisation must > be > applied before hashing. > > > Second, the hash also depends on the Normalization Form applied. I > > would suggest NFC, which is recommended anyway for XML. > > It would be great if that could also be specified (or at least > > referenced) in the XEP. > > I think the unicode normalization form which is to be used is also > covered by > RFC 7622, even though maybe only implicitly in PRECIS. Luckily, it is > NFC :-). > > Adding a note that the RFC 7622 rules for Comparison need to be > applied before > hashing probably makes sense. > > kind regards, > Jonas > _______________________________________________ > Standards mailing list > Info: https://mail.jabber.org/mailman/listinfo/standards > Unsubscribe: standards-unsubscr...@xmpp.org > _______________________________________________ >
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ Standards mailing list Info: https://mail.jabber.org/mailman/listinfo/standards Unsubscribe: standards-unsubscr...@xmpp.org _______________________________________________