[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: discuss: draft-ietf-pkix-pi-07.txt



--On 16. oktober 2003 04:27 -0400 "Steven M. Bellovin" <smb@research.att.com> wrote:

i think there are some rituals that have to be followed re utf8, and
they are not being followed.  but i am not an expert in this.  e.g. i
suspect this is deadly

     The Alphanumeric Identifier Match rule compares for equality a
     presented value with an attribute value of type UTF8String
     or IA5String, which is interpreted as a series of alphanumeric
     characters.  The rules for matching are that a working comparison
     value is constructed from each of the two values by including
     only the digits and alphabetic characters appearing in the value;
     and then the two comparison values are compared using
     CaseIgnoreMatch.  This rule is intended for use only with
     identifiers in variants of the Latin, Greek, and Cyrillic scripts.


I raised that issue the first time we saw that document.  If I recall
correctly, Leslie -- who knows far more about that stuff than I do --
persuaded me that it wasn't a problem here.

I'm holding a DISCUSS on this paragraph. If they solve the problem by limiting the string to ASCII, that's fine with me.

Note (something I realized only on re-reading Leslie's comment): this matching rule will match http://com.net.org/ with http://com.net/org. This may be a bad idea.

Harald