SwA profile header canonicalization issue and proposed resolution

[<Frederick.Hirsch@nokia.com>]

I propose we add the following issue to the issues list for the SwA Profile:

"When doing Attachment-Complete-Transform for signatures should the "<>" associated with the Content-ID value be included in the MIME header canonicalization"

This is based on the SwA interop #2 issues message that Blake sent out

http://www.oasis-open.org/apps/org/workgroup/wss/email/archives/200411/msg00082.html

I propose we resolve this issue by stating that the "<>" IS to be included, by adding the following text to draft 14 in section 4.4.1 (MIME header canonicalization), as follows:

Add a new step after step 6:

 

"7. When a Content-ID header is processed, the "<>" characters associated with the msg-id MUST be included in the transform. The reason is that although semantically these angle bracket characters are not part of the msg-id (RFC 2822) they are a standard part of the header lexographic representation."

 

It appears that in Interop the common interpretation was to include the "<>" characters.

 

(To address the Oracle concern, it was not the intent of the SwA profile to imply that the "<>" should not be included.)

 

Additionally, it seems appropriate to remove step 1 in 4.4.1  entirely since header canonicalization only applies to headers, which by definition precede the content.

 

Comments? Please send comment to the list this week so that this issue may be resolved on the list this week if possible.  This will allow a new draft to be produced with time for review before the next WSS meeting.

 

Thanks

regards, Frederick

Frederick Hirsch
Nokia