[Top] [All Lists]

(lack of) message header field ordering

2005-03-10 15:21:58

On Thu March 10 2005 13:21, Tony Finch wrote:

Good point. I need a word to describe the set of fields that are permitted
to appear at the start of the message header interleaved with the trace
fields. At the moment this is only Resent- fields, which is obviously not

Resent- fields are not required to be interleaved with other fields.
RFC 2822, at Proposed Standard, introduces a recommendation (N.B. not
a requirement) that Resent- fields, when used, be maintained in a block.
However, 2822 notes:

"  It is important to note that the header fields are not guaranteed to
   be in a particular order.  They may appear in any order, and they
   have been known to be reordered occasionally when transported over
   the Internet."

So a message-parsing application cannot justifiably assume that
fields appear in any particular order, or that fields have not been
reordered, both for the reasons noted in 2822 and because there exist
(a large number of diverse) applications that have not been revised to
conform to 2822.