I think this is absolutely preposterous. I think that the uniqueness
requirement means that the string "--outer<CRLF>" must not appear as body
contents, but that "--outer-inner<CRLF>" is perfectly OK ...
I think you're absolutely right, Steve. But I must confess to
having written code that would fail to properly parse a "lenghtened
boundary" multipart as you would generate. (because I was using a
stock library instead of counting bytes)
What say you all? Have I misread the RFC? Has the other vendor?
Others have already pointed out the weakness in the wording
of the RFC. I would urge you to "code for the worst" in the future,
but whether or not you hand-out free Pentiums depends upon how many
customers complain and how big those customers are. Really, *I*
don't think you've done anything to be embarrassed about in your code.
Does the RFC need to be clarified?
(I still haven't seen the rest of the thread to know the consensus)
Since you used the <CRLF> terminology, I really REALLY think
that trailing white space should be ignored IN PROCESSING. That is,
you don't want to throw it away; keep every byte you get. But when
scanning, don't count trailing white space for boundary uniqueness.
Other than that, you've done quite well.
--
Steve Dorner, Qualcomm Incorporated. "Oog make mission statement."
--
Rick Troth <troth(_at_)ua1vm(_dot_)ua(_dot_)edu>, Houston, Texas, USA
http://ua1vm.ua.edu/~troth/