Kind of OT: guess what I found!

[Steve Freegard]

David Lee wrote:
> I wonder whether if might be the "Message-ID" in Julian's emails?
> 
> Here are some of the headers.  The 'Message-ID' is wrapped ("folded" in
> RFC822 terminology), contains spaces, commas and a tab;  the tab is
> probably the result of the folding.  RFC822 suggests that 'msg-id'
> should be "<" addr-spec ">".  (As I read it, if its syntax wouldn't be a
> valid email address, then it wouldn't be a valid 'msg-id' either.)
> 
> Chasing down the syntax from there suggests that such a 'Message-ID'
> would seem borderline at best; my suspicion is that it is beyond the
> borders. Even if such a Message-ID is technically legal, the "be
> conservative in what you send" principle might suggest considering a
> revision.

Hmmm - we recently made changes to the method used for modifying the
Message-ID; however something appears to be mangling it.  The actual
Message-ID is sent as:

<EMEW,l1KAsk0c15677e2728c1a4bcf90ab66afb3922,MailScanner%ecs.soton.ac.uk,499FDD6E.9030908 at ecs.soton.ac.uk>

e.g. there are no spaces in the string (as you correct point out this
would not be RFC valid if there were spaces within it).

I've had several e-mail to my account directly from Jules and I've been
through all of them; none of the MIDs have any spaces in them either.

So something is mangling these on input to the list (MailMan??).  I'll
do some investigation and see if I can replicate this behaviour and see
if I can change the delimeter to something other than a comma.

Kind regards,
Steve.