[Top] [All Lists]

Re: New I-D: draft-koch-subject-tags-considered-00.txt

2004-11-19 15:32:00

moore(_at_)cs(_dot_)utk(_dot_)edu (Keith Moore)  wrote on 19.11.04 in 

Here is a Sieve language snippet for
identifying an ietf-822 list message:

if anyof (address :contains ["From", "Sender", "Return-Path", "To",
"CC", "Bcc"] "ietf-822", header :contains ["List-Post", "List-ID"]
"ietf-822") {
        fileinto "INBOX.IETF-822";

arrgh.  if this is sufficient for your purposes, great - but it's not a
good idea in general to recommend using header fields for this purpose.
it will not do the right thing with forwarded or resent messages or
messages sent to multiple lists to which you are subscribed.

you really want to file things based on envelope information, like MAIL
FROM or (with subaddresses) RCPT TO.

My personal experience is that, no, I do *NOT* want to sort on envelope in  
preference to List-Id - envelope information, in my experience, is  
significantly less reliable than List-Id (if there is a List-Id - and a  
pox on lists where there isn't!).

Any use of the address the list uses to reach you can have any amount of  
spam coming in, not to mention legitimate answers, completely independent  
of if you even used it in a list context or just forgot to change it to  
whatever other context you mistakenly used it in. None of those will have  
the List-Id: set, however.

And then think of all the people - still the vast majority, I suspect -  
who *can't* use list-specific addresses.

Or did you mean envelope-*from* (Return-Path:)? That breaks even more  
easily. A simple list-side mail system change that's completely unrelated  
to the list might change that!

There's a reason we have List-Id:.

I do agree that other headers are less useful - I just include envelope  
fields in that category.

MfG Kai