On 8 Jun 2000, Jason L Tibbitts III wrote:
> >>>>> "DL" == Dan Liston <firstname.lastname@example.org> writes:
> DL> I would be happy to forward a file of nearly 4000 bounced messages for
> DL> you to parse, examine differences in headers, error messages,
> DL> fatal/non-fatal, etc. Just say the word, and I will email or ftp the
> DL> "large" mbox file to you.
> This would be nice; feel free to mail the file to me.
> Although it would be even nicer if someone (i.e. besides me) could do some
> of the legwork by figuring out which of these bounces we handle properly,
> which we don't, and how we can parse them. It's pretty simple to write a
> parse routine; there are already several in the source and I'm adding
> support for the broken DSNs generated by Netscape Messaging Server (no
> per-recipient-fields as required by RFC1894) and various MTAs that put
> things like "250 OK" as the failure diagnostic (AOL,Yahoo). Once these are
> done, I'll see about 95% of the bounces I get being parsed properly.
> I can do some of it, but going through 4000 bounces is going to bog me down
> a bit.
Stupid question, but how many of those 4000 are 'unique'? Maybe Dan could
do an initial pass through and determine that? How about some sort of doc
on what we already know? MTA, Message, etc? Then ppl could scan that and
send in anything "new" they come across?