Jason L Tibbitts III wrote:
> EC> As I said, I think trying to do full RFC822 support is a lost cause.
> EC> 822 is IMHO way too bizarre. I think it is a case of trying to please
> EC> everyone, and in the end missing the point of the entire exercise.
>
> To play devil's advocate, this is a mailing list program. It's supposed to
> support mailing to various and sundry addresses, regardless of their level
> of bogosity so long as they can be sent out by your MTA. We should not try
> to limit the addresses to which mail can be sent.
Sorry, I was refering to 822 being a case of trying to please everyone
and in the end missing the point of what an RFC is supposed to do
(propose a _standard_ way of doing something). Majordomo should try to
support anything _reasonable_ from 822 but I don't think we should go
any further than that.
> Well, it helps those who are stuck behind weird systems.
After trying to clean up my large mailing list and having to deal with
said weird systems, it gives me a whole new perspective on what should
be done to support them. I think a thermonuclear strike would be the
best bet :-)
> 1. 822 allows the following: jason. (mj_weenie) tibbitts@hpc.uh.edu, with
> the parenthesized text and surrounding space being an inconsequential
> comment. This is disgusting, and not allowed in 1036; parenthesized
> comments are only allowed to come after the address.
>
> 2. 1036 defined what you can't have in commented text, and what you
> probably shouldn't have in there.
>
I fully understand what 1036 specifies and IMHO it is a _very good
thing_.
Regarding the first case, I did a search of mail sent in to majordomo@
and the -request addresses since May 29, 1996. On 423700 messages, I
found some truly bizarre addresses. In a preliminary search, however,
I wasn't able to find an example of the case you illustrated. The vast
majority are in the three 1036 categories, though some have the
following modification:
From: Foo Bar <foo@bar> (Yes, it is foo, and it is bar, at the same
time!)
In order to get something reasonable, we may have to do something like
kill everything in brackets.
Regarding the second, it is to make it easier to find which part of the
address actually contains the e-mail address. Ie: it doesn't allow @ in
a comment, so you could do searches to grab out the only word with an @
in it. 1036 sets restrictions on addresses in other ways, however, that
make it easier to pick out the address part of the string. We also
can't depend on an @ being in a valid e-mail address due to local users
on mailers that don't append a domain to all mail.
> The following address is _not legal_ in RFC1036: <tibbs@hpc.uh.edu>, but I
> think that it's necessary to allow it.
That is one exception that I would be perfectly happy supporting.
Evan
--
Evan Champion * Director, Network Operations
mailto:evanc@synapse.net * Directeur, Exploitation du reseau
http://www.synapse.net/ * Synapse Internet
References:
|
|
