Re: [exim-dev] [Bug 1066] interpret some 4xx error codes fro…

Page principale
Supprimer ce message
Répondre à ce message
Auteur: Ted Cooper
Date:  
À: exim-dev
CC: Andrey N. Oktyabrski
Sujet: Re: [exim-dev] [Bug 1066] interpret some 4xx error codes from remote server as permanent errors (5xx). Sometimes the Postfix MTA returns 4xx error when mailbox does not exist
On 17/01/11 04:29, Andrey N. Oktyabrski wrote:
> ------- You are receiving this mail because: -------
> You are on the CC list for the bug.
>
> http://bugs.exim.org/show_bug.cgi?id=1066
>
> --- Comment #17 from Andrey N. Oktyabrski <ano@???> 2011-01-16 18:29:01 ---
> (In reply to comment #16)
>> I'm sorry, but this is insane. We're talking about deliberately coding to allow
>> for blatant abuse of the SMTP protocol here. Even if it's hidden away, this
>> just allows for too much abuse.
> Yes, this idea (write any bug reports here) is insane. Theoretically, nothing
> must be added in exim. Practically, I have all necessary patches. Adieu.


I'm sorry you have taken this in a negative way as it was not meant to
be a personal attack on you or the original idea. You have provided some
very useful patches - things that have been (or will be) added to Exim -
and I thank you for that.

The insane part of it all was when talk of adding a feature to turn 5xx
into 4xx started. That's one amazing feature creep. 4xx -> 5xx was
almost tenable with an existing compatibility issue, but with special
retry rules it is already possible. However, since this has been fixed
in Postfix, the reason for the fix becomes mute. This feature then
becomes a liability and extending it to 5xx -> 4xx is two steps too far.

There is a difference between adding useful features and adding a
feature to break the SMTP protocol.

There are many things that not everyone has agreed with that have been
added to not only Exim but many MTA - SPF, DKIM, Content scanning, etc.
All of these things were never in the first versions and had to be added
in later. None of them however, degraded the protocol. Some of them took
advantage of existing features but they didn't outright rewrite or break
it. In the end, everyone still speaks the same protocol and respects
machine readable return codes. A bug like the one in Postfix was fixed
in Postfix instead of everyone writing something to get around it.

That's pretty much it for me on this subject. Again, I'm sorry if you
took my rather strong dislike of the extended idea as an attack on yourself.