[Top] [All Lists]

Re: rfc2821bis-01 Issue 18: Usability of 1yz replies

2007-04-23 13:45:45

Tony Hansen wrote:
Tony Hansen wrote:
A MUST will support and codify current behavior. A SHOULD will not and
can not. Therefore a MUST is necessary to move 2821bis to draft standard.

<process hat on>

I'm going to declare consensus on this issue. There has been much
discussion, and not everyone is totally happy with the final
conclusions, but enough evidence has been brought forward that I am
convinced that this is the only viable way to move forward to draft

</process hat off>

When I brought up the issue, I gambled that this would be the ultimate conclusion. I wasn't going to bring it up knowing perfectly well there will be perfectly logically reasoning to lock this down. However, my 1st mistake was mis-reading you as I thought you were be the #1 person who would understand the entire picture because of OPES. If this synergy was not present, I would be 100% on board with the consensus. My 2nd mistake was underestimated how broken the few but very popular code are which is unfortunately enough to make endorsing decisions which is not otherwise done. Its unfair, but thats reality.

In any case, I will be happy with this if the TEXT makes it very clear for clients to be more correct on reading its response codes.

In other words, they should FIX their code and simply stating servers MUST send persistent reply codes regardless if they are continuation lines or not should be coupled with clients focusing on the last reply-code[sp text] ABNF. Otherwise, there is no point in any text regarding multi-line parsing.

        The first 3 digit in the receive buffer is
        the primary response code.


This reminds me of the similar issue with the white space after the MAIL FROM: command. An undocumented acceptable practice.

Thanks Tony (and John) for your patience.


<Prev in Thread] Current Thread [Next in Thread>