Re: [Imap-protocol] Working around the evils of LITERAL+

On 28 Mar 2013 18:52, "Tim Showalter" <[email protected]> wrote:
>
> On 3/27/13 7:35 PM, Lyndon Nerenberg wrote:
>
>>> I'm hesitant to suggest it, but is it worth sending BYE and dropping
the connection? It's sailing just within the lines of the specification, I
think.
>>
>>
>> I think this is perfectly valid behaviour, and well within the spec.  A
smart server might choose whether the eat-and-continue vs. bye-and-punt
based on the size of the proposed literal+, but punting is entirely valid.
>
>
> I would recommend against this, or at least advise caution.  A lot of
clients, particularly mobile phone clients of a couple years ago, are both
stupid and single-minded.  Hanging up is going to be treated as an obvious
temporary failure and they will retry.
>

Right, but to no damage. I think the best pattern is:

1) Send an tagged(?) NO, with an ALERT, explaining to the user what's
happening.

2) Send an untagged BYE, with the same message.

3) Terminate the connection. I suspect a partial shutdown might help get
the message through in some cases, but I'm not sure it's worth the pain.

Finally, you want to trigger this behaviour only for abusive cases -
multi-megabyte messages you'd have to reject anyway, not over 50k, however
desirable that might be.

By all that's holy - that is, RFC 3501 - the error will get to the user
and/or developer. At the least the constant failure will cause some users
to contact the developer. The weight of the Google Gorilla should be
sufficient to get the client fixed.

Dave.

_______________________________________________
Imap-protocol mailing list
[email protected]
http://mailman2.u.washington.edu/mailman/listinfo/imap-protocol