Re: [imap5] Feature set? - was Re: Designing a new replacement protocol for IMAP

[Adrien de Croy <[e-mail]>]

On 18/02/2012 8:36 p.m., Bron Gondwana wrote:
> On Sat, Feb 18, 2012, at 08:13 PM, Adrien de Croy wrote:
>> AFAIR Bcc should never hit the wire.  That's how it can be "blind"
>> carbon copy.
>>
>> Otherwise you rely on a server to strip it.
> I entirely disagree.  I was thinking quite a bit about the "upload to
> the IMAP server and then send it via SMTP while telling it who to
> send to", and I actually don't like it.

with respect, others may like it and want to do it.

If you don't like it, you can design your server and client to behave 
the way you like without making it impossible for others to do it in the 
protocol.

others may prefer to submit the SMTP envelope information when they tell 
the server to send a message, and it could be stored as some meta-data 
on the message.

Sure, if there's a huge desire to do this then fall-back to SMTP may be 
possible... but we're doing submission in IMAP so that client authors no 
longer have to support that protocol right?

> Certainly in the XSENT I implemented, once it has sent, it sets an
> \XSent flag on the immutable message.  Not an \XSent-to-one-address.

server is free to do whatever else it likes, like moving it to sent, or 
making a copy or whatever.

>
> The message with an \XSent flag is a record that it was sent to the
> addresses listed in that header.
>
> And the copy that went "over the wire" up to the IMAP server is your
> record of what you did.  That _should_ include the BCC, so you can go
> back later and check who you sent it to.  Obviously, the BCC should
> be stripped as it gets handed over to the delivery agent.
I guess it's a grey area there - if you consider IMAP to be a remote 
mail client that you're driving from a distance, then it's ok to store a 
BCC header, just like you would if you were storing it in a local 
store.  But you're relying on the IMAP server to strip it when it 
submits the mail for delivery, which is a bit of work for the IMAP server.

If you don't have to parse and re-build messages all the time, you'll 
scale much better.

>
> Bron.

-- 
Adrien de Croy - WinGate Proxy Server - http://www.wingate.com