Re: The IMAP APPENDLIMIT Extension

Hi Pete,

On 03/01/2015 13:11, Pete Maclean wrote:
> Jay,
>
> I have not said anything on APPENDLIMIT before but, reading your 
> latest draft now, there are a few things I wonder about.  I don't 
> think any of these necessarily requires any changes to the draft but 
> some may merit additional consideration.
>
> In section 5 you make reference to "the server's fixed maximum mail 
> upload size".  What is the intended scope of said fixedness. Should a 
> client assume that a maximum upload size obtained in one session 
> applies in a later session?  I think it should not.
I don't think this matters, because the limit would be always advertised 
in CAPABILITY response and/or on SELECT anyway.
> Hardware or other changes at the server could surely lead to a maximum 
> upload size changing over time.
>
> How should a client behave if it obtains a maximum upload size for a 
> given mailbox in a LIST command and then later in the same session it 
> does a SELECT/EXAMINE for the same mailbox that quotes a different 
> maximum upload size?  For that matter, is it valid for a server to 
> return different maximum upload sizes in CAPABILITY responses within a 
> single session?  In cases where a maximum upload size is dependent on 
> a quota, it could make sense for these things to happen.
Yes, it is probably sensible to allow for that.
> Surely it would be very valuable to include in this extension a means 
> for a server to indicate that it does not accept APPENDs at all.
APPENDLIMIT=0?
> For IMAP servers fronting email archives, it is entirely reasonable 
> for APPENDs to be unsupported.  In other cases, it seems reasonable 
> for APPENDs to be disallowed for certain mailboxes (as, for example, 
> in a proxy situation or when a mailbox reaches maximum capacity).  An 
> easy way to deal with this would perhaps be to change the syntax to 
> allow an append limit of 0 with the meaning, not that zero-length 
> messages may be APPENDed (something that my checks show that many 
> servers disallow anyway), but that APPENDs are not allowed period.  
> This could avoid a lot of trouble, I believe.
Yes, I think this is a good idea.

_______________________________________________
imapext mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/imapext