[mailmaint] Re: WGLC on draft-ietf-mailmaint-imap-uidbatches-18

[Alexey Melnikov <[e-mail]>]

Hi all,

On 23/10/2025 11:24, Arnt Gulbrandsen wrote:
> October 23, 2025 at 1:41 PM, "Ken Murchison" <[e-mail] 
> <mailto:[e-mail]?to=%22Ken%20Murchison%22%20%3Cmurch%40fastmail.com%3E>> 
> wrote:
>
>     Arnt, Alexey,
>
>     Since you both have read the doc and provided feedback previously,
>     are you ok with the current version?
>
>
> Yes.

I am not entirely. I think the document went too far the other way to 
become unnecessarily flexible on server side.

More specifically:

In 3.1.3.1. Batch Size Summary

    *  Servers SHOULD return batches close to the requested size (>= 90%
       and <= 110% when possible)

Why would server ever return more data than asked by the client, if the client is not interested in it? This waste bandwidth and would be easily enough to filter out on the server before sending.

(also in the table above this text)


In 3.1.3.3.1.  Hard Constraints

    If the server supports the [RFC9738] MESSAGELIMIT and UIDBATCHES
    extensions, the server MUST NOT return ranges that contain more than
    the number of messages per batch requested by the client.  This is a
    strict upper bound that cannot be exceeded.  See Section 3.2 for
    details.

So this text was not tied to MESSAGELIMIT extension in the previous 
revision. Why is this restriction tied to MESSAGELIMIT?

Not having batch sizes larger than announced MESSAGELIMIT would make 
sense, but this is not what the above text is saying.


Best Regards,

Alexey