Re: Gen-ART review of draft-ietf-qresync-rfc5162bis-09

Alexey Melnikov <alexey.melnikov@xxxxxxxxx> · Thu, 30 Jan 2014 18:19:17 +0000

Hi David,

On 27/01/2014 14:20, Black, David wrote:
Hi Eliot,

Regarding line length:

As the draft indicates, this is problematic for the two extensions in
question. We have discussed this limitation in the working group, and
nobody seems to be concerned.  That's due to two factors, I think:
first, RFC 2683 is not normative (informational).  I am unaware of any
strict line length limitation in RFC 3501, for instance.  Instead, there
is a requirement for strict syntax parsing.  If the client blows it in
any way, the server SHOULD return an error with a BAD response.
The latter two sentences would be good to add to the draft in some form.

My primary concern is what happens if the client sends a rather long
line to a server that isn't prepared to cope, and the latter sentence
would address it.  That could be accompanied by some sort of statement
that servers that implement this extension need to be prepared to cope
with longer lines.
I added some text on this issue. It will appear in -10 which I will post 
shortly.
I'll leave the downref as a process concern for you (WG chair) and Barry (AD)
to work out.

Thanks,
--David

-----Original Message-----
From: Eliot Lear [mailto:lear@xxxxxxxxx]
Sent: Monday, January 27, 2014 8:01 AM
To: Black, David; Alexey Melnikov (alexey.melnikov@xxxxxxxxx);
dcridland@xxxxxxxxxx; General Area Review Team (gen-art@xxxxxxxx)
Cc: Barry Leiba (barryleiba@xxxxxxxxxxxx);ietf@xxxxxxxx;imapext@xxxxxxxx
Subject: Re: Gen-ART review of draft-ietf-qresync-rfc5162bis-09

Hi Dave,

First, thank you for your review.  I personally appreciate to the time
you put in to all of your reviews.

On 1/27/14, 3:06 AM, Black, David wrote:
I am the assigned Gen-ART reviewer for this draft. For background on
Gen-ART, please see the FAQ at

<http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>.

Please resolve these comments along with any other Last Call comments
you may receive.

Document: draft-ietf-qresync-rfc5162bis-09
Reviewer: David L. Black
Review Date: January 26, 2014
IETF LC End Date: January 27, 2014

Summary:
This draft is basically ready for publication, but has minor issues that
should be fixed before publication.

This draft consolidates RFC 4551 (IMAP Conditional STORE) and RFC 5162
(IMAP Quick Mailbox Resynchronization) into a single document and makes
a few minor updates.  It also updates the command line length
recommendation in RFC 2683 to support the longer command lines that
can occur with these extensions.

As this is an update, I checked the diffs against RFC 4551 and RFC 5162;
they look reasonable.  I found one minor issue that should be relatively
easy to address.

Minor Issues:

The command line length change in Section 4 applies to all IMAP commands,
and hence affects old servers, including those that don't implement either
of the extensions in this draft. That raises a backwards compatibility
concern - what happens when a new client sends a > 1000 character command
line to an old server that isn't expecting more than 1000 characters?
As the draft indicates, this is problematic for the two extensions in
question. We have discussed this limitation in the working group, and
nobody seems to be concerned.  That's due to two factors, I think:
first, RFC 2683 is not normative (informational).  I am unaware of any
strict line length limitation in RFC 3501, for instance.  Instead, there
is a requirement for strict syntax parsing.  If the client blows it in
any way, the server SHOULD return an error with a BAD response.

This having been said, while the working group has discussed this
matter, it would be good to hear any additional voices of concern, in
the interests of interoperability.

This takes us to your next point:

One possibility would be to apply the larger length recommendation only
after the client determines that the server supports at least one of
these extensions.

Nits/editorial comments:

The update to RFC 2683 would be easier to find from the table of contents
if the title of Section 4 were changed to "Long Command Lines (Update to
RFC 2683)".

idnits 2.13.01 got confused by the command line examples, and flagged a
downref:

   ** Downref: Normative reference to an Informational RFC: RFC 2683

That downref appears to be ok and intended, as noted in the shepherd's
writeup.
This will be addressed in a forthcoming update.  Barry has "educated" us
that really this is NOT an update to RFC2683, where the advice given in
that document is superceded by a specific option.  As such, this error
will evaporate.