Resent-Date: Mon, 15 Jul 2013 19:45:00 +0000
Resent-Message-Id: <E1Uyohg-00037W-Cv@frink.w3.org>
MIME-Version: 1.0
Sender: patrick.ducksong@gmail.com
Date: Mon, 15 Jul 2013 15:44:24 -0400
Message-ID: <CAOdDvNoPRmM-8hpbrCoQ4GQFJTd0qPjONTyJuF6Pu2UhoyQ_zA@mail.gmail.com>
From: Patrick McManus <pmcmanus@mozilla.com>
To: HTTP Working Group <ietf-http-wg@w3.org>
Content-Type: multipart/alternative; boundary=047d7b3a928668f7ac04e1921617
Received-SPF: pass client-ip=209.85.214.177;
 envelope-from=patrick.ducksong@gmail.com; helo=mail-ob0-f177.google.com
Subject: SETTINGS error handling
Archived-At: <http://www.w3.org/mid/CAOdDvNoPRmM-8hpbrCoQ4GQFJTd0qPjONTyJuF6Pu2UhoyQ_zA@mail.gmail.com>
Resent-From: ietf-http-wg@w3.org
Resent-Sender: ietf-http-wg-request@w3.org
Precedence: list

--047d7b3a928668f7ac04e1921617
Content-Type: text/plain; charset=ISO-8859-1

Section 6.5
   single SETTINGS frame MUST NOT contain multiple values for the same
   ID.  If the receiver of a SETTINGS frame discovers multiple values
   for the same ID, it MUST ignore all values for that ID except the
   first one.



I'm wondering why the text proscribes error handling of MUST ignore in
response to violation of the MUST NOT send provision. I'd rather my
implementation be strict and enforce the MUST NOT send via PROTOCOL ERROR
when receiving this. The kindler gentler days of being liberal in what we
receive have sailed. There is not yet an argument for backwards
compatibility to not do that, but if I did so I would also be in non
compliance.

Can we either change it to PROTOCOL ERROR (preferred) or just be silent on
handling of the error?

-Patrick

--047d7b3a928668f7ac04e1921617
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div><div><pre class=3D"">Section 6.5
   single SETTINGS frame MUST NOT contain multiple values for the same
   ID.  If the receiver of a SETTINGS frame discovers multiple values
   for the same ID, it MUST ignore all values for that ID except the
   first one.</pre><br><br>I&#39;m wondering why the text proscribes error =
handling of MUST ignore in response to violation of the MUST NOT send provi=
sion. I&#39;d rather my implementation be strict and enforce the MUST NOT s=
end via PROTOCOL ERROR when receiving this. The kindler gentler days of bei=
ng liberal in what we receive have sailed. There is not yet an argument for=
 backwards compatibility to not do that, but if I did so I would also be in=
 non compliance.<br>
<br></div>Can we either change it to PROTOCOL ERROR (preferred) or just be =
silent on handling of the error?<br><br></div>-Patrick<br></div>

--047d7b3a928668f7ac04e1921617--