Re: What is WARNING: suspicious CR characters in HTTP header ?

From: Anand Kumria <wildfire@dont-contact.us>
Date: Wed, 16 Feb 2005 03:27:56 +1100

On Tue, 15 Feb 2005 01:45:46 +0300, Evgeny Kotsuba wrote:

> On Mon, 14 Feb 2005 23:35:02 +0100 (CET)
> Henrik Nordstrom <hno@squid-cache.org> wrote:
>> On Tue, 15 Feb 2005, Evgeny Kotsuba wrote:
>>
>>> Well, why it is parsed properly without return
>>
>> It didn't. Earlier versions just did not notice it could not parse
>>this malformed header properly..
>>
>>> and why Mozilla/IE with
>>> direct connection can parse it properly ?
>>
>> Question is if they can.
>>
>> Yes, they probably show the expected content, but did they parse the
>>header proper?
>
> Users now are standing up on these rakes and say: "Why today I can't do
> thing that I can do yesterday ?" They don't want to know - did they can
> parse somethig or not , they know that it woked yesterday.

The server is broken. Basically servers are supposed to be:

        'liberal in what they accept and conservative in what they send'

However where to draw the line is the hard part. This was returned as a
warning -- not as a hard error.

What if the next server you encounter send multiple response(s) per
request (I've seen these in the wild) like:

HTTP/1.1 200 OK
<some headers>
HTTP/1.0 200 OK
<some more headers>

What should any server do in this situation? Or in yours?

Anand
Received on Tue Feb 15 2005 - 09:26:48 MST

This archive was generated by hypermail pre-2.1.9 : Fri Feb 25 2005 - 12:00:03 MST