Re: line folding - ABNF vs prose

 On Thu, 28 Jul 2011 00:02:45 +0200, Willy Tarreau wrote:
> On Wed, Jul 27, 2011 at 08:20:58AM -0700, Roy T. Fielding wrote:
>> > But I thought we were only allowed to fold lines for headers which 
>> are
>> > defined as comma-separated lists, and to add a comma. So by 
>> definition
>> > we should not do that on a header that is not a list, and since a 
>> header
>> > which is not a list must not appear multiple times, this should 
>> not be
>> > a problem.
>>
>> No, (un)folding long field values has nothing to do with combining 
>> multiple
>> header fields of the same name into a single value.
>
> Sorry Roy, I said totally stupid things here, I was mixing two 
> things,
> multiple header procesing and multiple lines. No of course I'm not 
> adding
> the comma here. I'm just replacing all LWS with SP, so basically we 
> get at
> least 2 SP in the result (in case of pure LF+SP) or 3 SP for the 
> common
> case (CRLF+SP).
>
> Sorry for the noise.
>
> Willy

 Thank you, that discussion explains this old bug which we have seen on 
 some networks:

 HTTP/1.1 200 Okay
 Connection: Keep-Alive
 Content-Type: multipart/byteranges; boundary=":gn1, Tl__lXvUYr73ia-"
 Date: Wed, 31 Jul 2010 16:12:17 GMT

 ---:gn1Tl__lXvUYr73ia-
 blah


 The agents involved there have now been fixed not to wrap at 50 octets. 
 But still, boundary may be very long and there are other cases like this 
 out there. Particularly amongst long quoted-string custom headers.
   http://bugs.squid-cache.org/show_bug.cgi?id=3061

 AYJ

Received on Wednesday, 27 July 2011 23:36:25 UTC