Re: Comments on Byte range draft
Simon Spero (ses@tipper.oit.unc.edu)
Mon, 13 Nov 1995 17:24:27 -0800 (PST)
On Mon, 13 Nov 1995, Ari Luotonen wrote:
>
> > Unfortunately, this isn't true. The rule is: if you see a response code
> > that you don't understand, you must treat it as if the code was <N>00,
> > where <N> is the first character of the response code. Thus, a conformant
> > HTTP/1.0 system should treat a code '205' as code '200'.
>
> Nevertheless, the current practice in proxies is to be safe and not
> cache anything that isn't 200. So it will be backwords compatible
> with existing implementations.
In that case, the HTTP/1.0 spec is incorrect, and must be changed, either
to put in a special case for 2XX series results; the general case is
still needed as it has been used in several places.
Simon