On Sat 15 Sep 2001, Andrew M. Bishop wrote:
> > :
> > : I asked for the "gzip" encoding, but WWWOFFLE replied with an
> > : "x-gzip" encoding. Those may be the same format but you cannot
> > : expect the client to know that.
>
> According to RFC2616 (HTTP/1.1) you SHOULD expect the client to know
> that.
>
> -------------------- RFC 2616 --------------------
> Use of program names for the identification of encoding formats
> is not desirable and is discouraged for future encodings. Their
> use here is representative of historical practice, not good
> design. For compatibility with previous implementations of HTTP,
> applications SHOULD consider "x-gzip" and "x-compress" to be
> equivalent to "gzip" and "compress" respectively.
> -------------------- RFC 2616 --------------------
>
> I suggest that it is a bug with the client and not with the proxy. A
> bug report should be raised with the browser producer.
>
> > : Indeed, MS Internet Explorer
> > : version 5.00.2614.3500 asks for "Accept-Encoding: gzip, deflate"
> > : and cannot parse the body with "Content-Encoding: x-gzip".
>
> Not much hope of Microsoft *ever* following RFCs or standards is
> there.
Unfortunately not, no :-(
> > : Please change WWWOFFLE so that it remembers whether the client
> > : asked for "gzip" or "x-gzip", and uses the same name in the
> > : response.
>
> WWWOFFLE is compliant with the RFC, the browser is not; where is the
> problem?
>
> I should take a stand on this and refuse to make the change.
Of course, the interpretation of the RFC's use of "application" could be
extended to proxies. Wwwoffle (correctly) considers gzip and x-gzip
equivalent, but it might be argued that it should return the same type
to the client that the client asked for. Indeed, although wwwoffle now
complies with the RFC, by "remembering" the type asked for and returning
that (i.e. noting whether "x-" was used or not) will not change that
compliance with the RFC, yet will work with more (broken) clients.
> But if I did that I would just lose more WWWOFFLE users and add
> another reason that they don't try and use the Win32 version.
That's true. And you can add in the docs that wwwoffle now caters to the
broken behaviour of Microsoft's I.E. in this respect :-)
Paul Slootman
