[11598] in www-talk@info.cern.ch
Content-Encoding confusion
daemon@ATHENA.MIT.EDU (Hugo Haas)
Tue May 2 16:11:23 2000
Date: Wed, 26 Apr 2000 19:50:53 -0400
From: Hugo Haas <hugo@w3.org>
To: www-talk@w3.org
Message-Id: <20000502155658.B4719@w3.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Resent-From: www-talk@w3.org
There are a lot of problems with the gzip'ed files downloaded.
If I request:
http://www.w3.org/TR/html4/html40.pdf
the server is replies:
Content-Encoding: gzip
Content-Length: 962329
Content-Location: html40.pdf.gz
Content-Type: application/pdf; qs=0.001
Now, when I ask for:
http://www.w3.org/TR/html4/html40.pdf.gz
the server is going to reply:
Content-Encoding: gzip
Content-Length: 962329
Content-Type: application/pdf; qs=0.001
This is the configuration on www.w3.org, and I believe that it is
Apache's default behavior.
When I request html40.pdf, it is obvious that my browser needs to decode
(i.e. uncompress) the file on the fly, and save it under html40.pdf.
What about html.pdf.gz? The HTTP headers are the same, so I guess that
it's why browsers usually uncompress the file. And since they asked for
html40.pdf.gz, they save it under this name (there's no evidence telling
them to get rid of the .gz extension). Under Windows, you get a PDF file
with a .gz extension, which PDF readers don't like.
I think that the reply to http://www.w3.org/TR/html4/html40.pdf.gz
should not specify any content encoding and the content type should be
set to application/x-gzip or something similar, and that application/pdf
should only be used when the resource is negotiated.
RFC2616 says (section 7.2.1):
Content-Type specifies the media type of the underlying data.
Content-Encoding may be used to indicate any additional content
codings applied to the data, usually for the purpose of data
compression, that are a property of the requested resource. There is
no default encoding.
My understanding of that is it is encoding of the transfer, not of the
data.
What is the correct behavior (of both clients and servers)?
--
Hugo Haas, Webmaster, Systems Team - W3C/MIT
mailto:hugo@w3.org - tel:+1-617-452-2092
