14.3 Accept-Encoding


The Accept-Encoding request-header field is similar to Accept, but restricts the content-codings (section 3.5) that are acceptable in the response.
       Accept-Encoding  = "Accept-Encoding" ":"
                          1#( codings [ ";" "q" "=" qvalue ] )
       codings          = ( content-coding | "*" )


Examples of its use are:
       Accept-Encoding: compress, gzip
       Accept-Encoding:
       Accept-Encoding: *
       Accept-Encoding: compress;q=0.5, gzip;q=1.0
       Accept-Encoding: gzip;q=1.0, identity; q=0.5, *;q=0


A server tests whether a content-coding is acceptable, according to an Accept-Encoding field, using these rules:
      1. If the content-coding is one of the content-codings listed in
         the Accept-Encoding field, then it is acceptable, unless it is
         accompanied by a qvalue of 0. (As defined in section 3.9, a
         qvalue of 0 means "not acceptable.")
      2. The special "*" symbol in an Accept-Encoding field matches any
         available content-coding not explicitly listed in the header
         field.
      3. If multiple content-codings are acceptable, then the acceptable
         content-coding with the highest non-zero qvalue is preferred.
      4. The "identity" content-coding is always acceptable, unless
         specifically refused because the Accept-Encoding field includes
         "identity;q=0", or because the field includes "*;q=0" and does
         not explicitly include the "identity" content-coding. If the
         Accept-Encoding field-value is empty, then only the "identity"
         encoding is acceptable.


If an Accept-Encoding field is present in a request, and if the server cannot send a response which is acceptable according to the Accept-Encoding header, then the server SHOULD send an error response with the 406 (Not Acceptable) status code.

If no Accept-Encoding field is present in a request, the server MAY assume that the client will accept any content coding. In this case, if "identity" is one of the available content-codings, then the server SHOULD use the "identity" content-coding, unless it has additional information that a different content-coding is meaningful to the client.
      Note: If the request does not include an Accept-Encoding field,
      and if the "identity" content-coding is unavailable, then
      content-codings commonly understood by HTTP/1.0 clients (i.e.,
      "gzip" and "compress") are preferred; some older clients
      improperly display messages sent with other content-codings.  The
      server might also make this decision based on information about
      the particular user-agent or client.
      Note: Most HTTP/1.0 applications do not recognize or obey qvalues
      associated with content-codings. This means that qvalues will not
      work and are not permitted with x-gzip or x-compress.