Proposed Accept-Encoding header filter change.

[JJoe]

(Dec. 29, 2010 03:36 PM)whenever Wrote:  1. Empty Accept-Encoding is permitted. See http://tools.ietf.org/html/rfc2616#section-14.3

The empty header could be sent when the User-Agent wasn't sending an empty header. Doesn't seem like the thing to do?

(Dec. 29, 2010 03:36 PM)whenever Wrote:  2. The filter changes

Code:

Accept-Encoding: pack200-gzip, compress

to

Code:

Accept-Encoding: gzip

Which I think is not expected.

I forgot about that one. I'll plug the hole, thanks.

(Dec. 29, 2010 03:36 PM)whenever Wrote:  3. Why not remove x-gzip?

Besides the issue that Graycode has pointed out, x-gzip seems to be outdated. See http://en.wikipedia.org/wiki/HTTP_compression & http://tools.ietf.org/html/rfc2616#section-3.5

Lack of knowledge. I thought about that. Also considered changing x-gzip to gzip when gzip is not present.
However, Opera still sends x-gzip and I found an app that added support years after its creators considered x-gzip to be deprecated. Why?

Am also considering forcing gzip only when deflate and gzip are present and or throwing a flag for deflate compressed files.

The new filter allows the user to do as they wish.
Question is what should the default be?