http: Decode transfer encoding first

[jbrobst]

The unencoding stack is added to as Transfer-Encoding and Content-Encoding fields are encountered with no distinction between the two, meaning the stack will be incorrect if, e.g., the message has both fields and a non-chunked Transfer-Encoding comes first. This PR fixes this by ordering the stack with transfer encodings first.

Additionally, something I did not change but I think may be a bug is that MAX_ENCODE_STACK doesn't limit writer_stack's length, just the number of encodings in a single field line.

[bagder]

@monnerat: any comments on this?

[monnerat]

Yes, this makes sense. Without it we depend on the ordering of headers within the received stream.
The special case of chunked encoding still remains though as it is handled externally (always first). I guess it should be taken as a known limitation, unless we move it to the regular unencoding stack, which is a big work.

[bagder]

Thanks!