request - response

[zjmarlow]

Resolves #226

Draft for reviewing eol handling changes to HTTP::Message and both subclasses.

Some things to note:

• most work draws from RFC2616 - need to check current RFCs
• 020-message.rakutest Str 1/2 testcase currently fails due to Content-Length now being auto-calculated / included
• 020-message.rakutest Str 2/2 testcase currently fails due to Content-Length now being auto-calculated / included
• 040-request.rakutest parse 6/6 testcase currently fails due to Content-Length now being auto-calculated / included
• 040-request.rakutest multipart implied by existing content-type subtest currently fails due to missing trailing CRLF
• 050-response.rakutest parse - Str 1/4 testcase currently fails due to Content-Length now being auto-calculated / included
• 170-request-common.rakutest POST - 3 currently fails due to Content-Length now being calculated when Str is called

Could content handling (generating, parsing) be separated out? It occurs in several places.