Parsing of Content-Type does not strictly follow the RFC

[mseri]

For reference see #725 and the discussion in #542

[dinosaure]

If someone wants to work on this part of CoHTTP, he should take a look on typebeat (which will be archived!). The work of typebeat comes from what we did about mrmime (email and HTTP share the same RFC about Content-Type - see RFC 2045 § 5).

Note that, even if the format seems quite close, Accept{-*} field has another format described in RFC 7231 § 5.3 which is not handle by typebeat.

A final note is about FWS and CFWS token which can be the part of values (according RFC 7230 § 3.2.4), Even if it's considered as an obsolete form of values, you should handle such layout. You should look into unstrctrd which folds these kind of whitespace - mrmime uses it to pre-process the value and delete all FWS/CFWS.