[go: up one dir, main page]

Skip to content
This repository has been archived by the owner on Dec 1, 2020. It is now read-only.
/ typebeat Public archive

Parsing of the Content-Type header in pure OCaml

License

Notifications You must be signed in to change notification settings

mirage/typebeat

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

72 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

TypeBeat - Agnostic parser of the Content-Type in OCaml

Build Status

TypeBeat is a pure implementation of the parsing of the Content-Type's value (see RFC822 and RFC2045). The reason of this light library is to compute a complex rule. Indeed, it's hard to parse the value of the Content-Type, believe me.

So it's a common library if you want to know the value of the Content-Type and don't worry, we respect the standard. We saved the IANA database too.

Instalation

TypeBeat can be installed with opam:

opam install type-beat

Explanation

TypeBeat uses the cool and funny Angstrom library to parse the value of the Content-Type. If you want to implement an email parser (like MrMime) or an HTTP server (CoHTTP), firstly, these already exist, too bad.

This parser handles complex rules like the CFWS token and other weird rules from old and stupid RFCs. The point is to centralize all these parsers in one library (because you can find the Content-Type crazy rule in some different protocols) .

Then, the API was designed to be easy to use:

val of_string : string -> (content, error) result
val of_string_raw : string -> int -> int -> (content * int, error) result

The first transforms its string argument into a Content-Type value. The second is generally used by another parser (like an HTTP protocol parser) to parse a part of the string and return how many bytes the parser consumed.

If you are a warrior of the Angstrom library, you can use the parser:

val parser : content Angstrom.t

But the parser does not terminate because we have the CFWS token at the end. What does that mean? The parser expects an End of input or any character other than wsp (and you can produce that by Angstrom.Unbuffered.Complete) to check that the hypothetical next line is a new field. Because, as you know, we can write something like:

Content-Type: text/html;^CRLF
 charset="utf-8"

And it is still valid (see RFC822)!

Another point is that this library has all of the IANA media types database (dated 2016-06-01), so we recognize the IANA media types automatically.

Build Requirements

  • OCaml >= 4.01.0
  • Angstrom
  • topkg, ocamlfind and ocamlbuild to build the project

Improvement

If you want something from the RFC822, I can provide that in this library.