miximka / mimeparser

Mime parsing in Swift | Relevant RFCs: RFC 822, RFC 2045, RFC 2046

About

MimeParser is a simple MIME (Multipurpose Internet Mail Extensions) parsing library written in Swift (to learn more about mimes refer to RFC 822, RFC 2045, RFC 2046)

Installation

CocoaPods

CocoaPods is a dependency manager for Cocoa projects. You can install it with the following command:

$ gem install cocoapods

To integrate MimeParser into your Xcode project using CocoaPods, specify it in your Podfile:

project '<Your Project Name>.xcodeproj'
platform :osx, '10.12'

target 'Test' do
  use_frameworks!
  pod 'MimeParser', '~> 0.1'
end

Then, run the following command:

$ pod install

Usage

Import MimeParser before using it:

import MimeParser

Create parser object:

let parser = MimeParser()

Let this be a simplest mime to be parsed:

let str = """
	Content-Type: text/plain
	
	Test
	"""

You are ready to parse the mime:

let mime = try parser.parse(str)

Returned mime object is a root of the mime tree and provides access to its header fields and content:

public enum MimeContent {
    case body(MimeBody)
    case mixed([Mime])
    case alternative([Mime])
}

public struct MimeHeader {
    public let contentTransferEncoding: ContentTransferEncoding?
    public let contentType: ContentType?
    public let contentDisposition: ContentDisposition?
    public let other: [RFC822HeaderField]
}

if let contentTypeString = mime.header.contentType?.raw {
	print("\(contentTypeString)")
	// "text/plain"
}

if case .body(let body) = mime.content {
	print("\(body.raw)")
	// "Test"
}

Decoded mime's content is simply to retrieve:

let content = try mime.decodedContentData()
// "Test"

License

MimeParser is available under the MIT license. See the LICENSE file for more info.

Contribution

MimeParser is still very simple and incomplete, so pull requests welcome!

mimeparser's Issues

Fails to decode utf8 text in some cases

In the odd event that an attachment has content-type like this:

Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit

This may be an odd combination, but it is definitely out there in the wild and should be handled even if it is "non-standard"

The library tries to decode it as ".ascii" here:

MimeParser/Sources/MimeParser/MimeContentDecoder.swift

Line 31 in 0903ca7

guard let decoded = raw.data(using: .ascii) else { throw Error.decodingFailed }

Even though the content-transfer-encoding is specified as 7bit, the data is actually utf-8 and standard ascii decoding will fail.
The charset should possibly override the transfer encoding, or possibly the decoder could fall-back to trying .utf8 if the .ascii decode fails. For example:

     guard let decoded = raw.data(using: .ascii) else {
            raw.data(using: .utf8) else { throw Error.decodingFailed }
     }

Header parser fails with empty headers

Parsing the attached NNTP message with MimeParser fails to include the "Subject" header, because it is preceded by a "Cc" header with no content:

Cc: 
Subject: Emacs implementations, list of, regular post [long, FAQ]

The problem is the regular expression used to parse the headers in RFC822HeaderFieldsPartitioner.fields():

        let regex = try! NSRegularExpression(pattern: "(.+?):\\s*(.+)", options: [])

It would probably be better to parse these 1 line at a time.

EmptyCc.txt

Does not handle multipart/alternative correctly

The parser doesn't seem to be handling multipart/alternative correctly. Please see https://github.com/dkontyko/MimeParser/blob/adding_tests/Tests/MimeParserTests/EmailParsingTests.swift for an example. In the test email, the number of alternative parts is 3, but the parser only parses 2, and testCanParseEmailWithMultipartAlternative fails.

Encoded words support

First of all thank you for this library. I've used it to retrieve some headers, and I stumbled upon a header that contained UTF-8 data, is there a way to decode this header using this library ?

Format seems to be ?=encoding?charset?mytexthere?=

Thank you.

Headers are not case sensitive

Great parser! But header names are not case sensitive, so looking for "Content-Type" in HeaderParser.parse is a bit too strict.