@ronomon/mime

mime

Fast, robust, standards-compliant MIME decoder. Ships with extensive tests and fuzz tests.

Installation

npm install @ronomon/mime

Fast

• Decodes on demand only as much as necessary to access a particular property. For example, if you need mime.subject, then MIME will search for the CRLF pair marking the end of the headers and decode only the subject header, without decoding any other headers and without decoding the body. This works well with the first few layers of spam defenses, which often only need to decode particular headers to reject an email.

• Caches decoded properties for subsequent use.

• Uses native fuzz-tested C++ Base64 and Quoted-Printable bindings. MIME's Base64 decoder in particular was developed for decoding wrapped Base64 more efficiently and detecting obvious corruption and character truncation.

• Uses custom lookup tables to minimize the cost of branching through too many conditionals when decoding.

• Avoids unnecessary string and buffer allocations. Algorithms accept and work with buffers directly, and allocate and copy buffers only when necessary.

• Avoids regular expression decoders.

Robust

• Provides detailed error messages which refer to the relevant RFCs to assist debugging, and which can be used directly as part of an SMTP reply.

• Accepts CRLF and LF line-endings (which are common) but not CR line-endings (which are rare).

• Accepts illegal transport padding frequently added by intermediaries (e.g. within the angle brackets of a msg-id or angle-addr, and between tokens in an encoded-word).

• Decodes a variety of malformed but common mailbox syntaxes (e.g. no angle brackets around the addr-spec, with a display-name present on the left or right).

• Removes balanced single quotes around the display-name or addr-spec in an email address (sometimes added by Outlook).

• Decodes encoded-words not separated by WSP.

• Decodes encoded-words with empty encoded-text.

• Decodes encoded-words found in Content-Type and Content-Disposition parameters (used by Outlook and Gmail contrary to RFC 2047 5 Use of encoded-words in message headers).

• Removes any directory path components from an attachment name or filename (when accessed via mime.filename, see Usage below).

• Decodes msg-ids not separated by whitespace or commas (i.e. separated only by angle brackets).

• Rejects unrecognized Content-Transfer-Encoding mechanisms contrary to RFC 2045 6.4 (e.g. anything other than 7bit, 8bit, binary, base64, or quoted-printable). This is to avoid accepting responsibility for content which will not display correctly, if at all. In contrast, the spec advocates silently altering the Content-Type.

• Rejects malicious RFC 2231 continuation indices designed to cause overallocation.

• Rejects Base64 data containing illegal characters (anything which is not a valid Base64 or whitespace character, e.g. null bytes which could cause security issues).

• Rejects Base64 data which is clearly truncated (as opposed to just missing padding).

• Corrects Quoted-Printable data containing illegal characters (anything which is not a valid Quoted-Printable character, e.g. null bytes which could cause security issues).

• Rejects illegal character sequences according to the specified charset.

• Rejects truncated character sequences according to the specified charset.

• Normalizes and aliases a variety of character sets to the canonical character set, (e.g. ks_c_5601-1987 is sometimes used by Outlook and is aliased to CP949 - Korean, otherwise the characters would decode from the wrong character set and be unintelligible).

• Rejects unknown character sets not supported by iconv.

• Decodes text/* body parts to UTF-8 buffers if the Content-Type indicates that the body is encoded in any other character set.

• Rejects unterminated comments and quoted-strings.

• Rejects invalid Content-Type syntax.

• Detects missing multipart parts (e.g. no terminating boundary delimiter).

• Rejects dangerous message/external-body and message/partial media types.

• Decodes a variety of time zones and year formats.

• Accepts missing time zone and assumes UTC to support email clients such as Blackberry which do not provide the required time zone in the Date header.

• Rejects invalid Date header syntax.

• Rejects missing From header.

• Rejects headers containing forbidden characters.

• Rejects folded header lines which exceed the 998 line length limit, but only after allowing for clients such as Outlook.com which exclude the field-name and colon from their character count, and which mistake the limit to be 1000 characters excluding the CRLF. The limit is in fact 998 characters excluding the CRLF.

• Rejects multipart boundaries containing forbidden characters.

• Rejects malicious data designed to cause CPU-intensive decoding and stack overflows.

• Rejects malicious multiple occurrences of crucial headers and parameters, which could cause clients to render an email differently from that scanned by anti-virus software.

Standards-Compliant

• RFC 5322 - Internet Message Format.

• RFC 5321 - Simple Mail Transfer Protocol.

• RFC 2045 - Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies.

• RFC 2046 - Multipurpose Internet Mail Extensions (MIME) Part Two: Media Types.

• RFC 2047 - MIME (Multipurpose Internet Mail Extensions) Part Three: Message Header Extensions for Non-ASCII Text.

• RFC 2183 - The Content-Disposition Header Field.

• RFC 2231 - MIME Parameter Value and Encoded Word Extensions: Character Sets, Languages, and Continuations.

Usage

var MIME = require('@ronomon/mime');

// Instantiate a new mime instance (no decoding will take place):

var mime = new MIME.Message(buffer);

// Decoding will take place when the following getter properties are accessed.
// These getter properties may throw an exception for malformed MIME data.

mime.headers;             // { 'received': [<Buffer>] }
mime.body;                // <Buffer>

mime.from;                // [ { name: <String>, email: <String> } ]
mime.sender;              //   { name: <String>, email: <String> } / undefined
mime.replyTo;             // [ { name: <String>, email: <String> } ]
mime.to;                  // [ { name: <String>, email: <String> } ]
mime.cc;                  // [ { name: <String>, email: <String> } ]
mime.bcc;                 // [ { name: <String>, email: <String> } ]

mime.messageID;           //   <String> / undefined
mime.references;          // [ <String>, <String> ]
mime.inReplyTo;           // [ <String>, <String> ]

mime.date;                // <Unix Timestamp Integer> / undefined
mime.subject;             // <String>

mime.contentDisposition;  // { value: <String>, parameters: {} }
mime.contentType;         // { value: <String>, parameters: {} }
mime.contentID;           // <String> / undefined
mime.filename;            // <String> / undefined

mime.parts;               // [ <MIME.Message>, <MIME.Message> ]

Tests

To run all included tests and fuzz tests:

node test.js