Пример #1
0
// IsAttachment returns true, if the given header defines an attachment.
// First it checks, if the Content-Disposition header defines an attachement.
// If this test is false, the Content-Type header is checked.
//
// Valid Attachment-Headers:
//
//    Content-Disposition: attachment; filename="frog.jpg"
//    Content-Type: attachment; filename="frog.jpg"
//
func IsAttachment(header mail.Header) bool {
	mediatype, _, _ := mime.ParseMediaType(header.Get("Content-Disposition"))
	if strings.ToLower(mediatype) == "attachment" {
		return true
	}

	mediatype, _, _ = mime.ParseMediaType(header.Get("Content-Type"))
	if strings.ToLower(mediatype) == "attachment" {
		return true
	}

	return false
}
Пример #2
0
// ParseMIME reads a MIME document from the provided reader and parses it into
// tree of MIMEPart objects.
func ParseMIME(reader *bufio.Reader) (MIMEPart, error) {
	tr := textproto.NewReader(reader)
	header, err := tr.ReadMIMEHeader()
	if err != nil {
		if !strings.HasPrefix(err.Error(), "malformed MIME header") {
			return nil, err
		}
	}
	mediatype, params, err := mime.ParseMediaType(header.Get("Content-Type"))
	if err != nil && mime.IsOkPMTError(err) != nil {
		return nil, err
	}
	root := &memMIMEPart{header: header, contentType: mediatype}

	if strings.HasPrefix(mediatype, "multipart/") {
		boundary := params["boundary"]
		err = parseParts(root, reader, boundary)
		if err != nil {
			return nil, err
		}
	} else {
		// Content is text or data, decode it
		content, err := decodeSection(header.Get("Content-Transfer-Encoding"), reader)
		if err != nil {
			return nil, err
		}
		root.content = content
	}

	return root, nil
}
Пример #3
0
func (p *Part) parseContentDisposition() {
	v := p.Header.Get("Content-Disposition")
	var err error
	p.disposition, p.dispositionParams, err = mime.ParseMediaType(v)
	if err != nil && mime.IsOkPMTError(err) != nil {
		p.dispositionParams = emptyParams
	}
}
Пример #4
0
// Returns a MIME message with only one Attachment, the parsed original mail body.
func binMIME(mailMsg *mail.Message) (*MIMEBody, error) {
	// Root Node of our tree
	ctype := mailMsg.Header.Get("Content-Type")
	mediatype, mparams, err := mime.ParseMediaType(ctype)
	if err != nil {
		mediatype = "attachment"
	}

	m := &MIMEBody{
		header:         mailMsg.Header,
		Root:           NewMIMEPart(nil, mediatype),
		IsTextFromHTML: false,
	}

	p := NewMIMEPart(nil, mediatype)
	p.content, err = decodeSection(mailMsg.Header.Get("Content-Transfer-Encoding"), mailMsg.Body)
	if err != nil {
		return nil, err
	}

	// get set headers
	p.header = make(textproto.MIMEHeader, 4)
	// Figure out our disposition, filename
	disposition, dparams, err := mime.ParseMediaType(mailMsg.Header.Get("Content-Disposition"))
	if err == nil {
		// Disposition is optional
		p.disposition = disposition
		p.fileName = DecodeHeader(dparams["filename"])
	}
	if p.fileName == "" && mparams["name"] != "" {
		p.fileName = DecodeHeader(mparams["name"])
	}
	if p.fileName == "" && mparams["file"] != "" {
		p.fileName = DecodeHeader(mparams["file"])
	}
	if p.charset == "" {
		p.charset = mparams["charset"]
	}

	p.header.Set("Content-Type", mailMsg.Header.Get("Content-Type"))
	p.header.Set("Content-Disposition", mailMsg.Header.Get("Content-Disposition"))

	m.Attachments = append(m.Attachments, p)
	return m, err
}
Пример #5
0
// IsPlain returns true, if the the mime headers define a valid
// 'text/plain' or 'text/html part'. Ff emptyContentTypeIsPlain is set
// to true, a missing Content-Type header will result in a positive
// plain part detection.
func IsPlain(header mail.Header, emptyContentTypeIsPlain bool) bool {
	ctype := header.Get("Content-Type")
	if ctype == "" && emptyContentTypeIsPlain {
		return true
	}

	mediatype, _, err := mime.ParseMediaType(ctype)
	if err != nil {
		return false
	}
	switch mediatype {
	case "text/plain",
		"text/html":
		return true
	}

	return false

}
Пример #6
0
// IsMultipartMessage returns true if the message has a recognized multipart Content-Type
// header.  You don't need to check this before calling ParseMIMEBody, it can handle
// non-multipart messages.
func IsMultipartMessage(mailMsg *mail.Message) bool {
	// Parse top-level multipart
	ctype := mailMsg.Header.Get("Content-Type")
	mediatype, _, err := mime.ParseMediaType(ctype)
	if err != nil && mime.IsOkPMTError(err) != nil {
		return false
	}
	switch mediatype {
	case "multipart/alternative",
		"multipart/mixed",
		"multipart/related",
		"multipart/signed":
		return true
	default:
		if strings.HasPrefix(mediatype, "multipart/") {
			// according to rfc2046#section-5.1.7 all other multipart should
			// be treated as multipart/mixed
			return true
		}
	}
	return false
}
Пример #7
0
// parseParts recursively parses a mime multipart document.
func parseParts(parent *memMIMEPart, reader io.Reader, boundary string) error {
	var prevSibling *memMIMEPart

	// Loop over MIME parts
	mr := multipart.NewReader(reader, boundary)
	for {
		// mrp is golang's built in mime-part
		mrp, err := mr.NextPart()
		if err != nil {
			if err == io.EOF {
				// This is a clean end-of-message signal
				break
			} else if strings.HasPrefix(err.Error(), "malformed MIME header") {
				// ignore this type of error and continue to process and valid MIME header
				//log.Println("debug: malformed MIME header - ignore it", len(mrp.Header))
			} else if strings.HasSuffix(err.Error(), "EOF") {
				//log.Println("debug: type of EOF failure:", err)
				if mrp == nil {
					//log.Println("debug: next part is empty")
					break
				}
			} else {
				return err
			}
		}
		if len(mrp.Header) == 0 {
			// // Empty header probably means the part didn't using the correct trailing "--"
			// // syntax to close its boundary.  We will let this slide if this this the
			// // last MIME part.
			// if _, err := mr.NextPart(); err != nil {
			// 	if err == io.EOF || strings.HasSuffix(err.Error(), "EOF") {
			// 		// This is what we were hoping for
			// 		break
			// 	} else {
			// 		return fmt.Errorf("Error at boundary %v: %v", boundary, err)
			// 	}
			// }
			// return fmt.Errorf("Empty header at boundary %v", boundary)

			if errEOF := mr.CheckNextPart(); errEOF != nil {
				if errEOF == io.EOF || strings.HasSuffix(errEOF.Error(), "EOF") {
					// This is what we were hoping for. And to remain the ability
					// to detect the empty MIME header caused by improper boundary
					// ending.
					break
				}
			}
			// empty header field inside mime part body should not treat as error as
			// MIME is allowed to have empty header and straight to the body content.
			mrp.Header.Add("Content-Type", default_content_type)
		}
		ctype := mrp.Header.Get("Content-Type")
		if ctype == "" {
			//return fmt.Errorf("Missing Content-Type at boundary %v", boundary)

			// can not find Content-Type header does not mean error
			//log.Println("debug: can not found content-type - use default")
			mrp.Header.Add("Content-Type", default_content_type)
			ctype = mrp.Header.Get("Content-Type")
		}
		mediatype, mparams, err := mime.ParseMediaType(ctype)
		if err != nil && mime.IsOkPMTError(err) != nil {
			//log.Println("debug: parse parts media type error")
			return err
		}

		// Insert ourselves into tree, p is enmime's mime-part
		p := NewMIMEPart(parent, mediatype)
		p.header = mrp.Header
		if prevSibling != nil {
			prevSibling.nextSibling = p
		} else {
			parent.firstChild = p
		}
		prevSibling = p

		// Figure out our disposition, filename
		disposition, dparams, err := mime.ParseMediaType(mrp.Header.Get("Content-Disposition"))
		if err == nil || mime.IsOkPMTError(err) == nil {
			// Disposition is optional
			p.disposition = disposition
			p.fileName = DecodeHeader(dparams["filename"])
		}
		if p.fileName == "" && mparams["name"] != "" {
			p.fileName = DecodeHeader(mparams["name"])
		}
		if p.fileName == "" && mparams["file"] != "" {
			p.fileName = DecodeHeader(mparams["file"])
		}
		if p.charset == "" {
			p.charset = mparams["charset"]
		}

		boundary := mparams["boundary"]
		if boundary != "" && !strings.HasPrefix(mediatype, "text/") {
			// Content is another multipart
			err = parseParts(p, mrp, boundary)
			if err != nil {
				return err
			}
		} else {
			// Content is text or data, decode it
			d := mrp.Header.Get("Content-Transfer-Encoding")
			if mediatype == "message/rfc822" {
				switch strings.ToLower(d) {
				case "7bit", "8bit", "binary":
				default:
					d = "" // force no decoding
				}
			}
			data, err := decodeSection(d, mrp)
			if err != nil {
				return err
			}
			p.content = data
		}
	}

	return nil
}
Пример #8
0
// ParseMIMEBody parses the body of the message object into a  tree of MIMEPart objects,
// each of which is aware of its content type, filename and headers.  If the part was
// encoded in quoted-printable or base64, it is decoded before being stored in the
// MIMEPart object.
func ParseMIMEBody(mailMsg *mail.Message) (*MIMEBody, error) {
	var gerr error
	mimeMsg := &MIMEBody{
		IsTextFromHTML: false,
		header:         mailMsg.Header,
	}

	if !IsMultipartMessage(mailMsg) {
		// Attachment only?
		if IsBinaryBody(mailMsg) {
			return binMIME(mailMsg)
		}

		// Parse as text only
		bodyBytes, err := decodeSection(mailMsg.Header.Get("Content-Transfer-Encoding"),
			mailMsg.Body)
		if err != nil {
			return nil, fmt.Errorf("Error decoding text-only message: %v", err)
		}
		// Handle plain ASCII text, content-type unspecified
		mimeMsg.Text = string(bodyBytes)

		// Check for HTML at top-level, eat errors quietly
		ctype := mailMsg.Header.Get("Content-Type")
		if ctype != "" {
			if mediatype, mparams, err := mime.ParseMediaType(ctype); err == nil || mime.IsOkPMTError(err) == nil {
				/*
				 *Content-Type: text/plain;\t charset="hz-gb-2312"
				 */
				if mparams["charset"] != "" {
					// Convert plain text to UTF8 if content type specified a charset
					newStr, err := ConvertToUTF8String(mparams["charset"], bodyBytes)
					if err != nil && newStr == "" {
						return nil, err
					} else {
						if err != nil {
							gerr = err
						}
						mimeMsg.Text = newStr
					}
				} else if mediatype == "text/html" { // charset is empty, look in html body for charset
					charset, err := charsetFromHTMLString(mimeMsg.Text)

					if charset != "" && err == nil {
						newStr, err := ConvertToUTF8String(charset, bodyBytes)
						if err != nil && newStr == "" {
							return nil, err
						} else {
							if err != nil {
								gerr = err
							}
							mimeMsg.Text = newStr
						}
					}
				}
				if mediatype == "text/html" {
					mimeMsg.HTML = mimeMsg.Text
					// Empty Text will trigger html2text conversion below
					mimeMsg.Text = ""
				}
			}
		}
	} else {
		// Parse top-level multipart
		ctype := mailMsg.Header.Get("Content-Type")
		mediatype, params, err := mime.ParseMediaType(ctype)
		if err != nil && mime.IsOkPMTError(err) != nil {
			return nil, fmt.Errorf("Unable to parse media type: %v", err)
		}
		if !strings.HasPrefix(mediatype, "multipart/") {
			return nil, fmt.Errorf("Unknown mediatype: %v", mediatype)
		}
		boundary := params["boundary"]
		if boundary == "" {
			return nil, fmt.Errorf("Unable to locate boundary param in Content-Type header")
		}
		// Root Node of our tree
		root := NewMIMEPart(nil, mediatype)
		mimeMsg.Root = root
		err = parseParts(root, mailMsg.Body, boundary)
		if err != nil {
			return nil, err
		}

		// Locate text body
		if mediatype == "multipart/altern" {
			log.Println("surPrise: should not be here go.enmime/mime.go")
			match := BreadthMatchFirst(root, func(p MIMEPart) bool {
				return p.ContentType() == "text/plain" && p.Disposition() != "attachment"
			})
			if match != nil {
				if match.Charset() != "" {
					newStr, err := ConvertToUTF8String(match.Charset(), match.Content())
					if err != nil {
						if newStr == "" {
							return nil, err
						} else {
							gerr = err
						}
					}
					mimeMsg.Text += newStr
				} else {
					mimeMsg.Text += string(match.Content())
				}
			}
		} else {
			// multipart is of a mixed type
			match := DepthMatchAll(root, func(p MIMEPart) bool {
				return p.ContentType() == "text/plain" && p.Disposition() != "attachment"
			})
			for i, m := range match {
				if i > 0 {
					mimeMsg.Text += "\n--\n"
				}
				if m.Charset() != "" {
					newStr, err := ConvertToUTF8String(m.Charset(), m.Content())
					if err != nil {
						if newStr == "" {
							return nil, err
						} else {
							gerr = err
						}
					}
					mimeMsg.Text += newStr
				} else {
					mimeMsg.Text += string(m.Content())
				}
			}
		}

		// Locate HTML body
		match := BreadthMatchFirst(root, func(p MIMEPart) bool {
			return p.ContentType() == "text/html" && p.Disposition() != "attachment"
		})
		if match != nil {
			if match.Charset() != "" {
				newStr, err := ConvertToUTF8String(match.Charset(), match.Content())
				if err != nil {
					if newStr == "" {
						return nil, err
					} else {
						gerr = err
					}
				}
				mimeMsg.HTML += newStr
			} else {
				mimeMsg.HTML = string(match.Content())
			}

		}

		// Locate attachments
		mimeMsg.Attachments = BreadthMatchAll(root, func(p MIMEPart) bool {
			return p.Disposition() == "attachment" || p.ContentType() == "application/octet-stream"
		})

		// Locate inlines
		mimeMsg.Inlines = BreadthMatchAll(root, func(p MIMEPart) bool {
			return p.Disposition() == "inline"
		})

		// Locate others parts not handled in "Attachments" and "inlines"
		mimeMsg.OtherParts = BreadthMatchAll(root, func(p MIMEPart) bool {
			if strings.HasPrefix(p.ContentType(), "multipart/") {
				return false
			}

			if p.Disposition() != "" {
				return false
			}

			if p.ContentType() == "application/octet-stream" {
				return false
			}

			return p.ContentType() != "text/plain" && p.ContentType() != "text/html"
		})
	}

	// Down-convert HTML to text if necessary
	if mimeMsg.Text == "" && mimeMsg.HTML != "" {
		mimeMsg.IsTextFromHTML = true
		var err error
		if mimeMsg.Text, err = html2text.FromString(mimeMsg.HTML); err != nil {
			// Fail gently
			mimeMsg.Text = ""
			return mimeMsg, err
		}
	}
	return mimeMsg, gerr
}