-;;; rfc2047.el --- Functions for encoding and decoding rfc2047 messages
-;; Copyright (C) 1998,99 Free Software Foundation, Inc.
+;;; rfc2047.el --- functions for encoding and decoding rfc2047 messages
+;; Copyright (C) 1998, 1999, 2000, 2002, 2003 Free Software Foundation, Inc.
;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
;; MORIOKA Tomohiko <morioka@jaist.ac.jp>
;;; Commentary:
+;; RFC 2047 is "MIME (Multipurpose Internet Mail Extensions) Part
+;; Three: Message Header Extensions for Non-ASCII Text".
+
;;; Code:
-(eval-and-compile
- (eval
- '(unless (fboundp 'base64-decode-string)
- (require 'base64))))
+(eval-when-compile
+ (require 'cl)
+ (defvar message-posting-charset))
(require 'qp)
(require 'mm-util)
-(require 'ietf-drums)
+;; Fixme: Avoid this (used for mail-parse-charset) mm dependence on gnus.
(require 'mail-prsvr)
+(require 'base64)
+(autoload 'mm-body-7-or-8 "mm-bodies")
(defvar rfc2047-header-encoding-alist
'(("Newsgroups" . nil)
+ ("Followup-To" . nil)
("Message-ID" . nil)
+ ("\\(Resent-\\)?\\(From\\|Cc\\|To\\|Bcc\\|Reply-To\\|Sender\
+\\|Mail-Followup-To\\|Mail-Copies-To\\|Approved\\)" . address-mime)
(t . mime))
"*Header/encoding method alist.
The list is traversed sequentially. The keys can either be
-header regexps or `t'.
+header regexps or t.
The values can be:
1) nil, in which case no encoding is done;
2) `mime', in which case the header will be encoded according to RFC2047;
-3) a charset, in which case it will be encoded as that charse;
-4) `default', in which case the field will be encoded as the rest
+3) `address-mime', like `mime', but takes account of the rules for address
+ fields (where quoted strings and comments must be treated separately);
+4) a charset, in which case it will be encoded as that charset;
+5) `default', in which case the field will be encoded as the rest
of the article.")
(defvar rfc2047-charset-encoding-alist
(iso-8859-4 . Q)
(iso-8859-5 . B)
(koi8-r . B)
- (iso-8859-7 . Q)
- (iso-8859-8 . Q)
+ (iso-8859-7 . B)
+ (iso-8859-8 . B)
(iso-8859-9 . Q)
+ (iso-8859-14 . Q)
+ (iso-8859-15 . Q)
(iso-2022-jp . B)
(iso-2022-kr . B)
(gb2312 . B)
+ (big5 . B)
+ (cn-big5 . B)
(cn-gb . B)
(cn-gb-2312 . B)
(euc-kr . B)
(iso-2022-jp-2 . B)
- (iso-2022-int-1 . B))
+ (iso-2022-int-1 . B)
+ (viscii . Q))
"Alist of MIME charsets to RFC2047 encodings.
-Valid encodings are nil, `Q' and `B'.")
+Valid encodings are nil, `Q' and `B'. These indicate binary (no) encoding,
+quoted-printable and base64 respectively.")
(defvar rfc2047-encoding-function-alist
'((Q . rfc2047-q-encode-region)
(nil . ignore))
"Alist of RFC2047 encodings to encoding functions.")
-(defvar rfc2047-q-encoding-alist
- '(("\\(From\\|Cc\\|To\\|Bcc\||Reply-To\\):" . "-A-Za-z0-9!*+/=_")
- ("." . "^\000-\007\013\015-\037\200-\377=_?"))
- "Alist of header regexps and valid Q characters.")
-
;;;
;;; Functions for encoding RFC2047 messages
;;;
(progn
(forward-line 1)
(if (re-search-forward "^[^ \n\t]" nil t)
- (progn
- (beginning-of-line)
- (point))
+ (point-at-bol)
(point-max))))
(goto-char (point-min)))
+(defun rfc2047-field-value ()
+ "Return the value of the field at point."
+ (save-excursion
+ (save-restriction
+ (rfc2047-narrow-to-field)
+ (re-search-forward ":[ \t\n]*" nil t)
+ (buffer-substring (point) (point-max)))))
+
+(defvar rfc2047-encoding-type 'address-mime
+ "The type of encoding done by `rfc2047-encode-region'.
+This should be dynamically bound around calls to
+`rfc2047-encode-region' to either `mime' or `address-mime'. See
+`rfc2047-header-encoding-alist', for definitions.")
+
(defun rfc2047-encode-message-header ()
"Encode the message header according to `rfc2047-header-encoding-alist'.
Should be called narrowed to the head of the message."
(interactive "*")
- (when (featurep 'mule)
- (save-excursion
- (goto-char (point-min))
- (let ((alist rfc2047-header-encoding-alist)
- elem method)
- (while (not (eobp))
- (save-restriction
- (rfc2047-narrow-to-field)
- (when (rfc2047-encodable-p t)
- ;; We found something that may perhaps be encoded.
- (while (setq elem (pop alist))
- (when (or (and (stringp (car elem))
- (looking-at (car elem)))
- (eq (car elem) t))
- (setq alist nil
- method (cdr elem))))
- (when method
- (cond
- ((eq method 'mime)
- (rfc2047-encode-region (point-min) (point-max))
- (rfc2047-fold-region (point-min) (point-max)))
- ;; Hm.
- (t))))
- (goto-char (point-max)))))
- (when mail-parse-charset
- (encode-coding-region (point-min) (point-max)
- mail-parse-charset)))))
-
-(defun rfc2047-encodable-p (&optional header)
- "Say whether the current (narrowed) buffer contains characters that need encoding.
-
-A non-nil value should be passed for the optional parameter HEADER if
-the region is part of a message header. In this case, the only
-permitted unencoded charset is us-ascii."
- (let ((charsets
- (mapcar
- 'mm-mime-charset
- (mm-find-charset-region (point-min) (point-max))))
- (cs (if header
- '(us-ascii)
- (list 'us-ascii mail-parse-charset)))
- found)
- (while charsets
- (unless (memq (pop charsets) cs)
- (setq found t)))
- found))
-
-(defun rfc2047-dissect-region (b e)
- "Dissect the region between B and E into words."
- (let ((all-specials (concat ietf-drums-tspecials " \t\n\r"))
- (special-list (mapcar 'identity ietf-drums-tspecials))
- (blank-list '(? ?\t ?\n ?\r))
- words current cs state mail-parse-mule-charset)
- (save-restriction
- (narrow-to-region b e)
- (goto-char (point-min))
- (skip-chars-forward all-specials)
- (setq b (point))
+ (save-excursion
+ (goto-char (point-min))
+ (let (alist elem method)
(while (not (eobp))
- (cond
- ((not state)
- (if (memq (char-after) blank-list)
- (setq state 'blank)
- (setq state 'word)
- (if (not (eq (setq cs (mm-charset-after)) 'ascii))
- (setq current cs)))
- (setq b (point)))
- ((eq state 'blank)
- (cond
- ((memq (char-after) special-list)
- (setq state nil))
- ((memq (char-after) blank-list))
- (t
- (setq state 'word)
- (if (not (eq (setq cs (mm-charset-after)) 'ascii))
- (setq current cs)))))
- ((eq state 'word)
- (cond
- ((memq (char-after) special-list)
- (setq state nil)
- (push (list b (point) current) words)
- (setq current nil))
- ((memq (char-after) blank-list)
- (setq state 'blank)
- (push (list b (point) current) words)
- (setq current nil)
- (setq b (point)))
- ((or (eq (setq cs (mm-charset-after)) 'ascii)
- (if current
- (eq current cs)
- (setq current cs))))
- (t
- (push (list b (point) current) words)
- (setq current cs)
- (setq b (point))))))
- (if state
- (forward-char)
- (skip-chars-forward all-specials)))
- (if (eq state 'word)
- (push (list b (point) current) words)))
- words))
+ (save-restriction
+ (rfc2047-narrow-to-field)
+ (if (not (rfc2047-encodable-p))
+ (prog1
+ (if (and (eq (mm-body-7-or-8) '8bit)
+ (mm-multibyte-p)
+ (mm-coding-system-p
+ (car message-posting-charset)))
+ ;; 8 bit must be decoded.
+ (mm-encode-coding-region
+ (point-min) (point-max)
+ (mm-charset-to-coding-system
+ (car message-posting-charset))))
+ ;; No encoding necessary, but folding is nice
+ (rfc2047-fold-region
+ (save-excursion
+ (goto-char (point-min))
+ (skip-chars-forward "^:")
+ (when (looking-at ": ")
+ (forward-char 2))
+ (point))
+ (point-max)))
+ ;; We found something that may perhaps be encoded.
+ (setq method nil
+ alist rfc2047-header-encoding-alist)
+ (while (setq elem (pop alist))
+ (when (or (and (stringp (car elem))
+ (looking-at (car elem)))
+ (eq (car elem) t))
+ (setq alist nil
+ method (cdr elem))))
+ (goto-char (point-min))
+ (re-search-forward "^[^:]+: *" nil t)
+ (cond
+ ((eq method 'address-mime)
+ (rfc2047-encode-region (point) (point-max)))
+ ((eq method 'mime)
+ (let ((rfc2047-encoding-type 'mime))
+ (rfc2047-encode-region (point) (point-max))))
+ ((eq method 'default)
+ (if (and (featurep 'mule)
+ (if (boundp 'default-enable-multibyte-characters)
+ default-enable-multibyte-characters)
+ mail-parse-charset)
+ (mm-encode-coding-region (point) (point-max)
+ mail-parse-charset)))
+ ;; We get this when CC'ing messsages to newsgroups with
+ ;; 8-bit names. The group name mail copy just got
+ ;; unconditionally encoded. Previously, it would ask
+ ;; whether to encode, which was quite confusing for the
+ ;; user. If the new behaviour is wrong, tell me. I have
+ ;; left the old code commented out below.
+ ;; -- Per Abrahamsen <abraham@dina.kvl.dk> Date: 2001-10-07.
+ ;; Modified by Dave Love, with the commented-out code changed
+ ;; in accordance with changes elsewhere.
+ ((null method)
+ (rfc2047-encode-region (point) (point-max)))
+;;; ((null method)
+;;; (if (or (message-options-get
+;;; 'rfc2047-encode-message-header-encode-any)
+;;; (message-options-set
+;;; 'rfc2047-encode-message-header-encode-any
+;;; (y-or-n-p
+;;; "Some texts are not encoded. Encode anyway?")))
+;;; (rfc2047-encode-region (point-min) (point-max))
+;;; (error "Cannot send unencoded text")))
+ ((mm-coding-system-p method)
+ (if (and (featurep 'mule)
+ (if (boundp 'default-enable-multibyte-characters)
+ default-enable-multibyte-characters))
+ (mm-encode-coding-region (point) (point-max) method)))
+ ;; Hm.
+ (t)))
+ (goto-char (point-max)))))))
+
+;; Fixme: This, and the require below may not be the Right Thing, but
+;; should be safe just before release. -- fx 2001-02-08
+(eval-when-compile (defvar message-posting-charset))
+
+(defun rfc2047-encodable-p ()
+ "Return non-nil if any characters in current buffer need encoding in headers.
+The buffer may be narrowed."
+ (require 'message) ; for message-posting-charset
+ (let ((charsets
+ (mm-find-mime-charset-region (point-min) (point-max))))
+ (and charsets
+ (not (equal charsets (list (car message-posting-charset)))))))
+
+;; Use this syntax table when parsing into regions that may need
+;; encoding. Double quotes are string delimiters, backslash is
+;; character quoting, and all other RFC 2822 special characters are
+;; treated as punctuation so we can use forward-sexp/forward-word to
+;; skip to the end of regions appropriately. Nb. ietf-drums does
+;; things differently.
+(defconst rfc2047-syntax-table
+ ;; (make-char-table 'syntax-table '(2)) only works in Emacs.
+ (let ((table (make-syntax-table)))
+ ;; The following is done to work for setting all elements of the table
+ ;; in Emacs 21 and 22 and XEmacs; it appears to be the cleanest way.
+ ;; Play safe and don't assume the form of the word syntax entry --
+ ;; copy it from ?a.
+ (if (fboundp 'set-char-table-range) ; Emacs
+ (funcall (intern "set-char-table-range")
+ table t (aref (standard-syntax-table) ?a))
+ (if (fboundp 'put-char-table)
+ (if (fboundp 'get-char-table) ; warning avoidance
+ (put-char-table t (get-char-table ?a (standard-syntax-table))
+ table))))
+ (modify-syntax-entry ?\\ "\\" table)
+ (modify-syntax-entry ?\" "\"" table)
+ (modify-syntax-entry ?\( "." table)
+ (modify-syntax-entry ?\) "." table)
+ (modify-syntax-entry ?\< "." table)
+ (modify-syntax-entry ?\> "." table)
+ (modify-syntax-entry ?\[ "." table)
+ (modify-syntax-entry ?\] "." table)
+ (modify-syntax-entry ?: "." table)
+ (modify-syntax-entry ?\; "." table)
+ (modify-syntax-entry ?, "." table)
+ (modify-syntax-entry ?@ "." table)
+ table))
(defun rfc2047-encode-region (b e)
- "Encode all encodable words in REGION."
- (let ((words (rfc2047-dissect-region b e))
- beg end current word)
- (while (setq word (pop words))
- (if (equal (nth 2 word) current)
- (setq beg (nth 0 word))
- (when current
- (rfc2047-encode beg end current))
- (setq current (nth 2 word)
- beg (nth 0 word)
- end (nth 1 word))))
- (when current
- (rfc2047-encode beg end current))))
+ "Encode words in region B to E that need encoding.
+By default, the region is treated as containing RFC2822 addresses.
+Dynamically bind `rfc2047-encoding-type' to change that."
+ (save-restriction
+ (narrow-to-region b e)
+ (if (eq 'mime rfc2047-encoding-type)
+ ;; Simple case. Treat as single word after any initial ASCII
+ ;; part and before any tailing ASCII part. The leading ASCII
+ ;; is relevant for instance in Subject headers with `Re:' for
+ ;; interoperability with non-MIME clients, and we might as
+ ;; well avoid the tail too.
+ (progn
+ (goto-char (point-min))
+ ;; Does it need encoding?
+ (skip-chars-forward "\000-\177")
+ (unless (eobp)
+ (skip-chars-backward "^ \n") ; beginning of space-delimited word
+ (rfc2047-encode (point) (progn
+ (goto-char e)
+ (skip-chars-backward "\000-\177")
+ (skip-chars-forward "^ \n")
+ ;; end of space-delimited word
+ (point)))))
+ ;; `address-mime' case -- take care of quoted words, comments.
+ (with-syntax-table rfc2047-syntax-table
+ (let ((start) ; start of current token
+ end ; end of current token
+ ;; Whether there's an encoded word before the current
+ ;; token, either immediately or separated by space.
+ last-encoded)
+ (goto-char (point-min))
+ (condition-case nil ; in case of unbalanced quotes
+ ;; Look for rfc2822-style: sequences of atoms, quoted
+ ;; strings, specials, whitespace. (Specials mustn't be
+ ;; encoded.)
+ (while (not (eobp))
+ (setq start (point))
+ ;; Skip whitespace.
+ (unless (= 0 (skip-chars-forward " \t\n"))
+ (setq start (point)))
+ (cond
+ ((not (char-after))) ; eob
+ ;; else token start
+ ((eq ?\" (char-syntax (char-after)))
+ ;; Quoted word.
+ (forward-sexp)
+ (setq end (point))
+ ;; Does it need encoding?
+ (goto-char start)
+ (skip-chars-forward "\000-\177" end)
+ (if (= end (point))
+ (setq last-encoded nil)
+ ;; It needs encoding. Strip the quotes first,
+ ;; since encoded words can't occur in quotes.
+ (goto-char end)
+ (delete-backward-char 1)
+ (goto-char start)
+ (delete-char 1)
+ (when last-encoded
+ ;; There was a preceding quoted word. We need
+ ;; to include any separating whitespace in this
+ ;; word to avoid it getting lost.
+ (skip-chars-backward " \t")
+ ;; A space is needed between the encoded words.
+ (insert ? )
+ (setq start (point)
+ end (1+ end)))
+ ;; Adjust the end position for the deleted quotes.
+ (rfc2047-encode start (- end 2))
+ (setq last-encoded t))) ; record that it was encoded
+ ((eq ?. (char-syntax (char-after)))
+ ;; Skip other delimiters, but record that they've
+ ;; potentially separated quoted words.
+ (forward-char)
+ (setq last-encoded nil))
+ (t ; normal token/whitespace sequence
+ ;; Find the end.
+ (forward-word 1)
+ (skip-chars-backward " \t")
+ (setq end (point))
+ ;; Deal with encoding and leading space as for
+ ;; quoted words.
+ (goto-char start)
+ (skip-chars-forward "\000-\177" end)
+ (if (= end (point))
+ (setq last-encoded nil)
+ (when last-encoded
+ (goto-char start)
+ (skip-chars-backward " \t")
+ (insert ? )
+ (setq start (point)
+ end (1+ end)))
+ (rfc2047-encode start end)
+ (setq last-encoded t)))))
+ (error
+ (error "Invalid data for rfc2047 encoding: %s"
+ (buffer-substring b e)))))))
+ (rfc2047-fold-region b (point))))
(defun rfc2047-encode-string (string)
- "Encode words in STRING."
+ "Encode words in STRING.
+By default, the string is treated as containing addresses (see
+`rfc2047-encoding-type')."
(with-temp-buffer
(insert string)
(rfc2047-encode-region (point-min) (point-max))
(buffer-string)))
-(defun rfc2047-encode (b e charset)
- "Encode the word in the region with CHARSET."
- (let* ((mime-charset (mm-mime-charset charset))
+(defun rfc2047-encode (b e)
+ "Encode the word(s) in the region B to E.
+By default, the region is treated as containing addresses (see
+`rfc2047-encoding-type')."
+ (let* ((mime-charset (mm-find-mime-charset-region b e))
+ (cs (if (> (length mime-charset) 1)
+ ;; Fixme: Instead of this, try to break region into
+ ;; parts that can be encoded separately.
+ (error "Can't rfc2047-encode `%s'"
+ (buffer-substring b e))
+ (setq mime-charset (car mime-charset))
+ (mm-charset-to-coding-system mime-charset)))
+ ;; Fixme: Better, calculate the number of non-ASCII
+ ;; characters, at least for 8-bit charsets.
(encoding (or (cdr (assq mime-charset
rfc2047-charset-encoding-alist))
- 'B))
+ ;; For the charsets that don't have a preferred
+ ;; encoding, choose the one that's shorter.
+ (save-restriction
+ (narrow-to-region b e)
+ (if (eq (mm-qp-or-base64) 'base64)
+ 'B
+ 'Q))))
(start (concat
"=?" (downcase (symbol-name mime-charset)) "?"
(downcase (symbol-name encoding)) "?"))
+ (factor (case mime-charset
+ ((iso-8859-5 iso-8859-7 iso-8859-8 koi8-r) 1)
+ ((big5 gb2312 euc-kr) 2)
+ (utf-8 4)
+ (t 8)))
+ (pre (- b (save-restriction
+ (widen)
+ (point-at-bol))))
+ ;; encoded-words must not be longer than 75 characters,
+ ;; including charset, encoding etc. This leaves us with
+ ;; 75 - (length start) - 2 - 2 characters. The last 2 is for
+ ;; possible base64 padding. In the worst case (iso-2022-*)
+ ;; each character expands to 8 bytes which is expanded by a
+ ;; factor of 4/3 by base64 encoding.
+ (length (floor (- 75 (length start) 4) (* factor (/ 4.0 3.0))))
+ ;; Limit line length to 76 characters.
+ (length1 (max 1 (floor (- 76 (length start) 4 pre)
+ (* factor (/ 4.0 3.0)))))
(first t))
+ (if mime-charset
+ (save-restriction
+ (narrow-to-region b e)
+ (when (eq encoding 'B)
+ ;; break into lines before encoding
+ (goto-char (point-min))
+ (while (not (eobp))
+ (if first
+ (progn
+ (goto-char (min (point-max) (+ length1 (point))))
+ (setq first nil))
+ (goto-char (min (point-max) (+ length (point)))))
+ (unless (eobp)
+ (insert ?\n)))
+ (setq first t))
+ (if (and (mm-multibyte-p)
+ (mm-coding-system-p cs))
+ (mm-encode-coding-region (point-min) (point-max) cs))
+ (funcall (cdr (assq encoding rfc2047-encoding-function-alist))
+ (point-min) (point-max))
+ (goto-char (point-min))
+ (while (not (eobp))
+ (unless first
+ (insert ? ))
+ (setq first nil)
+ (insert start)
+ (end-of-line)
+ (insert "?=")
+ (forward-line 1))))))
+
+(defun rfc2047-fold-field ()
+ "Fold the current header field."
+ (save-excursion
(save-restriction
- (narrow-to-region b e)
- (when (eq encoding 'B)
- ;; break into lines before encoding
- (goto-char (point-min))
- (while (not (eobp))
- (goto-char (min (point-max) (+ 15 (point))))
- (unless (eobp)
- (insert "\n"))))
- (mm-encode-coding-region (point-min) (point-max) mime-charset)
- (funcall (cdr (assq encoding rfc2047-encoding-function-alist))
- (point-min) (point-max))
- (goto-char (point-min))
- (while (not (eobp))
- (unless first
- (insert " "))
- (setq first nil)
- (insert start)
- (end-of-line)
- (insert "?=")
- (forward-line 1)))))
+ (rfc2047-narrow-to-field)
+ (rfc2047-fold-region (point-min) (point-max)))))
(defun rfc2047-fold-region (b e)
- "Fold the long lines in the region."
+ "Fold long lines in region B to E."
(save-restriction
(narrow-to-region b e)
(goto-char (point-min))
- (let ((break nil))
+ (let ((break nil)
+ (qword-break nil)
+ (first t)
+ (bol (save-restriction
+ (widen)
+ (point-at-bol))))
(while (not (eobp))
+ (when (and (or break qword-break)
+ (> (- (point) bol) 76))
+ (goto-char (or break qword-break))
+ (setq break nil
+ qword-break nil)
+ (if (looking-at "[ \t]")
+ (insert ?\n)
+ (insert "\n "))
+ (setq bol (1- (point)))
+ ;; Don't break before the first non-LWSP characters.
+ (skip-chars-forward " \t")
+ (unless (eobp)
+ (forward-char 1)))
(cond
+ ((eq (char-after) ?\n)
+ (forward-char 1)
+ (setq bol (point)
+ break nil
+ qword-break nil)
+ (skip-chars-forward " \t")
+ (unless (or (eobp) (eq (char-after) ?\n))
+ (forward-char 1)))
+ ((eq (char-after) ?\r)
+ (forward-char 1))
((memq (char-after) '(? ?\t))
- (setq break (point)))
- ((and (not break)
- (looking-at "=\\?"))
- (setq break (point)))
- ((and (looking-at "\\?=")
- (> (- (point) (save-excursion (beginning-of-line) (point))) 76))
- (goto-char break)
- (setq break nil)
- (insert "\n ")))
+ (skip-chars-forward " \t")
+ (if first
+ ;; Don't break just after the header name.
+ (setq first nil)
+ (setq break (1- (point)))))
+ ((not break)
+ (if (not (looking-at "=\\?[^=]"))
+ (if (eq (char-after) ?=)
+ (forward-char 1)
+ (skip-chars-forward "^ \t\n\r="))
+ ;; Don't break at the start of the field.
+ (unless (= (point) b)
+ (setq qword-break (point)))
+ (skip-chars-forward "^ \t\n\r")))
+ (t
+ (skip-chars-forward "^ \t\n\r"))))
+ (when (and (or break qword-break)
+ (> (- (point) bol) 76))
+ (goto-char (or break qword-break))
+ (setq break nil
+ qword-break nil)
+ (if (looking-at "[ \t]")
+ (insert ?\n)
+ (insert "\n "))
+ (setq bol (1- (point)))
+ ;; Don't break before the first non-LWSP characters.
+ (skip-chars-forward " \t")
(unless (eobp)
(forward-char 1))))))
+(defun rfc2047-unfold-field ()
+ "Fold the current line."
+ (save-excursion
+ (save-restriction
+ (rfc2047-narrow-to-field)
+ (rfc2047-unfold-region (point-min) (point-max)))))
+
+(defun rfc2047-unfold-region (b e)
+ "Unfold lines in region B to E."
+ (save-restriction
+ (narrow-to-region b e)
+ (goto-char (point-min))
+ (let ((bol (save-restriction
+ (widen)
+ (point-at-bol)))
+ (eol (point-at-eol)))
+ (forward-line 1)
+ (while (not (eobp))
+ (if (and (looking-at "[ \t]")
+ (< (- (point-at-eol) bol) 76))
+ (delete-region eol (progn
+ (goto-char eol)
+ (skip-chars-forward "\r\n")
+ (point)))
+ (setq bol (point-at-bol)))
+ (setq eol (point-at-eol))
+ (forward-line 1)))))
+
(defun rfc2047-b-encode-region (b e)
- "Encode the header contained in REGION with the B encoding."
+ "Base64-encode the header contained in region B to E."
(save-restriction
(narrow-to-region (goto-char b) e)
(while (not (eobp))
(forward-line))))
(defun rfc2047-q-encode-region (b e)
- "Encode the header contained in REGION with the Q encoding."
+ "Quoted-printable-encode the header in region B to E."
(save-excursion
(save-restriction
(narrow-to-region (goto-char b) e)
- (let ((alist rfc2047-q-encoding-alist))
- (while alist
- (when (looking-at (caar alist))
- (quoted-printable-encode-region b e nil (cdar alist))
- (subst-char-in-region (point-min) (point-max) ? ?_)
- (setq alist nil))
- (pop alist))
- (goto-char (point-min))
- (while (not (eobp))
- (goto-char (min (point-max) (+ 64 (point))))
- (search-backward "=" (- (point) 2) t)
- (unless (eobp)
- (insert "\n")))))))
+ (let ((bol (save-restriction
+ (widen)
+ (point-at-bol))))
+ (quoted-printable-encode-region
+ b e nil
+ ;; = (\075), _ (\137), ? (\077) are used in the encoded word.
+ ;; Avoid using 8bit characters.
+ ;; This list excludes `especials' (see the RFC2047 syntax),
+ ;; meaning that some characters in non-structured fields will
+ ;; get encoded when they con't need to be. The following is
+ ;; what it used to be.
+;;; ;; Equivalent to "^\000-\007\011\013\015-\037\200-\377=_?"
+;;; "\010\012\014\040-\074\076\100-\136\140-\177")
+ "-\b\n\f !#-'*+0-9A-Z\\^`-~\d")
+ (subst-char-in-region (point-min) (point-max) ? ?_)
+ ;; The size of QP encapsulation is about 20, so set limit to
+ ;; 56=76-20.
+ (unless (< (- (point-max) (point-min)) 56)
+ ;; Don't break if it could fit in one line.
+ ;; Let rfc2047-encode-region break it later.
+ (goto-char (1+ (point-min)))
+ (while (and (not (bobp)) (not (eobp)))
+ (goto-char (min (point-max) (+ 56 bol)))
+ (search-backward "=" (- (point) 2) t)
+ (unless (or (bobp) (eobp))
+ (insert ?\n)
+ (setq bol (point)))))))))
;;;
;;; Functions for decoding RFC2047 messages
;;;
-(defvar rfc2047-encoded-word-regexp
- "=\\?\\([^][\000-\040()<>@,\;:\\\"/?.=]+\\)\\?\\(B\\|Q\\)\\?\\([!->@-~ +]+\\)\\?=")
+(eval-and-compile
+ (defconst rfc2047-encoded-word-regexp
+ "=\\?\\([^][\000-\040()<>@,\;:\\\"/?.=]+\\)\\?\\(B\\|Q\\)\
+\\?\\([!->@-~ +]*\\)\\?="))
+
+;; Fixme: This should decode in place, not cons intermediate strings.
+;; Also check whether it needs to worry about delimiting fields like
+;; encoding.
+
+;; In fact it's reported that (invalid) encoding of mailboxes in
+;; addr-specs is in use, so delimiting fields might help. Probably
+;; not decoding a word which isn't properly delimited is good enough
+;; and worthwhile (is it more correct or not?), e.g. something like
+;; `=?iso-8859-1?q?foo?=@'.
(defun rfc2047-decode-region (start end)
"Decode MIME-encoded words in region between START and END."
(goto-char (point-min))
;; Remove whitespace between encoded words.
(while (re-search-forward
- (concat "\\(" rfc2047-encoded-word-regexp "\\)"
- "\\(\n?[ \t]\\)+"
- "\\(" rfc2047-encoded-word-regexp "\\)")
+ (eval-when-compile
+ (concat "\\(" rfc2047-encoded-word-regexp "\\)"
+ "\\(\n?[ \t]\\)+"
+ "\\(" rfc2047-encoded-word-regexp "\\)"))
nil t)
(delete-region (goto-char (match-end 1)) (match-beginning 6)))
;; Decode the encoded words.
(prog1
(match-string 0)
(delete-region (match-beginning 0) (match-end 0)))))
+ ;; Remove newlines between decoded words, though such things
+ ;; essentially must not be there.
+ (save-restriction
+ (narrow-to-region e (point))
+ (goto-char e)
+ (while (re-search-forward "[\n\r]+" nil t)
+ (replace-match " "))
+ (goto-char (point-max)))
(when (and (mm-multibyte-p)
mail-parse-charset
+ (not (eq mail-parse-charset 'us-ascii))
(not (eq mail-parse-charset 'gnus-decoded)))
(mm-decode-coding-region b e mail-parse-charset))
(setq b (point)))
(defun rfc2047-decode-string (string)
"Decode the quoted-printable-encoded STRING and return the results."
(let ((m (mm-multibyte-p)))
- (with-temp-buffer
- (when m
- (mm-enable-multibyte))
- (insert string)
- (inline
- (rfc2047-decode-region (point-min) (point-max)))
- (buffer-string))))
+ (if (string-match "=\\?" string)
+ (with-temp-buffer
+ ;; Fixme: This logic is wrong, but seems to be required by
+ ;; Gnus summary buffer generation. The value of `m' depends
+ ;; on the current buffer, not global multibyteness or that
+ ;; of the string. Also the string returned should always be
+ ;; multibyte in a multibyte session, i.e. the buffer should
+ ;; be multibyte before `buffer-string' is called.
+ (when m
+ (mm-enable-multibyte))
+ (insert string)
+ (inline
+ (rfc2047-decode-region (point-min) (point-max)))
+ (buffer-string))
+ ;; Fixme: As above, `m' here is inappropriate.
+ (if (and m
+ mail-parse-charset
+ (not (eq mail-parse-charset 'us-ascii))
+ (not (eq mail-parse-charset 'gnus-decoded)))
+ ;; `decode-coding-string' in Emacs offers a third optional
+ ;; arg NOCOPY to avoid consing a new string if the decoding
+ ;; is "trivial". Unfortunately it currently doesn't
+ ;; consider anything else than a `nil' coding system
+ ;; trivial.
+ ;; `rfc2047-decode-string' is called multiple times for each
+ ;; article during summary buffer generation, and we really
+ ;; want to avoid unnecessary consing. So we bypass
+ ;; `decode-coding-string' if the string is purely ASCII.
+ (if (and (fboundp 'detect-coding-string)
+ ;; string is purely ASCII
+ (eq (detect-coding-string string t) 'undecided))
+ string
+ (mm-decode-coding-string string mail-parse-charset))
+ (mm-string-as-multibyte string)))))
(defun rfc2047-parse-and-decode (word)
"Decode WORD and return it if it is an encoded word.
-Return WORD if not."
+Return WORD if it is not not an encoded word or if the charset isn't
+decodable."
(if (not (string-match rfc2047-encoded-word-regexp word))
word
(or
(condition-case nil
(rfc2047-decode
(match-string 1 word)
- (upcase (match-string 2 word))
+ (string-to-char (match-string 2 word))
(match-string 3 word))
(error word))
- word)))
+ word))) ; un-decodable
+
+(defun rfc2047-pad-base64 (string)
+ "Pad STRING to quartets."
+ ;; Be more liberal to accept buggy base64 strings. If
+ ;; base64-decode-string accepts buggy strings, this function could
+ ;; be aliased to identity.
+ (case (mod (length string) 4)
+ (0 string)
+ (1 string) ;; Error, don't pad it.
+ (2 (concat string "=="))
+ (3 (concat string "="))))
(defun rfc2047-decode (charset encoding string)
- "Decode STRING that uses CHARSET with ENCODING.
-Valid ENCODINGs are \"B\" and \"Q\".
-If your Emacs implementation can't decode CHARSET, it returns nil."
+ "Decode STRING from the given MIME CHARSET in the given ENCODING.
+Valid ENCODINGs are the characters \"B\" and \"Q\".
+If your Emacs implementation can't decode CHARSET, return nil."
(if (stringp charset)
(setq charset (intern (downcase charset))))
- (if (or (not charset)
+ (if (or (not charset)
(eq 'gnus-all mail-parse-ignored-charsets)
(memq 'gnus-all mail-parse-ignored-charsets)
(memq charset mail-parse-ignored-charsets))
(setq charset mail-parse-charset))
(let ((cs (mm-charset-to-coding-system charset)))
- (if (and (not cs) charset
+ (if (and (not cs) charset
(listp mail-parse-ignored-charsets)
(memq 'gnus-unknown mail-parse-ignored-charsets))
(setq cs (mm-charset-to-coding-system mail-parse-charset)))
(setq cs mail-parse-charset))
(mm-decode-coding-string
(cond
- ((equal "B" encoding)
- (base64-decode-string string))
- ((equal "Q" encoding)
+ ((char-equal ?B encoding)
+ (base64-decode-string
+ (rfc2047-pad-base64 string)))
+ ((char-equal ?Q encoding)
(quoted-printable-decode-string
- (mm-replace-chars-in-string string ?_ ? )))
- (t (error "Invalid encoding: %s" encoding)))
+ (mm-subst-char-in-string ?_ ? string t)))
+ (t (error "Invalid encoding: %c" encoding)))
cs))))
(provide 'rfc2047)