Fix typo in last commit.

[gnus] / lisp / rfc2047.el

diff --git a/lisp/rfc2047.el b/lisp/rfc2047.el
index e4a7626..d4d6e96 100644 (file)
--- a/lisp/rfc2047.el
+++ b/lisp/rfc2047.el
@@ -71,8 +71,8 @@ Value is what BODY returns."
   '(("Newsgroups" . nil)
     ("Followup-To" . nil)
     ("Message-ID" . nil)
-    ("\\(Resent-\\)?\\(From\\|Cc\\|To\\|Bcc\\|Reply-To\\|Sender\\)" .
-     address-mime)
+    ("\\(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
@@ -110,7 +110,8 @@ The values can be:
     (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'.  These indicate binary (no) encoding,
 quoted-printable and base64 respectively.")
@@ -133,9 +134,7 @@ quoted-printable and base64 respectively.")
    (progn
      (forward-line 1)
      (if (re-search-forward "^[^ \n\t]" nil t)
-        (progn
-          (beginning-of-line)
-          (point))
+        (rfc2047-point-at-bol)
        (point-max))))
   (goto-char (point-min)))
 
@@ -378,8 +377,9 @@ Dynamically bind `rfc2047-encoding-type' to change that."
                            end (1+ end)))
                    (rfc2047-encode start end)
                    (setq last-encoded t)))))
-           (error (error "Invalid data for rfc2047 encoding: %s"
-                         (buffer-substring b e)))))))
+           (error
+            (error "Invalid data for rfc2047 encoding: %s"
+                   (buffer-substring b e)))))))
     (rfc2047-fold-region b (point))))
 
 (defun rfc2047-encode-string (string)
@@ -405,14 +405,36 @@ By default, the region is treated as containing addresses (see
               (mm-charset-to-coding-system mime-charset)))
         ;; Fixme: Better, calculate the number of non-ASCII
         ;; characters, at least for 8-bit charsets.
-        (encoding (if (assq mime-charset
-                            rfc2047-charset-encoding-alist)
-                      (cdr (assq mime-charset
+        (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)
+                    (rfc2047-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
@@ -421,9 +443,14 @@ By default, the region is treated as containing addresses (see
            ;; break into lines before encoding
            (goto-char (point-min))
            (while (not (eobp))
-             (goto-char (min (point-max) (+ 15 (point))))
+             (if first
+                 (progn
+                   (goto-char (min (point-max) (+ length1 (point))))
+                   (setq first nil))
+               (goto-char (min (point-max) (+ length (point)))))
              (unless (eobp)
-               (insert ?\n))))
+               (insert ?\n)))
+           (setq first t))
          (if (and (mm-multibyte-p)
                   (mm-coding-system-p cs))
              (mm-encode-coding-region (point-min) (point-max) cs))
@@ -493,7 +520,9 @@ By default, the region is treated as containing addresses (see
              (if (eq (char-after) ?=)
                  (forward-char 1)
                (skip-chars-forward "^ \t\n\r="))
-           (setq qword-break (point))
+           ;; 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"))))
@@ -561,8 +590,13 @@ By default, the region is treated as containing addresses (see
         b e nil
         ;; = (\075), _ (\137), ? (\077) are used in the encoded word.
         ;; Avoid using 8bit characters.
-        ;; Equivalent to "^\000-\007\011\013\015-\037\200-\377=_?"
-        "\010\012\014\040-\074\076\100-\136\140-\177")
+        ;; 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.