;;; nndoc.el --- single file access for Gnus
-;; Copyright (C) 1995,96,97,98 Free Software Foundation, Inc.
+;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002
+;; Free Software Foundation, Inc.
;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
-;; Masanobu UMEDA <umerin@flab.flab.fujitsu.junet>
+;; Masanobu UMEDA <umerin@flab.flab.fujitsu.junet>
;; Keywords: news
;; This file is part of GNU Emacs.
;;; Commentary:
+;; For Outlook mail boxes format, see http://mbx2mbox.sourceforge.net/
+
;;; Code:
(require 'nnheader)
(require 'message)
(require 'nnmail)
(require 'nnoo)
+(require 'gnus-util)
+(require 'mm-util)
(eval-when-compile (require 'cl))
(nnoo-declare nndoc)
(defvoo nndoc-article-type 'guess
"*Type of the file.
One of `mbox', `babyl', `digest', `news', `rnews', `mmdf', `forward',
-`rfc934', `rfc822-forward', `mime-digest', `standard-digest',
-`slack-digest', `clari-briefs' or `guess'.")
+`rfc934', `rfc822-forward', `mime-parts', `standard-digest',
+`slack-digest', `clari-briefs', `nsmail', `outlook', `oe-dbx',
+`mailman', `exim-bounce', or `guess'.")
(defvoo nndoc-post-type 'mail
"*Whether the nndoc group is `mail' or `post'.")
(defvoo nndoc-open-document-hook 'nnheader-ms-strip-cr
"Hook run after opening a document.
The default function removes all trailing carriage returns
-from the document.")
+from the document.")
(defvar nndoc-type-alist
`((mmdf
(article-begin . "^\^A\^A\^A\^A\n")
(body-end . "^\^A\^A\^A\^A\n"))
+ (exim-bounce
+ (article-begin . "^------ This is a copy of the message, including all the headers. ------\n\n")
+ (body-end-function . nndoc-exim-bounce-body-end-function))
+ (nsmail
+ (article-begin . "^From - "))
(news
(article-begin . "^Path:"))
(rnews
(body-end . "\^_")
(body-begin-function . nndoc-babyl-body-begin)
(head-begin-function . nndoc-babyl-head-begin))
- (forward
- (article-begin . "^-+ Start of forwarded message -+\n+")
- (body-end . "^-+ End of forwarded message -+$")
- (prepare-body-function . nndoc-unquote-dashes))
(rfc934
(article-begin . "^--.*\n+")
(body-end . "^--.*$")
(prepare-body-function . nndoc-unquote-dashes))
+ (mailman
+ (article-begin . "^--__--__--\n\nMessage:")
+ (body-end . "^--__--__--$")
+ (prepare-body-function . nndoc-unquote-dashes))
(clari-briefs
(article-begin . "^ \\*")
(body-end . "^\t------*[ \t]^*\n^ \\*")
(article-transform-function . nndoc-transform-clari-briefs))
(mime-digest
(article-begin . "")
+ (head-begin . "^ ?\n")
(head-end . "^ ?$")
(body-end . "")
(file-end . "")
(subtype digest guess))
+ (mime-parts
+ (generate-head-function . nndoc-generate-mime-parts-head)
+ (article-transform-function . nndoc-transform-mime-parts))
(standard-digest
(first-article . ,(concat "^" (make-string 70 ?-) "\n *\n+"))
(article-begin . ,(concat "^\n" (make-string 30 ?-) "\n *\n+"))
(head-begin . "^Paper.*:")
(head-end . "\\(^\\\\\\\\.*\n\\|-----------------\\)")
(body-begin . "")
- (body-end . "-------------------------------------------------")
- (file-end . "^Title: Recent Seminal")
+ (body-end . "\\(-------------------------------------------------\\|%-%-%-%-%-%-%-%-%-%-%-%-%-%-\\|%%--%%--%%--%%--%%--%%--%%--%%--\\|%%%---%%%---%%%---%%%---\\)")
+ (file-end . "\\(^Title: Recent Seminal\\|%%%---%%%---%%%---%%%---\\)")
(generate-head-function . nndoc-generate-lanl-gov-head)
(article-transform-function . nndoc-transform-lanl-gov-announce)
(subtype preprints guess))
(rfc822-forward
(article-begin . "^\n")
(body-end-function . nndoc-rfc822-forward-body-end-function))
+ (outlook
+ (article-begin-function . nndoc-outlook-article-begin)
+ (body-end . "\0"))
+ (oe-dbx ;; Outlook Express DBX format
+ (dissection-function . nndoc-oe-dbx-dissection)
+ (generate-head-function . nndoc-oe-dbx-generate-head)
+ (generate-article-function . nndoc-oe-dbx-generate-article))
+ (forward
+ (article-begin . "^-+ \\(Start of \\)?forwarded message.*\n+")
+ (body-end . "^-+ End \\(of \\)?forwarded message.*$")
+ (prepare-body-function . nndoc-unquote-dashes))
+ (mail-in-mail ;; Wild guess on mailer daemon's messages or others
+ (article-begin-function . nndoc-mail-in-mail-article-begin))
(guess
(guess . t)
(subtype nil))
(guess . t)
(subtype nil))))
-\f
+(defvar nndoc-binary-file-names ".[Dd][Bb][Xx]$"
+ "Regexp for binary nndoc file names.")
+\f
(defvoo nndoc-file-begin nil)
(defvoo nndoc-first-article nil)
-(defvoo nndoc-article-end nil)
(defvoo nndoc-article-begin nil)
(defvoo nndoc-head-begin nil)
(defvoo nndoc-head-end nil)
(defvoo nndoc-body-begin-function nil)
(defvoo nndoc-head-begin-function nil)
(defvoo nndoc-body-end nil)
+;; nndoc-dissection-alist is a list of sublists. Each sublist holds the
+;; following items. ARTICLE acts as the association key and is an ordinal
+;; starting at 1. HEAD-BEGIN [0], HEAD-END [1], BODY-BEGIN [2] and BODY-END
+;; [3] are positions in the `nndoc' buffer. LINE-COUNT [4] is a count of
+;; lines in the body. For MIME dissections only, ARTICLE-INSERT [5] and
+;; SUMMARY-INSERT [6] give headers to insert for full article or summary line
+;; generation, respectively. Other headers usually follow directly from the
+;; buffer. Value `nil' means no insert.
(defvoo nndoc-dissection-alist nil)
(defvoo nndoc-prepare-body-function nil)
(defvoo nndoc-generate-head-function nil)
(defvoo nndoc-article-transform-function nil)
(defvoo nndoc-article-begin-function nil)
+(defvoo nndoc-generate-article-function nil)
+(defvoo nndoc-dissection-function nil)
(defvoo nndoc-status-string "")
(defvoo nndoc-group-alist nil)
(insert-buffer-substring
nndoc-current-buffer (car entry) (nth 1 entry)))
(goto-char (point-max))
- (unless (= (char-after (1- (point))) ?\n)
+ (unless (eq (char-after (1- (point))) ?\n)
(insert "\n"))
(insert (format "Lines: %d\n" (nth 4 entry)))
(insert ".\n")))
(set-buffer buffer)
(erase-buffer)
(when entry
- (if (stringp article)
- nil
+ (cond
+ ((stringp article) nil)
+ (nndoc-generate-article-function
+ (funcall nndoc-generate-article-function article))
+ (t
(insert-buffer-substring
nndoc-current-buffer (car entry) (nth 1 entry))
(insert "\n")
(funcall nndoc-prepare-body-function))
(when nndoc-article-transform-function
(funcall nndoc-article-transform-function article))
- t)))))
+ t))))))
(deffoo nndoc-request-group (group &optional server dont-check)
"Select news GROUP."
(deffoo nndoc-request-type (group &optional article)
(cond ((not article) 'unknown)
- (nndoc-post-type nndoc-post-type)
- (t 'unknown)))
+ (nndoc-post-type nndoc-post-type)
+ (t 'unknown)))
(deffoo nndoc-close-group (group &optional server)
(nndoc-possibly-change-buffer group server)
(setq nndoc-dissection-alist nil)
(save-excursion
(set-buffer nndoc-current-buffer)
- (buffer-disable-undo (current-buffer))
(erase-buffer)
- (if (stringp nndoc-address)
- (nnheader-insert-file-contents nndoc-address)
- (insert-buffer-substring nndoc-address))
- (run-hooks 'nndoc-open-document-hook))))
+ (if (and (stringp nndoc-address)
+ (string-match nndoc-binary-file-names nndoc-address))
+ (let ((coding-system-for-read 'binary))
+ (mm-insert-file-contents nndoc-address))
+ (if (stringp nndoc-address)
+ (nnheader-insert-file-contents nndoc-address)
+ (insert-buffer-substring nndoc-address))
+ (run-hooks 'nndoc-open-document-hook)))))
;; Initialize the nndoc structures according to this new document.
(when (and nndoc-current-buffer
(not nndoc-dissection-alist))
(save-excursion
(set-buffer nndoc-current-buffer)
(nndoc-set-delims)
- (nndoc-dissect-buffer)))
+ (if (eq nndoc-article-type 'mime-parts)
+ (nndoc-dissect-mime-parts)
+ (nndoc-dissect-buffer))))
(unless nndoc-current-buffer
(nndoc-close-server))
;; Return whether we managed to select a file.
"Set the nndoc delimiter variables according to the type of the document."
(let ((vars '(nndoc-file-begin
nndoc-first-article
- nndoc-article-end nndoc-head-begin nndoc-head-end
+ nndoc-article-begin-function
+ nndoc-head-begin nndoc-head-end
nndoc-file-end nndoc-article-begin
nndoc-body-begin nndoc-body-end-function nndoc-body-end
nndoc-prepare-body-function nndoc-article-transform-function
nndoc-generate-head-function nndoc-body-begin-function
- nndoc-head-begin-function)))
+ nndoc-head-begin-function
+ nndoc-generate-article-function
+ nndoc-dissection-function)))
(while vars
(set (pop vars) nil)))
(let (defs)
(setq entry (pop alist)))
(when (memq subtype (or (cdr (assq 'subtype entry)) '(guess)))
(goto-char (point-min))
+ ;; Remove blank lines.
+ (while (eq (following-char) ?\n)
+ (delete-char 1))
(when (numberp (setq result (funcall (intern
(format "nndoc-%s-type-p"
(car entry))))))
t))
(defun nndoc-forward-type-p ()
- (when (and (re-search-forward "^-+ Start of forwarded message -+\n+" nil t)
- (not (re-search-forward "^Subject:.*digest" nil t))
- (not (re-search-backward "^From:" nil t 2))
- (not (re-search-forward "^From:" nil t 2)))
+ (when (and (re-search-forward "^-+ \\(Start of \\)?forwarded message.*\n+"
+ nil t)
+ (looking-at "[\r\n]*[a-zA-Z][a-zA-Z0-9-]*:"))
t))
(defun nndoc-rfc934-type-p ()
(not (re-search-forward "^From:" nil t 2)))
t))
+(defun nndoc-mailman-type-p ()
+ (when (re-search-forward "^--__--__--\n+" nil t)
+ t))
+
(defun nndoc-rfc822-forward-type-p ()
(save-restriction
(message-narrow-to-head)
(defun nndoc-rfc822-forward-body-end-function ()
(goto-char (point-max)))
+(defun nndoc-mime-parts-type-p ()
+ (let ((case-fold-search t)
+ (limit (search-forward "\n\n" nil t)))
+ (goto-char (point-min))
+ (when (and limit
+ (re-search-forward
+ (concat "\
+^Content-Type:[ \t]*multipart/[a-z]+ *; *\\(\\(\n[ \t]\\)?.*;\\)*"
+ "\\(\n[ \t]\\)?[ \t]*boundary=\"?[^\"\n]*[^\" \t\n]")
+ limit t))
+ t)))
+
+(defun nndoc-transform-mime-parts (article)
+ (let* ((entry (cdr (assq article nndoc-dissection-alist)))
+ (headers (nth 5 entry)))
+ (when headers
+ (goto-char (point-min))
+ (insert headers))))
+
+(defun nndoc-generate-mime-parts-head (article)
+ (let* ((entry (cdr (assq article nndoc-dissection-alist)))
+ (headers (nth 6 entry)))
+ (save-restriction
+ (narrow-to-region (point) (point))
+ (insert-buffer-substring
+ nndoc-current-buffer (car entry) (nth 1 entry))
+ (goto-char (point-max)))
+ (when headers
+ (insert headers))))
+
(defun nndoc-clari-briefs-type-p ()
(when (let ((case-fold-search nil))
(re-search-forward "^\t[^a-z]+ ([^a-z]+) --" nil t))
(insert "From: " "clari@clari.net (" (or from "unknown") ")"
"\nSubject: " (or subject "(no subject)") "\n")))
+(defun nndoc-exim-bounce-type-p ()
+ (and (re-search-forward "^------ This is a copy of the message, including all the headers. ------" nil t)
+ t))
+
+(defun nndoc-exim-bounce-body-end-function ()
+ (goto-char (point-max)))
+
+
(defun nndoc-mime-digest-type-p ()
(let ((case-fold-search t)
boundary-id b-delimiter entry)
(when (and
(re-search-forward
(concat "^Content-Type: *multipart/digest;[ \t\n]*[ \t]"
- "boundary=\"\\([^\"\n]*[^\" \t\n]\\)\"")
+ "boundary=\"?\\([^\"\n]*[^\" \t\n]\\)")
nil t)
(match-beginning 1))
(setq boundary-id (match-string 1)
- b-delimiter (concat "\n--" boundary-id "[\n \t]+"))
+ b-delimiter (concat "\n--" boundary-id "[ \t]*$"))
(setq entry (assq 'mime-digest nndoc-type-alist))
(setcdr entry
(list
+ (cons 'head-begin "^ ?\n")
(cons 'head-end "^ ?$")
(cons 'body-begin "^ ?\n")
(cons 'article-begin b-delimiter)
(cons 'body-end-function 'nndoc-digest-body-end)
- (cons 'file-end (concat "\n--" boundary-id "--[ \t]*$"))))
+ (cons 'file-end (concat "^--" boundary-id "--[ \t]*$"))))
t)))
(defun nndoc-standard-digest-type-p ()
(defun nndoc-lanl-gov-announce-type-p ()
(when (let ((case-fold-search nil))
- (re-search-forward "^\\\\\\\\\nPaper: [a-z-]+/[0-9]+" nil t))
+ (re-search-forward "^\\\\\\\\\nPaper\\( (\\*cross-listing\\*)\\)?: [a-zA-Z-\\.]+/[0-9]+" nil t))
t))
(defun nndoc-transform-lanl-gov-announce (article)
(goto-char (point-max))
- (when (re-search-backward "^\\\\\\\\ +(\\([^ ]*\\) , *\\([^ ]*\\))" nil t)
+ (when (re-search-backward "^\\\\\\\\ +( *\\([^ ]*\\) , *\\([^ ]*\\))" nil t)
(replace-match "\n\nGet it at \\1 (\\2)" t nil))
- ;; (when (re-search-backward "^\\\\\\\\$" nil t)
- ;; (replace-match "" t t))
- )
+ (goto-char (point-min))
+ (while (re-search-forward "^\\\\\\\\$" nil t)
+ (replace-match "" t nil))
+ (goto-char (point-min))
+ (when (re-search-forward "^replaced with revised version +\\(.*[^ ]\\) +" nil t)
+ (replace-match "Date: \\1 (revised) " t nil))
+ (goto-char (point-min))
+ (unless (re-search-forward "^From" nil t)
+ (goto-char (point-min))
+ (when (re-search-forward "^Authors?: \\(.*\\)" nil t)
+ (goto-char (point-min))
+ (insert "From: " (match-string 1) "\n"))))
(defun nndoc-generate-lanl-gov-head (article)
(let ((entry (cdr (assq article nndoc-dissection-alist)))
- (e-mail "no address given")
- subject from)
+ (from "<no address given>")
+ subject date)
(save-excursion
(set-buffer nndoc-current-buffer)
(save-restriction
- (narrow-to-region (car entry) (nth 1 entry))
- (goto-char (point-min))
- (when (looking-at "^Paper.*: \\([a-z-]+/[0-9]+\\)")
- (setq subject (concat " (" (match-string 1) ")"))
- (when (re-search-forward "^From: \\([^ ]+\\)" nil t)
- (setq e-mail (match-string 1)))
- (when (re-search-forward "^Title: \\([^\f]*\\)\nAuthors?: \\(.*\\)"
- nil t)
- (setq subject (concat (match-string 1) subject))
- (setq from (concat (match-string 2) " <" e-mail ">"))))
- ))
+ (narrow-to-re