lisp/nndoc.el

   1 ;;; nndoc.el --- single file access for Gnus
   2 ;; Copyright (C) 1995,96 Free Software Foundation, Inc.
   3
   4 ;; Author: Lars Magne Ingebrigtsen <larsi@ifi.uio.no>
   5 ;;      Masanobu UMEDA <umerin@flab.flab.fujitsu.junet>
   6 ;; Keywords: news
   7
   8 ;; This file is part of GNU Emacs.
   9
  10 ;; GNU Emacs is free software; you can redistribute it and/or modify
  11 ;; it under the terms of the GNU General Public License as published by
  12 ;; the Free Software Foundation; either version 2, or (at your option)
  13 ;; any later version.
  14
  15 ;; GNU Emacs is distributed in the hope that it will be useful,
  16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18 ;; GNU General Public License for more details.
  19
  20 ;; You should have received a copy of the GNU General Public License
  21 ;; along with GNU Emacs; see the file COPYING.  If not, write to the
  22 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  23 ;; Boston, MA 02111-1307, USA.
  24
  25 ;;; Commentary:
  26
  27 ;;; Code:
  28
  29 (require 'nnheader)
  30 (require 'message)
  31 (require 'nnmail)
  32 (require 'nnoo)
  33 (eval-when-compile (require 'cl))
  34
  35 (nnoo-declare nndoc)
  36
  37 (defvoo nndoc-article-type 'guess
  38   "*Type of the file.
  39 One of `mbox', `babyl', `digest', `news', `rnews', `mmdf', `forward',
  40 `mime-digest', `standard-digest', `slack-digest', `clari-briefs' or
  41 `guess'.")
  42
  43 (defvoo nndoc-post-type 'mail
  44   "*Whether the nndoc group is `mail' or `post'.")
  45
  46 (defvar nndoc-type-alist
  47   `((mmdf
  48      (article-begin .  "^\^A\^A\^A\^A\n")
  49      (body-end .  "^\^A\^A\^A\^A\n"))
  50     (news
  51      (article-begin . "^Path:"))
  52     (rnews
  53      (article-begin . "^#! *rnews +\\([0-9]+\\) *\n")
  54      (body-end-function . nndoc-rnews-body-end))
  55     (mbox
  56      (article-begin . "^From \\([^ \n]*\\(\\|\".*\"[^ \n]*\\)\\)  ?\\([^ \n]*\\) *\\([^ ]*\\) *\\([0-9]*\\) *\\([0-9:]*\\) *\\([A-Z]?[A-Z]?[A-Z][A-Z]\\( DST\\)?\\|[-+]?[0-9][0-9][0-9][0-9]\\|\\) * [0-9][0-9]\\([0-9]*\\) *\\([A-Z]?[A-Z]?[A-Z][A-Z]\\( DST\\)?\\|[-+]?[0-9][0-9][0-9][0-9]\\|\\) *\\(remote from .*\\)?\n")
  57      (article-begin-function . nndoc-mbox-article-begin)
  58      (body-end-function . nndoc-mbox-body-end))
  59     (babyl
  60      (article-begin . "\^_\^L *\n")
  61      (body-end . "\^_")
  62      (body-begin-function . nndoc-babyl-body-begin)
  63      (head-begin-function . nndoc-babyl-head-begin))
  64     (forward
  65      (article-begin . "^-+ Start of forwarded message -+\n+")
  66      (body-end . "^-+ End of forwarded message -+$")
  67      (prepare-body . nndoc-unquote-dashes))
  68     (clari-briefs
  69      (article-begin . "^ \\*")
  70      (body-end . "^\t------*[ \t]^*\n^ \\*")
  71      (body-begin . "^\t")
  72      (head-end . "^\t")
  73      (generate-head . nndoc-generate-clari-briefs-head)
  74      (article-transform . nndoc-transform-clari-briefs))
  75     (slack-digest
  76      (article-begin . "^------------------------------*[\n \t]+")
  77      (head-end . "^ ?$")
  78      (body-end-function . nndoc-digest-body-end)
  79      (body-begin . "^ ?$")
  80      (file-end . "^End of")
  81      (prepare-body . nndoc-unquote-dashes))
  82     (mime-digest
  83      (article-begin . "")
  84      (head-end . "^ ?$")
  85      (body-end . "")
  86      (file-end . ""))
  87     (standard-digest
  88      (first-article . ,(concat "^" (make-string 70 ?-) "\n\n+"))
  89      (article-begin . ,(concat "\n\n" (make-string 30 ?-) "\n\n+"))
  90      (prepare-body . nndoc-unquote-dashes)
  91      (body-end-function . nndoc-digest-body-end)
  92      (head-end . "^ ?$")
  93      (body-begin . "^ ?\n")
  94      (file-end . "^End of .*digest.*[0-9].*\n\\*\\*\\|^End of.*Digest *$"))
  95     (guess
  96      (guess . nndoc-guess-type))
  97     (digest
  98      (guess . nndoc-guess-digest-type))
  99     ))
 100
 101 \f
 102
 103 (defvoo nndoc-file-begin nil)
 104 (defvoo nndoc-first-article nil)
 105 (defvoo nndoc-article-end nil)
 106 (defvoo nndoc-article-begin nil)
 107 (defvoo nndoc-article-begin-function nil)
 108 (defvoo nndoc-head-begin nil)
 109 (defvoo nndoc-head-end nil)
 110 (defvoo nndoc-file-end nil)
 111 (defvoo nndoc-body-begin nil)
 112 (defvoo nndoc-body-end-function nil)
 113 (defvoo nndoc-body-begin-function nil)
 114 (defvoo nndoc-head-begin-function nil)
 115 (defvoo nndoc-body-end nil)
 116 (defvoo nndoc-dissection-alist nil)
 117 (defvoo nndoc-prepare-body nil)
 118 (defvoo nndoc-generate-head nil)
 119 (defvoo nndoc-article-transform nil)
 120
 121 (defvoo nndoc-status-string "")
 122 (defvoo nndoc-group-alist nil)
 123 (defvoo nndoc-current-buffer nil
 124   "Current nndoc news buffer.")
 125 (defvoo nndoc-address nil)
 126
 127 (defconst nndoc-version "nndoc 1.0"
 128   "nndoc version.")
 129
 130 \f
 131
 132 ;;; Interface functions
 133
 134 (nnoo-define-basics nndoc)
 135
 136 (deffoo nndoc-retrieve-headers (articles &optional newsgroup server fetch-old)
 137   (when (nndoc-possibly-change-buffer newsgroup server)
 138     (save-excursion
 139       (set-buffer nntp-server-buffer)
 140       (erase-buffer)
 141       (let (article entry)
 142         (if (stringp (car articles))
 143             'headers
 144           (while articles
 145             (when (setq entry (cdr (assq (setq article (pop articles))
 146                                          nndoc-dissection-alist)))
 147               (insert (format "221 %d Article retrieved.\n" article))
 148               (if nndoc-generate-head
 149                   (funcall nndoc-generate-head article)
 150                 (insert-buffer-substring
 151                  nndoc-current-buffer (car entry) (nth 1 entry)))
 152               (goto-char (point-max))
 153               (or (= (char-after (1- (point))) ?\n) (insert "\n"))
 154               (insert (format "Lines: %d\n" (nth 4 entry)))
 155               (insert ".\n")))
 156
 157           (nnheader-fold-continuation-lines)
 158           'headers)))))
 159
 160 (deffoo nndoc-request-article (article &optional newsgroup server buffer)
 161   (nndoc-possibly-change-buffer newsgroup server)
 162   (save-excursion
 163     (let ((buffer (or buffer nntp-server-buffer))
 164           (entry (cdr (assq article nndoc-dissection-alist)))
 165           beg)
 166       (set-buffer buffer)
 167       (erase-buffer)
 168       (if (stringp article)
 169           nil
 170         (insert-buffer-substring
 171          nndoc-current-buffer (car entry) (nth 1 entry))
 172         (insert "\n")
 173         (setq beg (point))
 174         (insert-buffer-substring
 175          nndoc-current-buffer (nth 2 entry) (nth 3 entry))
 176         (goto-char beg)
 177         (when nndoc-prepare-body
 178           (funcall nndoc-prepare-body))
 179         (when nndoc-article-transform
 180           (funcall nndoc-article-transform article))
 181         t))))
 182
 183 (deffoo nndoc-request-group (group &optional server dont-check)
 184   "Select news GROUP."
 185   (let (number)
 186     (cond
 187      ((not (nndoc-possibly-change-buffer group server))
 188       (nnheader-report 'nndoc "No such file or buffer: %s"
 189                        nndoc-address))
 190      (dont-check
 191       (nnheader-report 'nndoc "Selected group %s" group)
 192       t)
 193      ((zerop (setq number (length nndoc-dissection-alist)))
 194       (nndoc-close-group group)
 195       (nnheader-report 'nndoc "No articles in group %s" group))
 196      (t
 197       (nnheader-insert "211 %d %d %d %s\n" number 1 number group)))))
 198
 199 (deffoo nndoc-request-type (group &optional article)
 200   (cond ((not article) 'unknown)
 201         (nndoc-post-type nndoc-post-type)
 202         (t 'unknown)))
 203
 204 (deffoo nndoc-close-group (group &optional server)
 205   (nndoc-possibly-change-buffer group server)
 206   (and nndoc-current-buffer
 207        (buffer-name nndoc-current-buffer)
 208        (kill-buffer nndoc-current-buffer))
 209   (setq nndoc-group-alist (delq (assoc group nndoc-group-alist)
 210                                 nndoc-group-alist))
 211   (setq nndoc-current-buffer nil)
 212   (nnoo-close-server 'nndoc server)
 213   (setq nndoc-dissection-alist nil)
 214   t)
 215
 216 (deffoo nndoc-request-list (&optional server)
 217   nil)
 218
 219 (deffoo nndoc-request-newgroups (date &optional server)
 220   nil)
 221
 222 (deffoo nndoc-request-list-newsgroups (&optional server)
 223   nil)
 224
 225 \f
 226 ;;; Internal functions.
 227
 228 (defun nndoc-possibly-change-buffer (group source)
 229   (let (buf)
 230     (cond
 231      ;; The current buffer is this group's buffer.
 232      ((and nndoc-current-buffer
 233            (buffer-name nndoc-current-buffer)
 234            (eq nndoc-current-buffer
 235                (setq buf (cdr (assoc group nndoc-group-alist))))))
 236      ;; We change buffers by taking an old from the group alist.
 237      ;; `source' is either a string (a file name) or a buffer object.
 238      (buf
 239       (setq nndoc-current-buffer buf))
 240      ;; It's a totally new group.
 241      ((or (and (bufferp nndoc-address)
 242                (buffer-name nndoc-address))
 243           (and (stringp nndoc-address)
 244                (file-exists-p nndoc-address)
 245                (not (file-directory-p nndoc-address))))
 246       (push (cons group (setq nndoc-current-buffer
 247                               (get-buffer-create
 248                                (concat " *nndoc " group "*"))))
 249             nndoc-group-alist)
 250       (setq nndoc-dissection-alist nil)
 251       (save-excursion
 252         (set-buffer nndoc-current-buffer)
 253         (buffer-disable-undo (current-buffer))
 254         (erase-buffer)
 255         (if (stringp nndoc-address)
 256             (insert-file-contents nndoc-address)
 257           (insert-buffer-substring nndoc-address)))))
 258     ;; Initialize the nndoc structures according to this new document.
 259     (when (and nndoc-current-buffer
 260                (not nndoc-dissection-alist))
 261       (save-excursion
 262         (set-buffer nndoc-current-buffer)
 263         (nndoc-set-delims)
 264         (nndoc-dissect-buffer)))
 265     (unless nndoc-current-buffer
 266       (nndoc-close-server))
 267     ;; Return whether we managed to select a file.
 268     nndoc-current-buffer))
 269
 270 ;; MIME (RFC 1341) digest hack by Ulrik Dickow <dickow@nbi.dk>.
 271 (defun nndoc-guess-digest-type ()
 272   "Guess what digest type the current document is."
 273   (let ((case-fold-search t)            ; We match a bit too much, keep it simple.
 274         boundary-id b-delimiter entry)
 275     (goto-char (point-min))
 276     (cond
 277      ;; MIME digest.
 278      ((and
 279        (re-search-forward
 280         (concat "^Content-Type: *multipart/digest;[ \t\n]*[ \t]"
 281                 "boundary=\"\\([^\"\n]*[^\" \t\n]\\)\"")
 282         nil t)
 283        (match-beginning 1))
 284       (setq boundary-id (match-string 1)
 285             b-delimiter (concat "\n--" boundary-id "[\n \t]+"))
 286       (setq entry (assq 'mime-digest nndoc-type-alist))
 287       (setcdr entry
 288               (list
 289                (cons 'head-end "^ ?$")
 290                (cons 'body-begin "^ ?\n")
 291                (cons 'article-begin b-delimiter)
 292                (cons 'body-end-function 'nndoc-digest-body-end)
 293 ;              (cons 'body-end
 294 ;                    (concat "\n--" boundary-id "\\(--\\)?[\n \t]+"))
 295                (cons 'file-end (concat "\n--" boundary-id "--[ \t]*$"))))
 296       'mime-digest)
 297      ;; Standard digest.
 298      ((and (re-search-forward (concat "^" (make-string 70 ?-) "\n\n") nil t)
 299            (re-search-forward
 300             (concat "\n\n" (make-string 30 ?-) "\n\n") nil t))
 301       'standard-digest)
 302      ;; Stupid digest.
 303      (t
 304       'slack-digest))))
 305
 306 (defun nndoc-guess-type ()
 307   "Guess what document type is in the current buffer."
 308   (goto-char (point-min))
 309   (cond
 310    ((looking-at message-unix-mail-delimiter)
 311     'mbox)
 312    ((looking-at "\^A\^A\^A\^A$")
 313     'mmdf)
 314    ((looking-at "^Path:.*\n")
 315     'news)
 316    ((looking-at "#! *rnews")
 317     'rnews)
 318    ((re-search-forward "\^_\^L *\n" nil t)
 319     'babyl)
 320    ((save-excursion
 321       (and (re-search-forward "^-+ Start of forwarded message -+\n+" nil t)
 322            (not (re-search-forward "^Subject:.*digest" nil t))))
 323     'forward)
 324    ((let ((case-fold-search nil))
 325       (re-search-forward "^\t[^a-z]+ ([^a-z]+) --" nil t))
 326     'clari-briefs)
 327    (t
 328     'digest)))
 329
 330 (defun nndoc-set-delims ()
 331   "Set the nndoc delimiter variables according to the type of the document."
 332   (let ((vars '(nndoc-file-begin
 333                 nndoc-first-article
 334                 nndoc-article-end nndoc-head-begin nndoc-head-end
 335                 nndoc-file-end nndoc-article-begin
 336                 nndoc-body-begin nndoc-body-end-function nndoc-body-end
 337                 nndoc-prepare-body nndoc-article-transform
 338                 nndoc-generate-head nndoc-body-begin-function
 339                 nndoc-head-begin-function nndoc-article-begin-function)))
 340     (while vars
 341       (set (pop vars) nil)))
 342   (let* (defs guess)
 343     ;; Guess away until we find the real file type.
 344     (while (setq defs (cdr (assq nndoc-article-type nndoc-type-alist))
 345                  guess (assq 'guess defs))
 346       (setq nndoc-article-type (funcall (cdr guess))))
 347     ;; Set the nndoc variables.
 348     (while defs
 349       (set (intern (format "nndoc-%s" (caar defs)))
 350            (cdr (pop defs))))))
 351
 352 (defun nndoc-search (regexp)
 353   (prog1
 354       (re-search-forward regexp nil t)
 355     (beginning-of-line)))
 356
 357 (defun nndoc-dissect-buffer ()
 358   "Go through the document and partition it into heads/bodies/articles."
 359   (let ((i 0)
 360         (first t)
 361         head-begin head-end body-begin body-end)
 362     (setq nndoc-dissection-alist nil)
 363     (save-excursion
 364       (set-buffer nndoc-current-buffer)
 365       (goto-char (point-min))
 366       ;; Find the beginning of the file.
 367       (when nndoc-file-begin
 368         (nndoc-search nndoc-file-begin))
 369       ;; Go through the file.
 370       (while (if (and first nndoc-first-article)
 371                  (nndoc-search nndoc-first-article)
 372                (if nndoc-article-begin-function
 373                    (funcall nndoc-article-begin-function)
 374                  (nndoc-search nndoc-article-begin)))
 375         (setq first nil)
 376         (cond (nndoc-head-begin-function
 377                (funcall nndoc-head-begin-function))
 378               (nndoc-head-begin
 379                (nndoc-search nndoc-head-begin)))
 380         (if (and nndoc-file-end
 381                  (looking-at nndoc-file-end))
 382             (goto-char (point-max))
 383           (setq head-begin (point))
 384           (nndoc-search (or nndoc-head-end "^$"))
 385           (setq head-end (point))
 386           (if nndoc-body-begin-function
 387               (funcall nndoc-body-begin-function)
 388             (nndoc-search (or nndoc-body-begin "^\n")))
 389           (setq body-begin (point))
 390           (or (and nndoc-body-end-function
 391                    (funcall nndoc-body-end-function))
 392               (and nndoc-body-end
 393                    (nndoc-search nndoc-body-end))
 394               (if nndoc-article-begin-function
 395                   (funcall nndoc-article-begin-function)
 396                 (nndoc-search nndoc-article-begin))
 397               (progn
 398                 (goto-char (point-max))
 399                 (when nndoc-file-end
 400                   (and (re-search-backward nndoc-file-end nil t)
 401                        (beginning-of-line)))))
 402           (setq body-end (point))
 403           (push (list (incf i) head-begin head-end body-begin body-end
 404                       (count-lines body-begin body-end))
 405                 nndoc-dissection-alist))))))
 406
 407 (defun nndoc-unquote-dashes ()
 408   "Unquote quoted non-separators in digests."
 409   (while (re-search-forward "^- -"nil t)
 410     (replace-match "-" t t)))
 411
 412 (defun nndoc-digest-body-end ()
 413   (and (re-search-forward nndoc-article-begin nil t)
 414        (goto-char (match-beginning 0))))
 415
 416 (defun nndoc-mbox-article-begin ()
 417   (when (re-search-forward nndoc-article-begin nil t)
 418     (goto-char (match-beginning 0))))
 419
 420 (defun nndoc-mbox-body-end ()
 421   (let ((beg (point))
 422         len end)
 423     (when
 424         (save-excursion
 425           (and (re-search-backward nndoc-article-begin nil t)
 426                (setq end (point))
 427                (search-forward "\n\n" beg t)
 428                (re-search-backward
 429                 "^Content-Length:[ \t]*\\([0-9]+\\) *$" end t)
 430                (setq len (string-to-int (match-string 1)))
 431                (search-forward "\n\n" beg t)
 432                (or (= (setq len (+ (point) len)) (point-max))
 433                    (and (< len (point-max))
 434                         (goto-char len)
 435                         (looking-at nndoc-article-begin)))))
 436       (goto-char len))))
 437
 438 (defun nndoc-rnews-body-end ()
 439   (and (re-search-backward nndoc-article-begin nil t)
 440        (forward-line 1)
 441        (goto-char (+ (point) (string-to-int (match-string 1))))))
 442
 443 (defun nndoc-transform-clari-briefs (article)
 444   (goto-char (point-min))
 445   (when (looking-at " *\\*\\(.*\\)\n")
 446     (replace-match "" t t))
 447   (nndoc-generate-clari-briefs-head article))
 448
 449 (defun nndoc-generate-clari-briefs-head (article)
 450   (let ((entry (cdr (assq article nndoc-dissection-alist)))
 451         subject from)
 452     (save-excursion
 453       (set-buffer nndoc-current-buffer)
 454       (save-restriction
 455         (narrow-to-region (car entry) (nth 3 entry))
 456         (goto-char (point-min))
 457         (when (looking-at " *\\*\\(.*\\)$")
 458           (setq subject (match-string 1))
 459           (when (string-match "[ \t]+$" subject)
 460             (setq subject (substring subject 0 (match-beginning 0)))))
 461         (when
 462             (let ((case-fold-search nil))
 463               (re-search-forward
 464                "^\t\\([^a-z]+\\(,[^(]+\\)? ([^a-z]+)\\) --" nil t))
 465           (setq from (match-string 1)))))
 466     (insert "From: " "clari@clari.net (" (or from "unknown") ")"
 467             "\nSubject: " (or subject "(no subject)") "\n")))
 468
 469 (defun nndoc-babyl-body-begin ()
 470   (re-search-forward "^\n" nil t)
 471   (when (looking-at "\*\*\* EOOH \*\*\*")
 472     (re-search-forward "^\n" nil t)))
 473
 474 (defun nndoc-babyl-head-begin ()
 475   (when (re-search-forward "^[0-9].*\n" nil t)
 476     (when (looking-at "\*\*\* EOOH \*\*\*")
 477       (forward-line 1))
 478     t))
 479
 480 (provide 'nndoc)
 481
 482 ;;; nndoc.el ends here