lisp/gnus/nnslashdot.el

   1 ;;; nnslashdot.el --- interfacing with Slashdot
   2 ;; Copyright (C) 1999, 2000, 2001 Free Software Foundation, Inc.
   3
   4 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
   5 ;; Keywords: news
   6
   7 ;; This file is part of GNU Emacs.
   8
   9 ;; GNU Emacs is free software; you can redistribute it and/or modify
  10 ;; it under the terms of the GNU General Public License as published by
  11 ;; the Free Software Foundation; either version 2, or (at your option)
  12 ;; any later version.
  13
  14 ;; GNU Emacs is distributed in the hope that it will be useful,
  15 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  16 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17 ;; GNU General Public License for more details.
  18
  19 ;; You should have received a copy of the GNU General Public License
  20 ;; along with GNU Emacs; see the file COPYING.  If not, write to the
  21 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  22 ;; Boston, MA 02111-1307, USA.
  23
  24 ;;; Commentary:
  25
  26 ;; Note: You need to have `url' and `w3' installed for this
  27 ;; backend to work.
  28
  29 ;;; Code:
  30
  31 (eval-when-compile (require 'cl))
  32
  33 (require 'nnoo)
  34 (require 'message)
  35 (require 'gnus-util)
  36 (require 'gnus)
  37 (require 'nnmail)
  38 (require 'mm-util)
  39 (eval-when-compile
  40   (ignore-errors
  41     (require 'nnweb)))
  42 ;; Report failure to find w3 at load time if appropriate.
  43 (eval '(require 'nnweb))
  44
  45 (nnoo-declare nnslashdot)
  46
  47 (defvoo nnslashdot-directory (nnheader-concat gnus-directory "slashdot/")
  48   "Where nnslashdot will save its files.")
  49
  50 (defvoo nnslashdot-active-url "http://slashdot.org/search.pl?section=&min=%d"
  51   "Where nnslashdot will fetch the active file from.")
  52
  53 (defvoo nnslashdot-comments-url "http://slashdot.org/comments.pl?sid=%s&threshold=%d&commentsort=%d&mode=flat&startat=%d"
  54   "Where nnslashdot will fetch comments from.")
  55
  56 (defvoo nnslashdot-article-url
  57     "http://slashdot.org/article.pl?sid=%s&mode=nocomment"
  58   "Where nnslashdot will fetch the article from.")
  59
  60 (defvoo nnslashdot-backslash-url "http://slashdot.org/slashdot.xml"
  61   "Where nnslashdot will fetch the stories from.")
  62
  63 (defvoo nnslashdot-threshold -1
  64   "The article threshold.")
  65
  66 (defvoo nnslashdot-threaded t
  67   "Whether the nnslashdot groups should be threaded or not.")
  68
  69 (defvoo nnslashdot-group-number 0
  70   "The number of non-fresh groups to keep updated.")
  71
  72 (defvoo nnslashdot-login-name ""
  73   "The login name to use when posting.")
  74
  75 (defvoo nnslashdot-password ""
  76   "The password to use when posting.")
  77
  78 ;;; Internal variables
  79
  80 (defvar nnslashdot-groups nil)
  81 (defvar nnslashdot-buffer nil)
  82 (defvar nnslashdot-headers nil)
  83
  84 ;;; Interface functions
  85
  86 (nnoo-define-basics nnslashdot)
  87
  88 (deffoo nnslashdot-retrieve-headers (articles &optional group server fetch-old)
  89   (nnslashdot-possibly-change-server group server)
  90   (condition-case why
  91       (unless gnus-nov-is-evil
  92         (nnslashdot-retrieve-headers-1 articles group))
  93     (search-failed (nnslashdot-lose why))))
  94
  95 (deffoo nnslashdot-retrieve-headers-1 (articles group)
  96   (let* ((last (car (last articles)))
  97          (start (if nnslashdot-threaded 1 (pop articles)))
  98          (entry (assoc group nnslashdot-groups))
  99          (sid (nth 2 entry))
 100          (first-comments t)
 101          headers article subject score from date lines parent point cid
 102          s startats changed)
 103     (save-excursion
 104       (set-buffer nnslashdot-buffer)
 105       (let ((case-fold-search t))
 106         (erase-buffer)
 107         (when (= start 1)
 108           (nnweb-insert (format nnslashdot-article-url
 109                                 (nnslashdot-sid-strip sid)) t)
 110           (goto-char (point-min))
 111           (re-search-forward "Posted by[ \t\r\n]+")
 112           (when (looking-at "\\(<a[^>]+>\\)?[ \t\r\n]*\\([^<\r\n]+\\)")
 113             (setq from (nnweb-decode-entities-string (match-string 2))))
 114           (search-forward "on ")
 115           (setq date (nnslashdot-date-to-date
 116                       (buffer-substring (point) (1- (search-forward "<")))))
 117           (setq lines (/ (- (point)
 118                             (progn (forward-line 1) (point)))
 119                          60))
 120           (push
 121            (cons
 122             1
 123             (make-full-mail-header
 124              1 group from date
 125              (concat "<" (nnslashdot-sid-strip sid) "%1@slashdot>")
 126              "" 0 lines nil nil))
 127            headers)
 128           (setq start (if nnslashdot-threaded 2 (pop articles))))
 129         (while (and start (<= start last))
 130           (setq point (goto-char (point-max)))
 131           (nnweb-insert
 132            (format nnslashdot-comments-url
 133                    (nnslashdot-sid-strip sid)
 134                    nnslashdot-threshold 0 (- start 2))
 135            t)
 136           (when (and nnslashdot-threaded first-comments)
 137             (setq first-comments nil)
 138             (goto-char (point-max))
 139             (while (re-search-backward "startat=\\([0-9]+\\)" nil t)
 140               (setq s (string-to-number (match-string 1)))
 141               (unless (memq s startats)
 142                 (push s startats)))
 143             (setq startats (sort startats '<)))
 144           (setq article (if (and article (< start article)) article start))
 145           (goto-char point)
 146           (while (re-search-forward
 147                   "<a name=\"\\([0-9]+\\)\"><\\(b\\|H4\\)>\\([^<]+\\)</\\(b\\|H4\\)>.*score:\\([^)]+\\))"
 148                   nil t)
 149             (setq cid (match-string 1)
 150                   subject (match-string 3)
 151                   score (match-string 5))
 152             (unless (assq article (nth 4 entry))
 153               (setcar (nthcdr 4 entry) (cons (cons article cid) (nth 4 entry)))
 154               (setq changed t))
 155             (when (string-match "^Re: *" subject)
 156               (setq subject (concat "Re: " (substring subject (match-end 0)))))
 157             (setq subject (nnweb-decode-entities-string subject))
 158             (search-forward "<BR>")
 159             (if (looking-at
 160                  "by[ \t\n]+<a[^>]+>\\([^<]+\\)</a>[ \t\n]*(\\(<[^>]+>\\)*\\([^<>)]+\\))")
 161                 (progn
 162                   (goto-char (- (match-end 0) 5))
 163                   (setq from (concat
 164                               (nnweb-decode-entities-string (match-string 1))
 165                               " <" (match-string 3) ">")))
 166               (setq from "")
 167               (when (looking-at "by \\([^<>]*\\) on ")
 168                 (goto-char (- (match-end 0) 5))
 169                 (setq from (nnweb-decode-entities-string (match-string 1)))))
 170             (search-forward " on ")
 171             (setq date
 172                   (nnslashdot-date-to-date
 173                    (buffer-substring (point) (progn (skip-chars-forward "^()<>\n\r") (point)))))
 174             (setq lines (/ (abs (- (search-forward "<td")
 175                                    (search-forward "</td>")))
 176                            70))
 177             (if (not
 178                  (re-search-forward ".*cid=\\([0-9]+\\)\">Parent</A>" nil t))
 179                 (setq parent nil)
 180               (setq parent (match-string 1))
 181               (when (string= parent "0")
 182                 (setq parent nil)))
 183             (push
 184              (cons
 185               article
 186               (make-full-mail-header
 187                article
 188                (concat subject " (" score ")")
 189                from date
 190                (concat "<" (nnslashdot-sid-strip sid) "%" cid "@slashdot>")
 191                (if parent
 192                    (concat "<" (nnslashdot-sid-strip sid) "%"
 193                            parent "@slashdot>")
 194                  "")
 195                0 lines nil nil))
 196              headers)
 197             (while (and articles (<= (car articles) article))
 198               (pop articles))
 199             (setq article (1+ article)))
 200           (if nnslashdot-threaded
 201               (progn
 202                 (setq start (pop startats))
 203                 (if start (setq start (+ start 2))))
 204             (setq start (pop articles))))))
 205     (if changed (nnslashdot-write-groups))
 206     (setq nnslashdot-headers (sort headers 'car-less-than-car))
 207     (save-excursion
 208       (set-buffer nntp-server-buffer)
 209       (erase-buffer)
 210       (mm-with-unibyte-current-buffer
 211        (dolist (header nnslashdot-headers)
 212          (nnheader-insert-nov (cdr header)))))
 213     'nov))
 214
 215 (deffoo nnslashdot-request-group (group &optional server dont-check)
 216   (nnslashdot-possibly-change-server nil server)
 217   (let ((elem (assoc group nnslashdot-groups)))
 218     (cond
 219      ((not elem)
 220       (nnheader-report 'nnslashdot "Group does not exist"))
 221      (t
 222       (nnheader-report 'nnslashdot "Opened group %s" group)
 223       (nnheader-insert
 224        "211 %d %d %d %s\n" (cadr elem) 1 (cadr elem)
 225        (prin1-to-string group))))))
 226
 227 (deffoo nnslashdot-close-group (group &optional server)
 228   (nnslashdot-possibly-change-server group server)
 229   (when (gnus-buffer-live-p nnslashdot-buffer)
 230     (save-excursion
 231       (set-buffer nnslashdot-buffer)
 232       (kill-buffer nnslashdot-buffer)))
 233   t)
 234
 235 (deffoo nnslashdot-request-article (article &optional group server buffer)
 236   (nnslashdot-possibly-change-server group server)
 237   (let (contents cid)
 238     (condition-case why
 239         (save-excursion
 240           (set-buffer nnslashdot-buffer)
 241           (let ((case-fold-search t))
 242             (goto-char (point-min))
 243             (when (and (stringp article)
 244                        (string-match "%\\([0-9]+\\)@" article))
 245               (setq cid (match-string 1 article))
 246               (let ((map (nth 4 (assoc group nnslashdot-groups))))
 247                 (while map
 248                   (if (equal (cdar map) cid)
 249                       (setq article (caar map)
 250                             map nil)
 251                     (setq map (cdr map))))))
 252             (when (numberp article)
 253               (if (= article 1)
 254                   (progn
 255                     (re-search-forward
 256                      "Posted by")
 257                     (search-forward "<BR>")
 258                     (setq contents
 259                           (buffer-substring
 260                            (point)
 261                            (progn
 262                              (re-search-forward
 263                               "&lt;&nbsp;[ \t\r\n]*<A HREF=\"\\(\\(http:\\)?//slashdot\\.org\\)?/article")
 264                              (match-beginning 0)))))
 265                 (setq cid (cdr (assq article
 266                                      (nth 4 (assoc group nnslashdot-groups)))))
 267                 (search-forward (format "<a name=\"%s\">" cid))
 268                 (setq contents
 269                       (buffer-substring
 270                        (re-search-forward "<td[^>]*>")
 271                        (search-forward "</td>")))))))
 272       (search-failed (nnslashdot-lose why)))
 273
 274     (when contents
 275       (save-excursion
 276         (set-buffer (or buffer nntp-server-buffer))
 277         (erase-buffer)
 278         (mm-with-unibyte-current-buffer
 279           (insert contents)
 280           (goto-char (point-min))
 281           (while (re-search-forward "\\(<br>\r?\\)+" nil t)
 282             (replace-match "<p>" t t))
 283           (goto-char (point-min))
 284           (insert "Content-Type: text/html\nMIME-Version: 1.0\n")
 285           (insert "Newsgroups: " (caddr (assoc group nnslashdot-groups))
 286                   "\n")
 287           (let ((header (cdr (assq article nnslashdot-headers))))
 288             (nnheader-insert-header header))
 289           (nnheader-report 'nnslashdot "Fetched article %s" article))
 290         (cons group article)))))
 291
 292 (deffoo nnslashdot-close-server (&optional server)
 293   (when (and (nnslashdot-server-opened server)
 294              (gnus-buffer-live-p nnslashdot-buffer))
 295     (save-excursion
 296       (set-buffer nnslashdot-buffer)
 297       (kill-buffer nnslashdot-buffer)))
 298   (nnoo-close-server 'nnslashdot server))
 299
 300 (deffoo nnslashdot-request-list (&optional server)
 301   (nnslashdot-possibly-change-server nil server)
 302   (let ((number 0)
 303         sid elem description articles gname)
 304     (condition-case why
 305         ;; First we do the Ultramode to get info on all the latest groups.
 306         (progn
 307           (mm-with-unibyte-buffer
 308             (nnweb-insert nnslashdot-backslash-url t)
 309             (goto-char (point-min))
 310             (while (search-forward "<story>" nil t)
 311               (narrow-to-region (point) (search-forward "</story>"))
 312               (goto-char (point-min))
 313               (re-search-forward "<title>\\([^<]+\\)</title>")
 314               (setq description
 315                     (nnweb-decode-entities-string (match-string 1)))
 316               (re-search-forward "<url>\\([^<]+\\)</url>")
 317               (setq sid (match-string 1))
 318               (string-match "sid=\\([0-9/]+\\)\\(.shtml\\|$\\)" sid)
 319               (setq sid (match-string 1 sid))
 320               (re-search-forward "<comments>\\([^<]+\\)</comments>")
 321               (setq articles (string-to-number (match-string 1)))
 322               (setq gname (concat description " (" sid ")"))
 323               (if (setq elem (assoc gname nnslashdot-groups))
 324                   (setcar (cdr elem) articles)
 325                 (push (list gname articles sid (current-time) nil)
 326                       nnslashdot-groups))
 327               (goto-char (point-max))
 328               (widen)))
 329           ;; Then do the older groups.
 330           (while (> (- nnslashdot-group-number number) 0)
 331             (mm-with-unibyte-buffer
 332               (let ((case-fold-search t))
 333                 (nnweb-insert (format nnslashdot-active-url number) t)
 334                 (goto-char (point-min))
 335                 (while (re-search-forward
 336                         "article.pl\\?sid=\\([^&]+\\).*<b>\\([^<]+\\)</b>"
 337                         nil t)
 338                   (setq sid (match-string 1)
 339                         description
 340                         (nnweb-decode-entities-string (match-string 2)))
 341                   (forward-line 1)
 342                   (when (re-search-forward "<b>\\([0-9]+\\)</b>" nil t)
 343                     (setq articles (string-to-number (match-string 1))))
 344                   (setq gname (concat description " (" sid ")"))
 345                   (if (setq elem (assoc gname nnslashdot-groups))
 346                       (setcar (cdr elem) articles)
 347                     (push (list gname articles sid (current-time) nil)
 348                           nnslashdot-groups)))))
 349             (incf number 30)))
 350       (search-failed (nnslashdot-lose why)))
 351     (nnslashdot-write-groups)
 352     (nnslashdot-generate-active)
 353     t))
 354
 355 (deffoo nnslashdot-request-newgroups (date &optional server)
 356   (nnslashdot-possibly-change-server nil server)
 357   (nnslashdot-generate-active)
 358   t)
 359
 360 (deffoo nnslashdot-request-post (&optional server)
 361   (nnslashdot-possibly-change-server nil server)
 362   (let ((sid (nnslashdot-sid-strip (message-fetch-field "newsgroups")))
 363         (subject (message-fetch-field "subject"))
 364         (references (car (last (split-string
 365                                 (message-fetch-field "references")))))
 366         body quoted pid)
 367     (string-match "%\\([0-9]+\\)@slashdot" references)
 368     (setq pid (match-string 1 references))
 369     (message-goto-body)
 370     (narrow-to-region (point) (progn (message-goto-signature) (point)))
 371     (goto-char (point-min))
 372     (while (not (eobp))
 373       (if (looking-at "> ")
 374           (progn
 375             (delete-region (point) (+ (point) 2))
 376             (unless quoted
 377               (insert "<blockquote>\n"))
 378             (setq quoted t))
 379         (when quoted
 380           (insert "</blockquote>\n")
 381           (setq quoted nil)))
 382       (forward-line 1))
 383     (goto-char (point-min))
 384     (while (re-search-forward "^ *\n" nil t)
 385       (replace-match "<p>\n"))
 386     (widen)
 387     (when (message-goto-signature)
 388       (forward-line -1)
 389       (insert "<p>\n")
 390       (while (not (eobp))
 391         (end-of-line)
 392         (insert "<br>")
 393         (forward-line 1)))
 394     (message-goto-body)
 395     (setq body (buffer-substring (point) (point-max)))
 396     (erase-buffer)
 397     (nnweb-fetch-form
 398      "http://slashdot.org/comments.pl"
 399      `(("sid" . ,sid)
 400        ("pid" . ,pid)
 401        ("rlogin" . "userlogin")
 402        ("unickname" . ,nnslashdot-login-name)
 403        ("upasswd" . ,nnslashdot-password)
 404        ("postersubj" . ,subject)
 405        ("op" . "Submit")
 406        ("postercomment" . ,body)
 407        ("posttype" . "html")))))
 408
 409 (deffoo nnslashdot-request-delete-group (group &optional force server)
 410   (nnslashdot-possibly-change-server group server)
 411   (setq nnslashdot-groups (delq (assoc group nnslashdot-groups)
 412                                 nnslashdot-groups))
 413   (nnslashdot-write-groups))
 414
 415 (deffoo nnslashdot-request-close ()
 416   (setq nnslashdot-headers nil
 417         nnslashdot-groups nil))
 418
 419 (deffoo nnslashdot-request-expire-articles
 420     (articles group &optional server force)
 421   (nnslashdot-possibly-change-server group server)
 422   (let ((item (assoc group nnslashdot-groups)))
 423     (when item
 424       (if (fourth item)
 425           (when (and (>= (length articles) (cadr item)) ;; All are expirable.
 426                      (nnmail-expired-article-p
 427                       group
 428                       (fourth item)
 429                       force))
 430             (setq nnslashdot-groups (delq item nnslashdot-groups))
 431             (nnslashdot-write-groups)
 432             (setq articles nil)) ;; all expired.
 433         (setcdr (cddr item) (list (current-time)))
 434         (nnslashdot-write-groups))))
 435   articles)
 436
 437 (nnoo-define-skeleton nnslashdot)
 438
 439 ;;; Internal functions
 440
 441 (defun nnslashdot-possibly-change-server (&optional group server)
 442   (nnslashdot-init server)
 443   (when (and server
 444              (not (nnslashdot-server-opened server)))
 445     (nnslashdot-open-server server))
 446   (unless nnslashdot-groups
 447     (nnslashdot-read-groups)))
 448
 449 (defun nnslashdot-make-tuple (tuple n)
 450   (prog1
 451       tuple
 452     (while (> n 1)
 453       (unless (cdr tuple)
 454         (setcdr tuple (list nil)))
 455       (setq tuple (cdr tuple)
 456             n (1- n)))))
 457
 458 (defun nnslashdot-read-groups ()
 459   (let ((file (expand-file-name "groups" nnslashdot-directory)))
 460     (when (file-exists-p file)
 461       (mm-with-unibyte-buffer
 462         (insert-file-contents file)
 463         (goto-char (point-min))
 464         (setq nnslashdot-groups (read (current-buffer))))
 465       (if (and nnslashdot-groups (< (length (car nnslashdot-groups)) 5))
 466           (let ((groups nnslashdot-groups))
 467             (while groups
 468               (nnslashdot-make-tuple (car groups) 5)
 469               (setq groups (cdr groups))))))))
 470
 471 (defun nnslashdot-write-groups ()
 472   (with-temp-file (expand-file-name "groups" nnslashdot-directory)
 473     (gnus-prin1 nnslashdot-groups)))
 474
 475 (defun nnslashdot-init (server)
 476   "Initialize buffers and such."
 477   (unless (file-exists-p nnslashdot-directory)
 478     (gnus-make-directory nnslashdot-directory))
 479   (unless (gnus-buffer-live-p nnslashdot-buffer)
 480     (setq nnslashdot-buffer
 481           (save-excursion
 482             (nnheader-set-temp-buffer
 483              (format " *nnslashdot %s*" server))))
 484     (push nnslashdot-buffer gnus-buffers)))
 485
 486 (defun nnslashdot-date-to-date (sdate)
 487   (condition-case err
 488       (let ((elem (delete "" (split-string sdate))))
 489         (concat (substring (nth 0 elem) 0 3) " "
 490                 (substring (nth 1 elem) 0 3) " "
 491                 (substring (nth 2 elem) 0 2) " "
 492                 (substring (nth 3 elem) 1 6) " "
 493                 (format-time-string "%Y") " "
 494                 (nth 4 elem)))
 495     (error "")))
 496
 497 (defun nnslashdot-generate-active ()
 498   (save-excursion
 499     (set-buffer nntp-server-buffer)
 500     (erase-buffer)
 501     (dolist (elem nnslashdot-groups)
 502       (insert (prin1-to-string (car elem))
 503               " " (number-to-string (cadr elem)) " 1 y\n"))))
 504
 505 (defun nnslashdot-lose (why)
 506   (error "Slashdot HTML has changed; please get a new version of nnslashdot"))
 507
 508 (defalias 'nnslashdot-sid-strip 'identity)
 509
 510 (provide 'nnslashdot)
 511
 512 ;;; nnslashdot.el ends here