lisp/gnus/nnslashdot.el

   1 ;;; nnslashdot.el --- interfacing with Slashdot
   2
   3 ;; Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004,
   4 ;;   2005 Free Software Foundation, Inc.
   5
   6 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
   7 ;; Keywords: news
   8
   9 ;; This file is part of GNU Emacs.
  10
  11 ;; GNU Emacs is free software; you can redistribute it and/or modify
  12 ;; it under the terms of the GNU General Public License as published by
  13 ;; the Free Software Foundation; either version 2, or (at your option)
  14 ;; any later version.
  15
  16 ;; GNU Emacs is distributed in the hope that it will be useful,
  17 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  18 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  19 ;; GNU General Public License for more details.
  20
  21 ;; You should have received a copy of the GNU General Public License
  22 ;; along with GNU Emacs; see the file COPYING.  If not, write to the
  23 ;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  24 ;; Boston, MA 02110-1301, USA.
  25
  26 ;;; Commentary:
  27
  28 ;;; Code:
  29
  30 (eval-when-compile (require 'cl))
  31
  32 (require 'nnoo)
  33 (require 'message)
  34 (require 'gnus-util)
  35 (require 'gnus)
  36 (require 'nnmail)
  37 (require 'mm-util)
  38 (require 'mm-url)
  39
  40 (nnoo-declare nnslashdot)
  41
  42 (defvoo nnslashdot-directory (nnheader-concat gnus-directory "slashdot/")
  43   "Where nnslashdot will save its files.")
  44
  45 (defvoo nnslashdot-active-url "http://slashdot.org/search.pl?section=&min=%d"
  46   "Where nnslashdot will fetch the active file from.")
  47
  48 (defvoo nnslashdot-comments-url "http://slashdot.org/comments.pl?sid=%s&threshold=%d&commentsort=%d&mode=flat&startat=%d"
  49   "Where nnslashdot will fetch comments from.")
  50
  51 (defvoo nnslashdot-article-url
  52     "http://slashdot.org/article.pl?sid=%s&mode=nocomment"
  53   "Where nnslashdot will fetch the article from.")
  54
  55 (defvoo nnslashdot-backslash-url "http://slashdot.org/slashdot.xml"
  56   "Where nnslashdot will fetch the stories from.")
  57
  58 (defvoo nnslashdot-use-front-page nil
  59   "Use the front page in addition to the backslash page.")
  60
  61 (defvoo nnslashdot-threshold -1
  62   "The article threshold.")
  63
  64 (defvoo nnslashdot-threaded t
  65   "Whether the nnslashdot groups should be threaded or not.")
  66
  67 (defvoo nnslashdot-group-number 0
  68   "The number of non-fresh groups to keep updated.")
  69
  70 (defvoo nnslashdot-login-name ""
  71   "The login name to use when posting.")
  72
  73 (defvoo nnslashdot-password ""
  74   "The password to use when posting.")
  75
  76 ;;; Internal variables
  77
  78 (defvar nnslashdot-groups nil)
  79 (defvar nnslashdot-buffer nil)
  80 (defvar nnslashdot-headers nil)
  81
  82 ;;; Interface functions
  83
  84 (nnoo-define-basics nnslashdot)
  85
  86 (deffoo nnslashdot-retrieve-headers (articles &optional group server fetch-old)
  87   (nnslashdot-possibly-change-server group server)
  88   (condition-case why
  89       (unless gnus-nov-is-evil
  90         (nnslashdot-retrieve-headers-1 articles group))
  91     (search-failed (nnslashdot-lose why))))
  92
  93 (deffoo nnslashdot-retrieve-headers-1 (articles group)
  94   (let* ((last (car (last articles)))
  95          (start (if nnslashdot-threaded 1 (pop articles)))
  96          (entry (assoc group nnslashdot-groups))
  97          (sid (nth 2 entry))
  98          (first-comments t)
  99          headers article subject score from date lines parent point cid
 100          s startats changed)
 101     (save-excursion
 102       (set-buffer nnslashdot-buffer)
 103       (let ((case-fold-search t))
 104         (erase-buffer)
 105         (when (= start 1)
 106           (mm-url-insert (format nnslashdot-article-url sid) t)
 107           (goto-char (point-min))
 108           (if (eobp)
 109               (error "Couldn't open connection to slashdot"))
 110           (re-search-forward "Posted by[ \t\r\n]+")
 111           (when (looking-at "\\(<a[^>]+>\\)?[ \t\r\n]*\\([^<\r\n]+\\)")
 112             (setq from (mm-url-decode-entities-string (match-string 2))))
 113           (search-forward "on ")
 114           (setq date (nnslashdot-date-to-date
 115                       (buffer-substring (point) (1- (search-forward "<")))))
 116           (setq lines (/ (- (point)
 117                             (progn (forward-line 1) (point)))
 118                          60))
 119           (push
 120            (cons
 121             1
 122             (make-full-mail-header
 123              1 group from date
 124              (concat "<" sid "%1@slashdot>")
 125              "" 0 lines nil nil))
 126            headers)
 127           (setq start (if nnslashdot-threaded 2 (pop articles))))
 128         (while (and start (<= start last))
 129           (setq point (goto-char (point-max)))
 130           (mm-url-insert
 131            (format nnslashdot-comments-url sid
 132                    nnslashdot-threshold 0 (- start 2))
 133            t)
 134           (when (and nnslashdot-threaded first-comments)
 135             (setq first-comments nil)
 136             (goto-char (point-max))
 137             (while (re-search-backward "startat=\\([0-9]+\\)" nil t)
 138               (setq s (string-to-number (match-string 1)))
 139               (unless (memq s startats)
 140                 (push s startats)))
 141             (setq startats (sort startats '<)))
 142           (setq article (if (and article (< start article)) article start))
 143           (goto-char point)
 144           (while (re-search-forward
 145                   "<a name=\"\\([0-9]+\\)\"><\\(b\\|H4\\)>\\([^<]+\\)</\\(b\\|H4\\)>.*score:\\([^)]+\\))"
 146                   nil t)
 147             (setq cid (match-string 1)
 148                   subject (match-string 3)
 149                   score (match-string 5))
 150             (unless (assq article (nth 4 entry))
 151               (setcar (nthcdr 4 entry) (cons (cons article cid) (nth 4 entry)))
 152               (setq changed t))
 153             (when (string-match "^Re: *" subject)
 154               (setq subject (concat "Re: " (substring subject (match-end 0)))))
 155             (setq subject (mm-url-decode-entities-string subject))
 156             (search-forward "<BR>")
 157             (cond
 158              ((looking-at
 159                "by[ \t\n]+<a[^>]+>\\([^<]+\\)</a>[ \t\n]*(\\(<[^>]+>\\)*\\([^<>)]+\\))")
 160               (goto-char (- (match-end 0) 5))
 161               (setq from (concat
 162                           (mm-url-decode-entities-string (match-string 1))
 163                           " <" (match-string 3) ">")))
 164              ((looking-at "by[ \t\n]+<a[^>]+>\\([^<(]+\\) (\\([0-9]+\\))</a>")
 165               (goto-char (- (match-end 0) 5))
 166               (setq from (concat
 167                           (mm-url-decode-entities-string (match-string 1))
 168                           " <" (match-string 2) ">")))
 169              ((looking-at "by \\([^<>]*\\)[\t\n\r ]+on ")
 170               (goto-char (- (match-end 0) 5))
 171               (setq from (mm-url-decode-entities-string (match-string 1))))
 172              (t
 173               (setq from "")))
 174             (search-forward "on ")
 175             (setq date
 176                   (nnslashdot-date-to-date
 177                    (buffer-substring (point) (progn (skip-chars-forward "^()<>\n\r") (point)))))
 178             (setq lines (/ (abs (- (search-forward "<td")
 179                                    (search-forward "</td>")))
 180                            70))
 181             (if (not
 182                  (re-search-forward ".*cid=\\([0-9]+\\)\">Parent</A>" nil t))
 183                 (setq parent nil)
 184               (setq parent (match-string 1))
 185               (when (string= parent "0")
 186                 (setq parent nil)))
 187             (push
 188              (cons
 189               article
 190               (make-full-mail-header
 191                article
 192                (concat subject " (" score ")")
 193                from date
 194                (concat "<" sid "%" cid "@slashdot>")
 195                (if parent
 196                    (concat "<" sid "%" parent "@slashdot>")
 197                  "")
 198                0 lines nil nil))
 199              headers)
 200             (while (and articles (<= (car articles) article))
 201               (pop articles))
 202             (setq article (1+ article)))
 203           (if nnslashdot-threaded
 204               (progn
 205                 (setq start (pop startats))
 206                 (if start (setq start (+ start 2))))
 207             (setq start (pop articles))))))
 208     (if changed (nnslashdot-write-groups))
 209     (setq nnslashdot-headers (sort headers 'car-less-than-car))
 210     (save-excursion
 211       (set-buffer nntp-server-buffer)
 212       (erase-buffer)
 213       (mm-with-unibyte-current-buffer
 214        (dolist (header nnslashdot-headers)
 215          (nnheader-insert-nov (cdr header)))))
 216     'nov))
 217
 218 (deffoo nnslashdot-request-group (group &optional server dont-check)
 219   (nnslashdot-possibly-change-server nil server)
 220   (let ((elem (assoc group nnslashdot-groups)))
 221     (cond
 222      ((not elem)
 223       (nnheader-report 'nnslashdot "Group does not exist"))
 224      (t
 225       (nnheader-report 'nnslashdot "Opened group %s" group)
 226       (nnheader-insert
 227        "211 %d %d %d %s\n" (cadr elem) 1 (cadr elem)
 228        (prin1-to-string group))))))
 229
 230 (deffoo nnslashdot-close-group (group &optional server)
 231   (nnslashdot-possibly-change-server group server)
 232   (when (gnus-buffer-live-p nnslashdot-buffer)
 233     (save-excursion
 234       (set-buffer nnslashdot-buffer)
 235       (kill-buffer nnslashdot-buffer)))
 236   t)
 237
 238 (deffoo nnslashdot-request-article (article &optional group server buffer)
 239   (nnslashdot-possibly-change-server group server)
 240   (let (contents cid)
 241     (condition-case why
 242         (save-excursion
 243           (set-buffer nnslashdot-buffer)
 244           (let ((case-fold-search t))
 245             (goto-char (point-min))
 246             (when (and (stringp article)
 247                        (string-match "%\\([0-9]+\\)@" article))
 248               (setq cid (match-string 1 article))
 249               (let ((map (nth 4 (assoc group nnslashdot-groups))))
 250                 (while map
 251                   (if (equal (cdar map) cid)
 252                       (setq article (caar map)
 253                             map nil)
 254                     (setq map (cdr map))))))
 255             (when (numberp article)
 256               (if (= article 1)
 257                   (progn
 258                     (re-search-forward
 259                      "Posted by")
 260                     (search-forward "<BR>")
 261                     (setq contents
 262                           (buffer-substring
 263                            (point)
 264                            (progn
 265                              (re-search-forward
 266                               "<IFRAME\\|<SCRIPT LANGUAGE=\"JAVASCRIPT\">\\|<!-- no ad 6 -->\\|&lt;&nbsp;[ \t\r\n]*<A HREF=\"\\(\\(http:\\)?//slashdot\\.org\\)?/article")
 267                              (match-beginning 0)))))
 268                 (setq cid (cdr (assq article
 269                                      (nth 4 (assoc group nnslashdot-groups)))))
 270                 (search-forward (format "<a name=\"%s\">" cid))
 271                 (setq contents
 272                       (buffer-substring
 273                        (re-search-forward "<td[^>]*>")
 274                        (search-forward "</td>")))))))
 275       (search-failed (nnslashdot-lose why)))
 276
 277     (when contents
 278       (save-excursion
 279         (set-buffer (or buffer nntp-server-buffer))
 280         (erase-buffer)
 281         (mm-with-unibyte-current-buffer
 282           (insert contents)
 283           (goto-char (point-min))
 284           (while (re-search-forward "\\(<br>\r?\\)+" nil t)
 285             (replace-match "<p>" t t))
 286           (goto-char (point-min))
 287           (insert "Content-Type: text/html\nMIME-Version: 1.0\n")
 288           (insert "Newsgroups: " (caddr (assoc group nnslashdot-groups))
 289                   "\n")
 290           (let ((header (cdr (assq article nnslashdot-headers))))
 291             (nnheader-insert-header header))
 292           (nnheader-report 'nnslashdot "Fetched article %s" article))
 293         (cons group article)))))
 294
 295 (deffoo nnslashdot-close-server (&optional server)
 296   (when (and (nnslashdot-server-opened server)
 297              (gnus-buffer-live-p nnslashdot-buffer))
 298     (save-excursion
 299       (set-buffer nnslashdot-buffer)
 300       (kill-buffer nnslashdot-buffer)))
 301   (nnoo-close-server 'nnslashdot server))
 302
 303 (deffoo nnslashdot-request-list (&optional server)
 304   (nnslashdot-possibly-change-server nil server)
 305   (let ((number 0)
 306         (first nnslashdot-use-front-page)
 307         sid elem description articles gname)
 308     (condition-case why
 309         ;; First we do the Ultramode to get info on all the latest groups.
 310         (progn
 311           (mm-with-unibyte-buffer
 312             (mm-url-insert nnslashdot-backslash-url t)
 313             (goto-char (point-min))
 314             (if (eobp)
 315                 (error "Couldn't open connection to slashdot"))
 316             (while (search-forward "<story>" nil t)
 317               (narrow-to-region (point) (search-forward "</story>"))
 318               (goto-char (point-min))
 319               (re-search-forward "<title>\\([^<]+\\)</title>")
 320               (setq description
 321                     (mm-url-decode-entities-string (match-string 1)))
 322               (re-search-forward "<url>\\([^<]+\\)</url>")
 323               (setq sid (match-string 1))
 324               (string-match "sid=\\([0-9/]+\\)\\(.shtml\\|$\\)" sid)
 325               (setq sid (match-string 1 sid))
 326               (re-search-forward "<comments>\\([^<]+\\)</comments>")
 327               (setq articles (string-to-number (match-string 1)))
 328               (setq gname (concat description " (" sid ")"))
 329               (if (setq elem (assoc gname nnslashdot-groups))
 330                   (setcar (cdr elem) articles)
 331                 (push (list gname articles sid (current-time) nil)
 332                       nnslashdot-groups))
 333               (goto-char (point-max))
 334               (widen)))
 335           ;; Then do the older groups.
 336           (while (or first
 337                      (> (- nnslashdot-group-number number) 0))
 338             (setq first nil)
 339             (mm-with-unibyte-buffer
 340               (let ((case-fold-search t))
 341                 (mm-url-insert (format nnslashdot-active-url number) t)
 342                 (goto-char (point-min))
 343                 (while (re-search-forward
 344                         "article.pl\\?sid=\\([^&]+\\).*>\\([^<]+\\)</a>"
 345                         nil t)
 346                   (setq sid (match-string 1)
 347                         description
 348                         (mm-url-decode-entities-string (match-string 2)))
 349                   (forward-line 1)
 350                   (when (re-search-forward "with \\([0-9]+\\) comment" nil t)
 351                     (setq articles (1+ (string-to-number (match-string 1)))))
 352                   (setq gname (concat description " (" sid ")"))
 353                   (if (setq elem (assoc gname nnslashdot-groups))
 354                       (setcar (cdr elem) articles)
 355                     (push (list gname articles sid (current-time) nil)
 356                           nnslashdot-groups)))))
 357             (incf number 30)))
 358       (search-failed (nnslashdot-lose why)))
 359     (nnslashdot-write-groups)
 360     (nnslashdot-generate-active)
 361     t))
 362
 363 (deffoo nnslashdot-request-newgroups (date &optional server)
 364   (nnslashdot-possibly-change-server nil server)
 365   (nnslashdot-generate-active)
 366   t)
 367
 368 (deffoo nnslashdot-request-post (&optional server)
 369   (nnslashdot-possibly-change-server nil server)
 370   (let ((sid (message-fetch-field "newsgroups"))
 371         (subject (message-fetch-field "subject"))
 372         (references (car (last (split-string
 373                                 (message-fetch-field "references")))))
 374         body quoted pid)
 375     (string-match "%\\([0-9]+\\)@slashdot" references)
 376     (setq pid (match-string 1 references))
 377     (message-goto-body)
 378     (narrow-to-region (point) (progn (message-goto-signature) (point)))
 379     (goto-char (point-min))
 380     (while (not (eobp))
 381       (if (looking-at "> ")
 382           (progn
 383             (delete-region (point) (+ (point) 2))
 384             (unless quoted
 385               (insert "<blockquote>\n"))
 386             (setq quoted t))
 387         (when quoted
 388           (insert "</blockquote>\n")
 389           (setq quoted nil)))
 390       (forward-line 1))
 391     (goto-char (point-min))
 392     (while (re-search-forward "^ *\n" nil t)
 393       (replace-match "<p>\n"))
 394     (widen)
 395     (when (message-goto-signature)
 396       (forward-line -1)
 397       (insert "<p>\n")
 398       (while (not (eobp))
 399         (end-of-line)
 400         (insert "<br>")
 401         (forward-line 1)))
 402     (message-goto-body)
 403     (setq body (buffer-substring (point) (point-max)))
 404     (erase-buffer)
 405     (mm-url-fetch-form
 406      "http://slashdot.org/comments.pl"
 407      `(("sid" . ,sid)
 408        ("pid" . ,pid)
 409        ("rlogin" . "userlogin")
 410        ("unickname" . ,nnslashdot-login-name)
 411        ("upasswd" . ,nnslashdot-password)
 412        ("postersubj" . ,subject)
 413        ("op" . "Submit")
 414        ("postercomment" . ,body)
 415        ("posttype" . "html")))))
 416
 417 (deffoo nnslashdot-request-delete-group (group &optional force server)
 418   (nnslashdot-possibly-change-server group server)
 419   (setq nnslashdot-groups (delq (assoc group nnslashdot-groups)
 420                                 nnslashdot-groups))
 421   (nnslashdot-write-groups))
 422
 423 (deffoo nnslashdot-request-close ()
 424   (setq nnslashdot-headers nil
 425         nnslashdot-groups nil))
 426
 427 (deffoo nnslashdot-request-expire-articles
 428     (articles group &optional server force)
 429   (nnslashdot-possibly-change-server group server)
 430   (let ((item (assoc group nnslashdot-groups)))
 431     (when item
 432       (if (fourth item)
 433           (when (and (>= (length articles) (cadr item)) ;; All are expirable.
 434                      (nnmail-expired-article-p
 435                       group
 436                       (fourth item)
 437                       force))
 438             (setq nnslashdot-groups (delq item nnslashdot-groups))
 439             (nnslashdot-write-groups)
 440             (setq articles nil)) ;; all expired.
 441         (setcdr (cddr item) (list (current-time)))
 442         (nnslashdot-write-groups))))
 443   articles)
 444
 445 (nnoo-define-skeleton nnslashdot)
 446
 447 ;;; Internal functions
 448
 449 (defun nnslashdot-possibly-change-server (&optional group server)
 450   (nnslashdot-init server)
 451   (when (and server
 452              (not (nnslashdot-server-opened server)))
 453     (nnslashdot-open-server server))
 454   (unless nnslashdot-groups
 455     (nnslashdot-read-groups)))
 456
 457 (defun nnslashdot-make-tuple (tuple n)
 458   (prog1
 459       tuple
 460     (while (> n 1)
 461       (unless (cdr tuple)
 462         (setcdr tuple (list nil)))
 463       (setq tuple (cdr tuple)
 464             n (1- n)))))
 465
 466 (defun nnslashdot-read-groups ()
 467   (let ((file (expand-file-name "groups" nnslashdot-directory)))
 468     (when (file-exists-p file)
 469       (mm-with-unibyte-buffer
 470         (insert-file-contents file)
 471         (goto-char (point-min))
 472         (setq nnslashdot-groups (read (current-buffer))))
 473       (if (and nnslashdot-groups (< (length (car nnslashdot-groups)) 5))
 474           (let ((groups nnslashdot-groups))
 475             (while groups
 476               (nnslashdot-make-tuple (car groups) 5)
 477               (setq groups (cdr groups))))))))
 478
 479 (defun nnslashdot-write-groups ()
 480   (with-temp-file (expand-file-name "groups" nnslashdot-directory)
 481     (gnus-prin1 nnslashdot-groups)))
 482
 483 (defun nnslashdot-init (server)
 484   "Initialize buffers and such."
 485   (unless (file-exists-p nnslashdot-directory)
 486     (gnus-make-directory nnslashdot-directory))
 487   (unless (gnus-buffer-live-p nnslashdot-buffer)
 488     (setq nnslashdot-buffer
 489           (save-excursion
 490             (nnheader-set-temp-buffer
 491              (format " *nnslashdot %s*" server))))
 492     (push nnslashdot-buffer gnus-buffers)))
 493
 494 (defun nnslashdot-date-to-date (sdate)
 495   (condition-case err
 496       (let ((elem (delete "" (split-string sdate))))
 497         (concat (substring (nth 0 elem) 0 3) " "
 498                 (substring (nth 1 elem) 0 3) " "
 499                 (substring (nth 2 elem) 0 2) " "
 500                 (substring (nth 3 elem) 1 6) " "
 501                 (format-time-string "%Y") " "
 502                 (nth 4 elem)))
 503     (error "")))
 504
 505 (defun nnslashdot-generate-active ()
 506   (save-excursion
 507     (set-buffer nntp-server-buffer)
 508     (erase-buffer)
 509     (dolist (elem nnslashdot-groups)
 510       (when (numberp (cadr elem))
 511         (insert (prin1-to-string (car elem))
 512                 " " (number-to-string (cadr elem)) " 1 y\n")))))
 513
 514 (defun nnslashdot-lose (why)
 515   (error "Slashdot HTML has changed; please get a new version of nnslashdot"))
 516
 517 (provide 'nnslashdot)
 518
 519 ;;; arch-tag: aa73df7a-f7e6-4eef-bdea-5ce2f8c691b3
 520 ;;; nnslashdot.el ends here