lisp/gnus/nnweb.el

   1 ;;; nnweb.el --- retrieving articles via web search engines
   2 ;; Copyright (C) 1996, 1997, 1998, 1999, 2000
   3 ;;        Free Software Foundation, Inc.
   4
   5 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
   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 ;; Note: You need to have `url' and `w3' installed for this
  28 ;; backend to work.
  29
  30 ;;; Code:
  31
  32 (eval-when-compile (require 'cl))
  33
  34 (require 'nnoo)
  35 (require 'message)
  36 (require 'gnus-util)
  37 (require 'gnus)
  38 (require 'nnmail)
  39 (require 'mm-util)
  40 (eval-when-compile
  41   (ignore-errors
  42     (require 'w3)
  43     (require 'url)
  44     (require 'w3-forms)))
  45
  46 ;; Report failure to find w3 at load time if appropriate.
  47 (unless noninteractive
  48   (eval '(progn
  49            (require 'w3)
  50            (require 'url)
  51            (require 'w3-forms))))
  52
  53 (nnoo-declare nnweb)
  54
  55 (defvoo nnweb-directory (nnheader-concat gnus-directory "nnweb/")
  56   "Where nnweb will save its files.")
  57
  58 (defvoo nnweb-type 'dejanews
  59   "What search engine type is being used.
  60 Valid types include `dejanews', `dejanewsold', `reference',
  61 and `altavista'.")
  62
  63 (defvar nnweb-type-definition
  64   '((dejanews
  65      (article . ignore)
  66      (id . "http://search.dejanews.com/msgid.xp?MID=%s&fmt=text")
  67      (map . nnweb-dejanews-create-mapping)
  68      (search . nnweb-dejanews-search)
  69      (address . "http://www.deja.com/=dnc/qs.xp")
  70      (identifier . nnweb-dejanews-identity))
  71     (dejanewsold
  72      (article . ignore)
  73      (map . nnweb-dejanews-create-mapping)
  74      (search . nnweb-dejanewsold-search)
  75      (address . "http://www.deja.com/dnquery.xp")
  76      (identifier . nnweb-dejanews-identity))
  77     (reference
  78      (article . nnweb-reference-wash-article)
  79      (map . nnweb-reference-create-mapping)
  80      (search . nnweb-reference-search)
  81      (address . "http://www.reference.com/cgi-bin/pn/go")
  82      (identifier . identity))
  83     (altavista
  84      (article . nnweb-altavista-wash-article)
  85      (map . nnweb-altavista-create-mapping)
  86      (search . nnweb-altavista-search)
  87      (address . "http://www.altavista.digital.com/cgi-bin/query")
  88      (id . "/cgi-bin/news?id@%s")
  89      (identifier . identity)))
  90   "Type-definition alist.")
  91
  92 (defvoo nnweb-search nil
  93   "Search string to feed to DejaNews.")
  94
  95 (defvoo nnweb-max-hits 999
  96   "Maximum number of hits to display.")
  97
  98 (defvoo nnweb-ephemeral-p nil
  99   "Whether this nnweb server is ephemeral.")
 100
 101 ;;; Internal variables
 102
 103 (defvoo nnweb-articles nil)
 104 (defvoo nnweb-buffer nil)
 105 (defvoo nnweb-group-alist nil)
 106 (defvoo nnweb-group nil)
 107 (defvoo nnweb-hashtb nil)
 108
 109 ;;; Interface functions
 110
 111 (nnoo-define-basics nnweb)
 112
 113 (deffoo nnweb-retrieve-headers (articles &optional group server fetch-old)
 114   (nnweb-possibly-change-server group server)
 115   (save-excursion
 116     (set-buffer nntp-server-buffer)
 117     (erase-buffer)
 118     (let (article header)
 119       (mm-with-unibyte-current-buffer
 120         (while (setq article (pop articles))
 121           (when (setq header (cadr (assq article nnweb-articles)))
 122             (nnheader-insert-nov header))))
 123       'nov)))
 124
 125 (deffoo nnweb-request-scan (&optional group server)
 126   (nnweb-possibly-change-server group server)
 127   (funcall (nnweb-definition 'map))
 128   (unless nnweb-ephemeral-p
 129     (nnweb-write-active)
 130     (nnweb-write-overview group)))
 131
 132 (deffoo nnweb-request-group (group &optional server dont-check)
 133   (nnweb-possibly-change-server nil server)
 134   (when (and group
 135              (not (equal group nnweb-group))
 136              (not nnweb-ephemeral-p))
 137     (let ((info (assoc group nnweb-group-alist)))
 138       (when info
 139         (setq nnweb-group group)
 140         (setq nnweb-type (nth 2 info))
 141         (setq nnweb-search (nth 3 info))
 142         (unless dont-check
 143           (nnweb-read-overview group)))))
 144   (cond
 145    ((not nnweb-articles)
 146     (nnheader-report 'nnweb "No matching articles"))
 147    (t
 148     (let ((active (if nnweb-ephemeral-p
 149                       (cons (caar nnweb-articles)
 150                             (caar (last nnweb-articles)))
 151                     (cadr (assoc group nnweb-group-alist)))))
 152       (nnheader-report 'nnweb "Opened group %s" group)
 153       (nnheader-insert
 154        "211 %d %d %d %s\n" (length nnweb-articles)
 155        (car active) (cdr active) group)))))
 156
 157 (deffoo nnweb-close-group (group &optional server)
 158   (nnweb-possibly-change-server group server)
 159   (when (gnus-buffer-live-p nnweb-buffer)
 160     (save-excursion
 161       (set-buffer nnweb-buffer)
 162       (set-buffer-modified-p nil)
 163       (kill-buffer nnweb-buffer)))
 164   t)
 165
 166 (deffoo nnweb-request-article (article &optional group server buffer)
 167   (nnweb-possibly-change-server group server)
 168   (save-excursion
 169     (set-buffer (or buffer nntp-server-buffer))
 170     (let* ((header (cadr (assq article nnweb-articles)))
 171            (url (and header (mail-header-xref header))))
 172       (when (or (and url
 173                      (mm-with-unibyte-current-buffer
 174                        (nnweb-fetch-url url)))
 175                 (and (stringp article)
 176                      (nnweb-definition 'id t)
 177                      (let ((fetch (nnweb-definition 'id))
 178                            art)
 179                        (when (string-match "^<\\(.*\\)>$" article)
 180                          (setq art (match-string 1 article)))
 181                        (and fetch
 182                             art
 183                             (mm-with-unibyte-current-buffer
 184                               (nnweb-fetch-url
 185                                (format fetch article)))))))
 186         (unless nnheader-callback-function
 187           (funcall (nnweb-definition 'article))
 188           (nnweb-decode-entities))
 189         (nnheader-report 'nnweb "Fetched article %s" article)
 190         (cons group (and (numberp article) article))))))
 191
 192 (deffoo nnweb-close-server (&optional server)
 193   (when (and (nnweb-server-opened server)
 194              (gnus-buffer-live-p nnweb-buffer))
 195     (save-excursion
 196       (set-buffer nnweb-buffer)
 197       (set-buffer-modified-p nil)
 198       (kill-buffer nnweb-buffer)))
 199   (nnoo-close-server 'nnweb server))
 200
 201 (deffoo nnweb-request-list (&optional server)
 202   (nnweb-possibly-change-server nil server)
 203   (save-excursion
 204     (set-buffer nntp-server-buffer)
 205     (nnmail-generate-active nnweb-group-alist)
 206     t))
 207
 208 (deffoo nnweb-request-update-info (group info &optional server)
 209   (nnweb-possibly-change-server group server))
 210
 211 (deffoo nnweb-asynchronous-p ()
 212   t)
 213
 214 (deffoo nnweb-request-create-group (group &optional server args)
 215   (nnweb-possibly-change-server nil server)
 216   (nnweb-request-delete-group group)
 217   (push `(,group ,(cons 1 0) ,@args) nnweb-group-alist)
 218   (nnweb-write-active)
 219   t)
 220
 221 (deffoo nnweb-request-delete-group (group &optional force server)
 222   (nnweb-possibly-change-server group server)
 223   (gnus-pull group nnweb-group-alist t)
 224   (nnweb-write-active)
 225   (gnus-delete-file (nnweb-overview-file group))
 226   t)
 227
 228 (nnoo-define-skeleton nnweb)
 229
 230 ;;; Internal functions
 231
 232 (defun nnweb-read-overview (group)
 233   "Read the overview of GROUP and build the map."
 234   (when (file-exists-p (nnweb-overview-file group))
 235     (mm-with-unibyte-buffer
 236       (nnheader-insert-file-contents (nnweb-overview-file group))
 237       (goto-char (point-min))
 238       (let (header)
 239         (while (not (eobp))
 240           (setq header (nnheader-parse-nov))
 241           (forward-line 1)
 242           (push (list (mail-header-number header)
 243                       header (mail-header-xref header))
 244                 nnweb-articles)
 245           (nnweb-set-hashtb header (car nnweb-articles)))))))
 246
 247 (defun nnweb-write-overview (group)
 248   "Write the overview file for GROUP."
 249   (with-temp-file (nnweb-overview-file group)
 250     (let ((articles nnweb-articles))
 251       (while articles
 252         (nnheader-insert-nov (cadr (pop articles)))))))
 253
 254 (defun nnweb-set-hashtb (header data)
 255   (gnus-sethash (nnweb-identifier (mail-header-xref header))
 256                 data nnweb-hashtb))
 257
 258 (defun nnweb-get-hashtb (url)
 259   (gnus-gethash (nnweb-identifier url) nnweb-hashtb))
 260
 261 (defun nnweb-identifier (ident)
 262   (funcall (nnweb-definition 'identifier) ident))
 263
 264 (defun nnweb-overview-file (group)
 265   "Return the name of the overview file of GROUP."
 266   (nnheader-concat nnweb-directory group ".overview"))
 267
 268 (defun nnweb-write-active ()
 269   "Save the active file."
 270   (gnus-make-directory nnweb-directory)
 271   (with-temp-file (nnheader-concat nnweb-directory "active")
 272     (prin1 `(setq nnweb-group-alist ',nnweb-group-alist) (current-buffer))))
 273
 274 (defun nnweb-read-active ()
 275   "Read the active file."
 276   (load (nnheader-concat nnweb-directory "active") t t t))
 277
 278 (defun nnweb-definition (type &optional noerror)
 279   "Return the definition of TYPE."
 280   (let ((def (cdr (assq type (assq nnweb-type nnweb-type-definition)))))
 281     (when (and (not def)
 282                (not noerror))
 283       (error "Undefined definition %s" type))
 284     def))
 285
 286 (defun nnweb-possibly-change-server (&optional group server)
 287   (nnweb-init server)
 288   (when server
 289     (unless (nnweb-server-opened server)
 290       (nnweb-open-server server)))
 291   (unless nnweb-group-alist
 292     (nnweb-read-active))
 293   (when group
 294     (when (and (not nnweb-ephemeral-p)
 295                (not (equal group nnweb-group)))
 296       (setq nnweb-hashtb (gnus-make-hashtable 4095))
 297       (nnweb-request-group group nil t))))
 298
 299 (defun nnweb-init (server)
 300   "Initialize buffers and such."
 301   (unless (gnus-buffer-live-p nnweb-buffer)
 302     (setq nnweb-buffer
 303           (save-excursion
 304             (mm-with-unibyte
 305               (nnheader-set-temp-buffer
 306                (format " *nnweb %s %s %s*"
 307                        nnweb-type nnweb-search server))
 308               (current-buffer))))))
 309
 310 (defun nnweb-fetch-url (url)
 311   (let (buf)
 312     (save-excursion
 313       (if (not nnheader-callback-function)
 314           (progn
 315             (with-temp-buffer
 316               (mm-enable-multibyte)
 317               (let ((coding-system-for-read 'binary)
 318                     (coding-system-for-write 'binary)
 319                     (default-process-coding-system 'binary))
 320                 (nnweb-insert url))
 321               (setq buf (buffer-string)))
 322             (erase-buffer)
 323             (insert buf)
 324             t)
 325         (nnweb-url-retrieve-asynch
 326          url 'nnweb-callback (current-buffer) nnheader-callback-function)
 327         t))))
 328
 329 (defun nnweb-callback (buffer callback)
 330   (when (gnus-buffer-live-p url-working-buffer)
 331     (save-excursion
 332       (set-buffer url-working-buffer)
 333       (funcall (nnweb-definition 'article))
 334       (nnweb-decode-entities)
 335       (set-buffer buffer)
 336       (goto-char (point-max))
 337       (insert-buffer-substring url-working-buffer))
 338     (funcall callback t)
 339     (gnus-kill-buffer url-working-buffer)))
 340
 341 (defun nnweb-url-retrieve-asynch (url callback &rest data)
 342   (let ((url-request-method "GET")
 343         (old-asynch url-be-asynchronous)
 344         (url-request-data nil)
 345         (url-request-extra-headers nil)
 346         (url-working-buffer (generate-new-buffer-name " *nnweb*")))
 347     (setq-default url-be-asynchronous t)
 348     (save-excursion
 349       (set-buffer (get-buffer-create url-working-buffer))
 350       (setq url-current-callback-data data
 351             url-be-asynchronous t
 352             url-current-callback-func callback)
 353       (url-retrieve url))
 354     (setq-default url-be-asynchronous old-asynch)))
 355
 356 ;;;
 357 ;;; DejaNews functions.
 358 ;;;
 359
 360 (defun nnweb-dejanews-create-mapping ()
 361   "Perform the search and create an number-to-url alist."
 362   (save-excursion
 363     (set-buffer nnweb-buffer)
 364     (erase-buffer)
 365     (when (funcall (nnweb-definition 'search) nnweb-search)
 366       (let ((i 0)
 367             (more t)
 368             (case-fold-search t)
 369             (active (or (cadr (assoc nnweb-group nnweb-group-alist))
 370                         (cons 1 0)))
 371             subject date from
 372             map url parse a table group text)
 373         (while more
 374           ;; Go through all the article hits on this page.
 375           (goto-char (point-min))
 376           (setq parse (w3-parse-buffer (current-buffer))
 377                 table (nth 1 (nnweb-parse-find-all 'table parse)))
 378           (dolist (row (nth 2 (car (nth 2 table))))
 379             (setq a (nnweb-parse-find 'a row)
 380                   url (cdr (assq 'href (nth 1 a)))
 381                   text (nreverse (nnweb-text row)))
 382             (when a
 383               (setq subject (nth 4 text)
 384                     group (nth 2 text)
 385                     date (nth 1 text)
 386                     from (nth 0 text))
 387               (if (string-match "\\([0-9]+\\)/\\([0-9]+\\)/\\([0-9]+\\)" date)
 388                   (setq date (format "%s %s 00:00:00 %s"
 389                                      (car (rassq (string-to-number
 390                                                   (match-string 2 date))
 391                                                  parse-time-months))
 392                                      (match-string 3 date)
 393                                      (match-string 1 date)))
 394                 (setq date "Jan 1 00:00:00 0000"))
 395               (incf i)
 396               (setq url (concat url "&fmt=text"))
 397               (when (string-match "&context=[^&]+" url)
 398                 (setq url (replace-match "" t t url)))
 399               (unless (nnweb-get-hashtb url)
 400                 (push
 401                  (list
 402                   (incf (cdr active))
 403                   (make-full-mail-header
 404                    (cdr active) (concat subject " (" group ")") from date
 405                    (concat "<" (nnweb-identifier url) "@dejanews>")
 406                    nil 0 0 url))
 407                  map)
 408                 (nnweb-set-hashtb (cadar map) (car map)))))
 409           ;; See whether there is a "Get next 20 hits" button here.
 410           (goto-char (point-min))
 411           (if (or (not (re-search-forward
 412                         "HREF=\"\\([^\"]+\\)\"[<>b]+Next result" nil t))
 413                   (>= i nnweb-max-hits))
 414               (setq more nil)
 415             ;; Yup -- fetch it.
 416             (setq more (match-string 1))
 417             (erase-buffer)
 418             (url-insert-file-contents more)))
 419         ;; Return the articles in the right order.
 420         (setq nnweb-articles
 421               (sort (nconc nnweb-articles map) 'car-less-than-car))))))
 422
 423 (defun nnweb-dejanews-search (search)
 424   (nnweb-insert
 425    (concat
 426     (nnweb-definition 'address)
 427     "?"
 428     (nnweb-encode-www-form-urlencoded
 429      `(("ST" . "PS")
 430        ("svcclass" . "dnyr")
 431        ("QRY" . ,search)
 432        ("defaultOp" . "AND")
 433        ("DBS" . "1")
 434        ("OP" . "dnquery.xp")
 435        ("LNG" . "ALL")
 436        ("maxhits" . "100")
 437        ("threaded" . "0")
 438        ("format" . "verbose2")
 439        ("showsort" . "date")
 440        ("agesign" . "1")
 441        ("ageweight" . "1")))))
 442   t)
 443
 444 (defun nnweb-dejanewsold-search (search)
 445   (nnweb-fetch-form
 446    (nnweb-definition 'address)
 447    `(("query" . ,search)
 448      ("defaultOp" . "AND")
 449      ("svcclass" . "dnold")
 450      ("maxhits" . "100")
 451      ("format" . "verbose2")
 452      ("threaded" . "0")
 453      ("showsort" . "date")
 454      ("agesign" . "1")
 455      ("ageweight" . "1")))
 456   t)
 457
 458 (defun nnweb-dejanews-identity (url)
 459   "Return an unique identifier based on URL."
 460   (if (string-match "AN=\\([0-9]+\\)" url)
 461       (match-string 1 url)
 462     url))
 463
 464 ;;;
 465 ;;; InReference
 466 ;;;
 467
 468 (defun nnweb-reference-create-mapping ()
 469   "Perform the search and create an number-to-url alist."
 470   (save-excursion
 471     (set-buffer nnweb-buffer)
 472     (erase-buffer)
 473     (when (funcall (nnweb-definition 'search) nnweb-search)
 474       (let ((i 0)
 475             (more t)
 476             (case-fold-search t)
 477             (active (or (cadr (assoc nnweb-group nnweb-group-alist))
 478                         (cons 1 0)))
 479             Subject Score Date Newsgroups From Message-ID
 480             map url)
 481         (while more
 482           ;; Go through all the article hits on this page.
 483           (goto-char (point-min))
 484           (search-forward "</pre><hr>" nil t)
 485           (delete-region (point-min) (point))
 486           (goto-char (point-min))
 487           (while (re-search-forward "^ +[0-9]+\\." nil t)
 488             (narrow-to-region
 489              (point)
 490              (if (re-search-forward "^$" nil t)
 491                  (match-beginning 0)
 492                (point-max)))
 493             (goto-char (point-min))
 494             (when (looking-at ".*href=\"\\([^\"]+\\)\"")
 495               (setq url (match-string 1)))
 496             (nnweb-remove-markup)
 497             (goto-char (point-min))
 498             (while (search-forward "\t" nil t)
 499               (replace-match " "))
 500             (goto-char (point-min))
 501             (while (re-search-forward "^\\([^:]+\\): \\(.*\\)$" nil t)
 502               (set (intern (match-string 1)) (match-string 2)))
 503             (widen)
 504             (search-forward "</pre>" nil t)
 505             (incf i)
 506             (unless (nnweb-get-hashtb url)
 507               (push
 508                (list
 509                 (incf (cdr active))
 510                 (make-full-mail-header
 511                  (cdr active) (concat  "(" Newsgroups ") " Subject) From Date
 512                  Message-ID
 513                  nil 0 (string-to-int Score) url))
 514                map)
 515               (nnweb-set-hashtb (cadar map) (car map))))
 516           (setq more nil))
 517         ;; Return the articles in the right order.
 518         (setq nnweb-articles
 519               (sort (nconc nnweb-articles map) 'car-less-than-car))))))
 520
 521 (defun nnweb-reference-wash-article ()
 522   (let ((case-fold-search t))
 523     (goto-char (point-min))
 524     (re-search-forward "^</center><hr>" nil t)
 525     (delete-region (point-min) (point))
 526     (search-forward "<pre>" nil t)
 527     (forward-line -1)
 528     (let ((body (point-marker)))
 529       (search-forward "</pre>" nil t)
 530       (delete-region (point) (point-max))
 531       (nnweb-remove-markup)
 532       (goto-char (point-min))
 533       (while (looking-at " *$")
 534         (gnus-delete-line))
 535       (narrow-to-region (point-min) body)
 536       (while (and (re-search-forward "^$" nil t)
 537                   (not (eobp)))
 538         (gnus-delete-line))
 539       (goto-char (point-min))
 540       (while (looking-at "\\(^[^ ]+:\\) *")
 541         (replace-match "\\1 " t)
 542         (forward-line 1))
 543       (goto-char (point-min))
 544       (when (re-search-forward "^References:" nil t)
 545         (narrow-to-region
 546          (point) (if (re-search-forward "^$\\|^[^:]+:" nil t)
 547                      (match-beginning 0)
 548                    (point-max)))
 549         (goto-char (point-min))
 550         (while (not (eobp))
 551           (unless (looking-at "References")
 552             (insert "\t")
 553             (forward-line 1)))
 554         (goto-char (point-min))
 555         (while (search-forward "," nil t)
 556           (replace-match " " t t)))
 557       (widen)
 558       (set-marker body nil))))
 559
 560 (defun nnweb-reference-search (search)
 561   (url-insert-file-contents
 562    (concat
 563     (nnweb-definition 'address)
 564     "?"
 565     (nnweb-encode-www-form-urlencoded
 566      `(("search" . "advanced")
 567        ("querytext" . ,search)
 568        ("subj" . "")
 569        ("name" . "")
 570        ("login" . "")
 571        ("host" . "")
 572        ("organization" . "")
 573        ("groups" . "")
 574        ("keywords" . "")
 575        ("choice" . "Search")
 576        ("startmonth" . "Jul")
 577        ("startday" . "25")
 578        ("startyear" . "1996")
 579        ("endmonth" . "Aug")
 580        ("endday" . "24")
 581        ("endyear" . "1996")
 582        ("mode" . "Quick")
 583        ("verbosity" . "Verbose")
 584        ("ranking" . "Relevance")
 585        ("first" . "1")
 586        ("last" . "25")
 587        ("score" . "50")))))
 588   (setq buffer-file-name nil)
 589   t)
 590
 591 ;;;
 592 ;;; Alta Vista
 593 ;;;
 594
 595 (defun nnweb-altavista-create-mapping ()
 596   "Perform the search and create an number-to-url alist."
 597   (save-excursion
 598     (set-buffer nnweb-buffer)
 599     (erase-buffer)
 600     (let ((part 0))
 601       (when (funcall (nnweb-definition 'search) nnweb-search part)
 602         (let ((i 0)
 603               (more t)
 604               (case-fold-search t)
 605               (active (or (cadr (assoc nnweb-group nnweb-group-alist))
 606                           (cons 1 0)))
 607               subject date from id group
 608               map url)
 609           (while more
 610             ;; Go through all the article hits on this page.
 611             (goto-char (point-min))
 612             (search-forward "<dt>" nil t)
 613             (delete-region (point-min) (match-beginning 0))
 614             (goto-char (point-min))
 615             (while (search-forward "<dt>" nil t)
 616               (replace-match "\n<blubb>"))
 617             (nnweb-decode-entities)
 618             (goto-char (point-min))
 619             (while (re-search-forward "<blubb>.*href=\"\\([^\"]+\\)\"><strong>\\([^>]*\\)</strong></a><dd>\\([^-]+\\)- <b>\\([^<]+\\)<.*href=\"news:\\([^\"]+\\)\">.*\">\\(.+\\)</a><P>"
 620                                       nil t)
 621               (setq url (match-string 1)
 622                     subject (match-string 2)
 623                     date (match-string 3)
 624                     group (match-string 4)
 625                     id (concat "<" (match-string 5) ">")
 626                     from (match-string 6))
 627               (incf i)
 628               (unless (nnweb-get-hashtb url)
 629                 (push
 630                  (list
 631                   (incf (cdr active))
 632                   (make-full-mail-header
 633                    (cdr active) (concat  "(" group ") " subject) from date
 634                    id nil 0 0 url))
 635                  map)
 636                 (nnweb-set-hashtb (cadar map) (car map))))
 637             ;; See if we want more.
 638             (when (or (not nnweb-articles)
 639                       (>= i nnweb-max-hits)
 640                       (not (funcall (nnweb-definition 'search)
 641                                     nnweb-search (incf part))))
 642               (setq more nil)))
 643           ;; Return the articles in the right order.
 644           (setq nnweb-articles
 645                 (sort (nconc nnweb-articles map) 'car-less-than-car)))))))
 646
 647 (defun nnweb-altavista-wash-article ()
 648   (goto-char (point-min))
 649   (let ((case-fold-search t))
 650     (when (re-search-forward "^<strong>" nil t)
 651       (delete-region (point-min) (match-beginning 0)))
 652     (goto-char (point-min))
 653     (while (looking-at "<strong>\\([^ ]+\\) +</strong> +\\(.*\\)$")
 654       (replace-match "\\1: \\2" t)
 655       (forward-line 1))
 656     (when (re-search-backward "^References:" nil t)
 657       (narrow-to-region (point) (progn (forward-line 1) (point)))
 658       (goto-char (point-min))
 659       (while (re-search-forward "<A.*\\?id@\\([^\"]+\\)\">[0-9]+</A>" nil t)
 660         (replace-match "&lt;\\1&gt; " t)))
 661     (widen)
 662     (nnweb-remove-markup)))
 663
 664 (defun nnweb-altavista-search (search &optional part)
 665   (url-insert-file-contents
 666    (concat
 667     (nnweb-definition 'address)
 668     "?"
 669     (nnweb-encode-www-form-urlencoded
 670      `(("pg" . "aq")
 671        ("what" . "news")
 672        ,@(when part `(("stq" . ,(int-to-string (* part 30)))))
 673        ("fmt" . "d")
 674        ("q" . ,search)
 675        ("r" . "")
 676        ("d0" . "")
 677        ("d1" . "")))))
 678   (setq buffer-file-name nil)
 679   t)
 680
 681 ;;;
 682 ;;; General web/w3 interface utility functions
 683 ;;;
 684
 685 (defun nnweb-insert-html (parse)
 686   "Insert HTML based on a w3 parse tree."
 687   (if (stringp parse)
 688       (insert parse)
 689     (insert "<" (symbol-name (car parse)) " ")
 690     (insert (mapconcat
 691              (lambda (param)
 692                (concat (symbol-name (car param)) "="
 693                        (prin1-to-string
 694                         (if (consp (cdr param))
 695                             (cadr param)
 696                           (cdr param)))))
 697              (nth 1 parse)
 698              " "))
 699     (insert ">\n")
 700     (mapcar 'nnweb-insert-html (nth 2 parse))
 701     (insert "</" (symbol-name (car parse)) ">\n")))
 702
 703 (defun nnweb-encode-www-form-urlencoded (pairs)
 704   "Return PAIRS encoded for forms."
 705   (mapconcat
 706    (function
 707     (lambda (data)
 708       (concat (w3-form-encode-xwfu (car data)) "="
 709               (w3-form-encode-xwfu (cdr data)))))
 710    pairs "&"))
 711
 712 (defun nnweb-fetch-form (url pairs)
 713   "Fetch a form from URL with PAIRS as the data using the POST method."
 714   (let ((url-request-data (nnweb-encode-www-form-urlencoded pairs))
 715         (url-request-method "POST")
 716         (url-request-extra-headers
 717          '(("Content-type" . "application/x-www-form-urlencoded"))))
 718     (url-insert-file-contents url)
 719     (setq buffer-file-name nil))
 720   t)
 721
 722 (defun nnweb-decode-entities ()
 723   "Decode all HTML entities."
 724   (goto-char (point-min))
 725   (while (re-search-forward "&\\(#[0-9]+\\|[a-z]+\\);" nil t)
 726     (replace-match (char-to-string
 727                     (if (eq (aref (match-string 1) 0) ?\#)
 728                         (let ((c
 729                                (string-to-number (substring
 730                                                   (match-string 1) 1))))
 731                           (if (mm-char-or-char-int-p c) c 32))
 732                       (or (cdr (assq (intern (match-string 1))
 733                                      w3-html-entities))
 734                           ?#)))
 735                    t t)))
 736
 737 (defun nnweb-decode-entities-string (str)
 738   (with-temp-buffer
 739     (insert str)
 740     (nnweb-decode-entities)
 741     (buffer-substring (point-min) (point-max))))
 742
 743 (defun nnweb-remove-markup ()
 744   "Remove all HTML markup, leaving just plain text."
 745   (goto-char (point-min))
 746   (while (search-forward "<!--" nil t)
 747     (delete-region (match-beginning 0)
 748                    (or (search-forward "-->" nil t)
 749                        (point-max))))
 750   (goto-char (point-min))
 751   (while (re-search-forward "<[^>]+>" nil t)
 752     (replace-match "" t t)))
 753
 754 (defun nnweb-insert (url &optional follow-refresh)
 755   "Insert the contents from an URL in the current buffer.
 756 If FOLLOW-REFRESH is non-nil, redirect refresh url in META."
 757   (let ((name buffer-file-name))
 758     (if follow-refresh
 759         (save-restriction
 760           (narrow-to-region (point) (point))
 761           (url-insert-file-contents url)
 762           (goto-char (point-min))
 763           (when (re-search-forward
 764                  "<meta[ \t\r\n]*http-equiv=\"Refresh\"[^>]*URL=\\([^\"]+\\)\"" nil t)
 765             (let ((url (match-string 1)))
 766               (delete-region (point-min) (point-max))
 767               (nnweb-insert url t))))
 768       (url-insert-file-contents url))
 769     (setq buffer-file-name name)))
 770
 771 (defun nnweb-parse-find (type parse &optional maxdepth)
 772   "Find the element of TYPE in PARSE."
 773   (catch 'found
 774     (nnweb-parse-find-1 type parse maxdepth)))
 775
 776 (defun nnweb-parse-find-1 (type contents maxdepth)
 777   (when (or (null maxdepth)
 778             (not (zerop maxdepth)))
 779     (when (consp contents)
 780       (when (eq (car contents) type)
 781         (throw 'found contents))
 782       (when (listp (cdr contents))
 783         (dolist (element contents)
 784           (when (consp element)
 785             (nnweb-parse-find-1 type element
 786                                 (and maxdepth (1- maxdepth)))))))))
 787
 788 (defun nnweb-parse-find-all (type parse)
 789   "Find all elements of TYPE in PARSE."
 790   (catch 'found
 791     (nnweb-parse-find-all-1 type parse)))
 792
 793 (defun nnweb-parse-find-all-1 (type contents)
 794   (let (result)
 795     (when (consp contents)
 796       (if (eq (car contents) type)
 797           (push contents result)
 798         (when (listp (cdr contents))
 799           (dolist (element contents)
 800             (when (consp element)
 801               (setq result
 802                     (nconc result (nnweb-parse-find-all-1 type element))))))))
 803     result))
 804
 805 (defvar nnweb-text)
 806 (defun nnweb-text (parse)
 807   "Return a list of text contents in PARSE."
 808   (let ((nnweb-text nil))
 809     (nnweb-text-1 parse)
 810     (nreverse nnweb-text)))
 811
 812 (defun nnweb-text-1 (contents)
 813   (dolist (element contents)
 814     (if (stringp element)
 815         (push element nnweb-text)
 816       (when (and (consp element)
 817                  (listp (cdr element)))
 818         (nnweb-text-1 element)))))
 819
 820 (provide 'nnweb)
 821
 822 ;;; nnweb.el ends here