Commit | Line | Data |
---|---|---|
eec82323 | 1 | ;;; nnweb.el --- retrieving articles via web search engines |
e84b4b86 TTN |
2 | |
3 | ;; Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, | |
46e8fe3d | 4 | ;; 2004, 2005, 2006 Free Software Foundation, Inc. |
eec82323 | 5 | |
6748645f | 6 | ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org> |
eec82323 LMI |
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 | |
3a35cf56 LK |
23 | ;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, |
24 | ;; Boston, MA 02110-1301, USA. | |
eec82323 LMI |
25 | |
26 | ;;; Commentary: | |
27 | ||
23f87bed | 28 | ;; Note: You need to have `w3' installed for some functions to work. |
eec82323 | 29 | |
46e8fe3d MB |
30 | ;; FIXME: Due to changes in the HTML output of Gmane, stuff related to Gmane |
31 | ;; web groups (`gnus-group-make-web-group') doesn't work anymore. | |
4a2358e9 | 32 | |
eec82323 LMI |
33 | ;;; Code: |
34 | ||
5ab7173c RS |
35 | (eval-when-compile (require 'cl)) |
36 | ||
eec82323 LMI |
37 | (require 'nnoo) |
38 | (require 'message) | |
39 | (require 'gnus-util) | |
40 | (require 'gnus) | |
eec82323 | 41 | (require 'nnmail) |
16409b0b | 42 | (require 'mm-util) |
23f87bed MB |
43 | (require 'mm-url) |
44 | (eval-and-compile | |
0d5dc4a5 | 45 | (ignore-errors |
23f87bed MB |
46 | (require 'url))) |
47 | (autoload 'w3-parse-buffer "w3-parse") | |
eec82323 LMI |
48 | |
49 | (nnoo-declare nnweb) | |
50 | ||
51 | (defvoo nnweb-directory (nnheader-concat gnus-directory "nnweb/") | |
52 | "Where nnweb will save its files.") | |
53 | ||
95fa1ff7 | 54 | (defvoo nnweb-type 'google |
6748645f | 55 | "What search engine type is being used. |
23f87bed | 56 | Valid types include `google', `dejanews', and `gmane'.") |
eec82323 | 57 | |
16409b0b | 58 | (defvar nnweb-type-definition |
23f87bed | 59 | |
4a2358e9 | 60 | (id . "http://www.google.com/groups?as_umsgid=%s&hl=en&dmode=source") |
46e8fe3d | 61 | (result . "http://groups.google.com/group/%s/msg/%s?dmode=source") |
4a2358e9 | 62 | (article . nnweb-google-wash-article) |
95fa1ff7 SZ |
63 | (reference . identity) |
64 | (map . nnweb-google-create-mapping) | |
65 | (search . nnweb-google-search) | |
4a2358e9 MB |
66 | (address . "http://groups.google.com/groups") |
67 | (base . "http://groups.google.com") | |
95fa1ff7 SZ |
68 | (identifier . nnweb-google-identity)) |
69 | (dejanews ;; alias of google | |
46e8fe3d MB |
70 | (id . "http://www.google.com/groups?as_umsgid=%s&hl=en&dmode=source") |
71 | (result . "http://groups.google.com/group/%s/msg/%s?dmode=source") | |
72 | (article . nnweb-google-wash-article) | |
95fa1ff7 SZ |
73 | (reference . identity) |
74 | (map . nnweb-google-create-mapping) | |
75 | (search . nnweb-google-search) | |
76 | (address . "http://groups.google.com/groups") | |
5f5475ac | 77 | (base . "http://groups.google.com") |
95fa1ff7 | 78 | (identifier . nnweb-google-identity)) |
23f87bed MB |
79 | (gmane |
80 | (article . nnweb-gmane-wash-article) | |
81 | (id . "http://gmane.org/view.php?group=%s") | |
82 | (reference . identity) | |
83 | (map . nnweb-gmane-create-mapping) | |
84 | (search . nnweb-gmane-search) | |
85 | (address . "http://gmane.org/") | |
86 | (identifier . nnweb-gmane-identity))) | |
eec82323 LMI |
87 | "Type-definition alist.") |
88 | ||
89 | (defvoo nnweb-search nil | |
23f87bed | 90 | "Search string to feed to Google.") |
eec82323 | 91 | |
6748645f | 92 | (defvoo nnweb-max-hits 999 |
eec82323 LMI |
93 | "Maximum number of hits to display.") |
94 | ||
95 | (defvoo nnweb-ephemeral-p nil | |
96 | "Whether this nnweb server is ephemeral.") | |
97 | ||
98 | ;;; Internal variables | |
99 | ||
100 | (defvoo nnweb-articles nil) | |
101 | (defvoo nnweb-buffer nil) | |
46e8fe3d | 102 | (defvar nnweb-group-alist nil) |
eec82323 LMI |
103 | (defvoo nnweb-group nil) |
104 | (defvoo nnweb-hashtb nil) | |
105 | ||
106 | ;;; Interface functions | |
107 | ||
108 | (nnoo-define-basics nnweb) | |
109 | ||
110 | (deffoo nnweb-retrieve-headers (articles &optional group server fetch-old) | |
111 | (nnweb-possibly-change-server group server) | |
112 | (save-excursion | |
113 | (set-buffer nntp-server-buffer) | |
114 | (erase-buffer) | |
115 | (let (article header) | |
16409b0b GM |
116 | (mm-with-unibyte-current-buffer |
117 | (while (setq article (pop articles)) | |
118 | (when (setq header (cadr (assq article nnweb-articles))) | |
119 | (nnheader-insert-nov header)))) | |
eec82323 LMI |
120 | 'nov))) |
121 | ||
122 | (deffoo nnweb-request-scan (&optional group server) | |
123 | (nnweb-possibly-change-server group server) | |
95fa1ff7 | 124 | (if nnweb-ephemeral-p |
46e8fe3d MB |
125 | (setq nnweb-hashtb (gnus-make-hashtable 4095)) |
126 | (unless nnweb-articles | |
127 | (nnweb-read-overview group))) | |
eec82323 LMI |
128 | (funcall (nnweb-definition 'map)) |
129 | (unless nnweb-ephemeral-p | |
130 | (nnweb-write-active) | |
131 | (nnweb-write-overview group))) | |
132 | ||
133 | (deffoo nnweb-request-group (group &optional server dont-check) | |
46e8fe3d MB |
134 | (nnweb-possibly-change-server group server) |
135 | (unless (or nnweb-ephemeral-p | |
6203370b MB |
136 | dont-check |
137 | nnweb-articles) | |
46e8fe3d | 138 | (nnweb-read-overview group)) |
eec82323 LMI |
139 | (cond |
140 | ((not nnweb-articles) | |
141 | (nnheader-report 'nnweb "No matching articles")) | |
142 | (t | |
143 | (let ((active (if nnweb-ephemeral-p | |
144 | (cons (caar nnweb-articles) | |
145 | (caar (last nnweb-articles))) | |
146 | (cadr (assoc group nnweb-group-alist))))) | |
147 | (nnheader-report 'nnweb "Opened group %s" group) | |
148 | (nnheader-insert | |
149 | "211 %d %d %d %s\n" (length nnweb-articles) | |
150 | (car active) (cdr active) group))))) | |
151 | ||
152 | (deffoo nnweb-close-group (group &optional server) | |
153 | (nnweb-possibly-change-server group server) | |
154 | (when (gnus-buffer-live-p nnweb-buffer) | |
155 | (save-excursion | |
156 | (set-buffer nnweb-buffer) | |
157 | (set-buffer-modified-p nil) | |
158 | (kill-buffer nnweb-buffer))) | |
159 | t) | |
160 | ||
161 | (deffoo nnweb-request-article (article &optional group server buffer) | |
162 | (nnweb-possibly-change-server group server) | |
163 | (save-excursion | |
164 | (set-buffer (or buffer nntp-server-buffer)) | |
165 | (let* ((header (cadr (assq article nnweb-articles))) | |
166 | (url (and header (mail-header-xref header)))) | |
167 | (when (or (and url | |
16409b0b | 168 | (mm-with-unibyte-current-buffer |
23f87bed | 169 | (mm-url-insert url))) |
eec82323 LMI |
170 | (and (stringp article) |
171 | (nnweb-definition 'id t) | |
172 | (let ((fetch (nnweb-definition 'id)) | |
95fa1ff7 | 173 | art active) |
eec82323 LMI |
174 | (when (string-match "^<\\(.*\\)>$" article) |
175 | (setq art (match-string 1 article))) | |
95fa1ff7 SZ |
176 | (when (and fetch art) |
177 | (setq url (format fetch art)) | |
178 | (mm-with-unibyte-current-buffer | |
23f87bed | 179 | (mm-url-insert url)) |
95fa1ff7 SZ |
180 | (if (nnweb-definition 'reference t) |
181 | (setq article | |
182 | (funcall (nnweb-definition | |
183 | 'reference) article))))))) | |
eec82323 | 184 | (unless nnheader-callback-function |
95fa1ff7 | 185 | (funcall (nnweb-definition 'article))) |
eec82323 | 186 | (nnheader-report 'nnweb "Fetched article %s" article) |
16409b0b | 187 | (cons group (and (numberp article) article)))))) |
eec82323 LMI |
188 | |
189 | (deffoo nnweb-close-server (&optional server) | |
190 | (when (and (nnweb-server-opened server) | |
191 | (gnus-buffer-live-p nnweb-buffer)) | |
192 | (save-excursion | |
193 | (set-buffer nnweb-buffer) | |
194 | (set-buffer-modified-p nil) | |
195 | (kill-buffer nnweb-buffer))) | |
196 | (nnoo-close-server 'nnweb server)) | |
197 | ||
198 | (deffoo nnweb-request-list (&optional server) | |
199 | (nnweb-possibly-change-server nil server) | |
200 | (save-excursion | |
201 | (set-buffer nntp-server-buffer) | |
46e8fe3d | 202 | (nnmail-generate-active (list (assoc server nnweb-group-alist))) |
eec82323 LMI |
203 | t)) |
204 | ||
205 | (deffoo nnweb-request-update-info (group info &optional server) | |
16409b0b | 206 | (nnweb-possibly-change-server group server)) |
eec82323 LMI |
207 | |
208 | (deffoo nnweb-asynchronous-p () | |
23f87bed | 209 | nil) |
eec82323 LMI |
210 | |
211 | (deffoo nnweb-request-create-group (group &optional server args) | |
212 | (nnweb-possibly-change-server nil server) | |
213 | (nnweb-request-delete-group group) | |
46e8fe3d | 214 | (push `(,group ,(cons 1 0)) nnweb-group-alist) |
eec82323 LMI |
215 | (nnweb-write-active) |
216 | t) | |
217 | ||
218 | (deffoo nnweb-request-delete-group (group &optional force server) | |
219 | (nnweb-possibly-change-server group server) | |
16409b0b GM |
220 | (gnus-pull group nnweb-group-alist t) |
221 | (nnweb-write-active) | |
eec82323 LMI |
222 | (gnus-delete-file (nnweb-overview-file group)) |
223 | t) | |
224 | ||
225 | (nnoo-define-skeleton nnweb) | |
226 | ||
227 | ;;; Internal functions | |
228 | ||
229 | (defun nnweb-read-overview (group) | |
230 | "Read the overview of GROUP and build the map." | |
231 | (when (file-exists-p (nnweb-overview-file group)) | |
16409b0b | 232 | (mm-with-unibyte-buffer |
eec82323 LMI |
233 | (nnheader-insert-file-contents (nnweb-overview-file group)) |
234 | (goto-char (point-min)) | |
235 | (let (header) | |
236 | (while (not (eobp)) | |
237 | (setq header (nnheader-parse-nov)) | |
238 | (forward-line 1) | |
239 | (push (list (mail-header-number header) | |
240 | header (mail-header-xref header)) | |
241 | nnweb-articles) | |
242 | (nnweb-set-hashtb header (car nnweb-articles))))))) | |
243 | ||
244 | (defun nnweb-write-overview (group) | |
245 | "Write the overview file for GROUP." | |
16409b0b | 246 | (with-temp-file (nnweb-overview-file group) |
eec82323 LMI |
247 | (let ((articles nnweb-articles)) |
248 | (while articles | |
249 | (nnheader-insert-nov (cadr (pop articles))))))) | |
250 | ||
251 | (defun nnweb-set-hashtb (header data) | |
252 | (gnus-sethash (nnweb-identifier (mail-header-xref header)) | |
253 | data nnweb-hashtb)) | |
254 | ||
255 | (defun nnweb-get-hashtb (url) | |
256 | (gnus-gethash (nnweb-identifier url) nnweb-hashtb)) | |
257 | ||
258 | (defun nnweb-identifier (ident) | |
259 | (funcall (nnweb-definition 'identifier) ident)) | |
260 | ||
261 | (defun nnweb-overview-file (group) | |
262 | "Return the name of the overview file of GROUP." | |
263 | (nnheader-concat nnweb-directory group ".overview")) | |
264 | ||
265 | (defun nnweb-write-active () | |
266 | "Save the active file." | |
16409b0b GM |
267 | (gnus-make-directory nnweb-directory) |
268 | (with-temp-file (nnheader-concat nnweb-directory "active") | |
eec82323 LMI |
269 | (prin1 `(setq nnweb-group-alist ',nnweb-group-alist) (current-buffer)))) |
270 | ||
271 | (defun nnweb-read-active () | |
272 | "Read the active file." | |
273 | (load (nnheader-concat nnweb-directory "active") t t t)) | |
274 | ||
275 | (defun nnweb-definition (type &optional noerror) | |
276 | "Return the definition of TYPE." | |
277 | (let ((def (cdr (assq type (assq nnweb-type nnweb-type-definition))))) | |
278 | (when (and (not def) | |
279 | (not noerror)) | |
280 | (error "Undefined definition %s" type)) | |
281 | def)) | |
282 | ||
283 | (defun nnweb-possibly-change-server (&optional group server) | |
eec82323 LMI |
284 | (when server |
285 | (unless (nnweb-server-opened server) | |
46e8fe3d MB |
286 | (nnweb-open-server server)) |
287 | (nnweb-init server)) | |
eec82323 LMI |
288 | (unless nnweb-group-alist |
289 | (nnweb-read-active)) | |
95fa1ff7 SZ |
290 | (unless nnweb-hashtb |
291 | (setq nnweb-hashtb (gnus-make-hashtable 4095))) | |
eec82323 | 292 | (when group |
46e8fe3d | 293 | (setq nnweb-group group))) |
eec82323 LMI |
294 | |
295 | (defun nnweb-init (server) | |
296 | "Initialize buffers and such." | |
297 | (unless (gnus-buffer-live-p nnweb-buffer) | |
298 | (setq nnweb-buffer | |
299 | (save-excursion | |
16409b0b GM |
300 | (mm-with-unibyte |
301 | (nnheader-set-temp-buffer | |
302 | (format " *nnweb %s %s %s*" | |
303 | nnweb-type nnweb-search server)) | |
304 | (current-buffer)))))) | |
eec82323 | 305 | |
95fa1ff7 | 306 | ;;; |
4a2358e9 | 307 | ;;; groups.google.com |
95fa1ff7 SZ |
308 | ;;; |
309 | ||
310 | (defun nnweb-google-wash-article () | |
4a2358e9 MB |
311 | ;; We have Google's masked e-mail addresses here. :-/ |
312 | (let ((case-fold-search t)) | |
95fa1ff7 | 313 | (goto-char (point-min)) |
d752cf53 MB |
314 | (if (save-excursion |
315 | (or (re-search-forward "The requested message.*could not be found." | |
316 | nil t) | |
317 | (not (and (re-search-forward "^<pre>" nil t) | |
318 | (re-search-forward "^</pre>" nil t))))) | |
319 | ;; FIXME: Don't know how to indicate "not found". | |
320 | ;; Should this function throw an error? --rsteib | |
321 | (progn | |
322 | (gnus-message 3 "Requested article not found") | |
323 | (erase-buffer)) | |
324 | (delete-region (point-min) | |
325 | (1+ (re-search-forward "^<pre>" nil t))) | |
326 | (goto-char (point-min)) | |
327 | (delete-region (- (re-search-forward "^</pre>" nil t) (length "</pre>")) | |
328 | (point-max)) | |
329 | (mm-url-decode-entities)))) | |
95fa1ff7 SZ |
330 | |
331 | (defun nnweb-google-parse-1 (&optional Message-ID) | |
46e8fe3d | 332 | "Parse search result in current buffer." |
95fa1ff7 SZ |
333 | (let ((i 0) |
334 | (case-fold-search t) | |
335 | (active (cadr (assoc nnweb-group nnweb-group-alist))) | |
336 | Subject Score Date Newsgroups From | |
337 | map url mid) | |
338 | (unless active | |
46e8fe3d | 339 | (push (list nnweb-group (setq active (cons 1 0))) |
95fa1ff7 SZ |
340 | nnweb-group-alist)) |
341 | ;; Go through all the article hits on this page. | |
342 | (goto-char (point-min)) | |
46e8fe3d MB |
343 | (while |
344 | (re-search-forward | |
345 | "a +href=\"/group/\\([^>\"]+\\)/browse_thread/[^>]+#\\([0-9a-f]+\\)" | |
346 | nil t) | |
347 | (setq Newsgroups (match-string-no-properties 1) | |
348 | ;; Note: Starting with Google Groups 2, `mid' is a Google-internal | |
349 | ;; ID, not a proper Message-ID. | |
350 | mid (match-string-no-properties 2) | |
debad045 | 351 | url (format |
46e8fe3d | 352 | (nnweb-definition 'result) Newsgroups mid)) |
95fa1ff7 SZ |
353 | (narrow-to-region (search-forward ">" nil t) |
354 | (search-forward "</a>" nil t)) | |
23f87bed MB |
355 | (mm-url-remove-markup) |
356 | (mm-url-decode-entities) | |
95fa1ff7 SZ |
357 | (setq Subject (buffer-string)) |
358 | (goto-char (point-max)) | |
359 | (widen) | |
46e8fe3d MB |
360 | (narrow-to-region (point) |
361 | (search-forward "</td" nil t)) | |
362 | ||
363 | (mm-url-remove-markup) | |
364 | (mm-url-decode-entities) | |
365 | (search-backward " - ") | |
95fa1ff7 | 366 | (when (looking-at |
46e8fe3d | 367 | " - \\([a-zA-Z]+\\) \\([0-9]+\\)\\(?: \\([0-9]\\{4\\}\\)\\)?, [^\n]+by \\([^<\n]+\\)\n") |
23f87bed MB |
368 | (setq From (match-string 4) |
369 | Date (format "%s %s 00:00:00 %s" | |
46e8fe3d MB |
370 | (match-string 1) |
371 | (match-string 2) | |
372 | (or (match-string 3) | |
373 | (substring (current-time-string) -4))))) | |
374 | ||
375 | (widen) | |
95fa1ff7 SZ |
376 | (forward-line 1) |
377 | (incf i) | |
378 | (unless (nnweb-get-hashtb url) | |
379 | (push | |
380 | (list | |
381 | (incf (cdr active)) | |
382 | (make-full-mail-header | |
383 | (cdr active) (if Newsgroups | |
384 | (concat "(" Newsgroups ") " Subject) | |
385 | Subject) | |
386 | From Date (or Message-ID mid) | |
387 | nil 0 0 url)) | |
388 | map) | |
389 | (nnweb-set-hashtb (cadar map) (car map)))) | |
390 | map)) | |
391 | ||
392 | (defun nnweb-google-reference (id) | |
393 | (let ((map (nnweb-google-parse-1 id)) header) | |
394 | (setq nnweb-articles | |
395 | (nconc nnweb-articles map)) | |
396 | (when (setq header (cadar map)) | |
397 | (mm-with-unibyte-current-buffer | |
23f87bed | 398 | (mm-url-insert (mail-header-xref header))) |
95fa1ff7 SZ |
399 | (caar map)))) |
400 | ||
401 | (defun nnweb-google-create-mapping () | |
debad045 | 402 | "Perform the search and create a number-to-url alist." |
95fa1ff7 SZ |
403 | (save-excursion |
404 | (set-buffer nnweb-buffer) | |
405 | (erase-buffer) | |
406 | (when (funcall (nnweb-definition 'search) nnweb-search) | |
23f87bed MB |
407 | (let ((more t) |
408 | (i 0)) | |
95fa1ff7 SZ |
409 | (while more |
410 | (setq nnweb-articles | |
411 | (nconc nnweb-articles (nnweb-google-parse-1))) | |
23f87bed MB |
412 | ;; Check if there are more articles to fetch |
413 | (goto-char (point-min)) | |
414 | (incf i 100) | |
415 | (if (or (not (re-search-forward | |
46e8fe3d | 416 | "<td><a href=\"\n\\([^>\"]+\\)\"><img src=\"/img/nav_next" nil t)) |
23f87bed MB |
417 | (>= i nnweb-max-hits)) |
418 | (setq more nil) | |
419 | ;; Yup, there are more articles | |
5f5475ac | 420 | (setq more (concat (nnweb-definition 'base) (match-string 1))) |
23f87bed MB |
421 | (when more |
422 | (erase-buffer) | |
423 | (mm-url-insert more)))) | |
95fa1ff7 SZ |
424 | ;; Return the articles in the right order. |
425 | (setq nnweb-articles | |
426 | (sort nnweb-articles 'car-less-than-car)))))) | |
427 | ||
428 | (defun nnweb-google-search (search) | |
23f87bed | 429 | (mm-url-insert |
95fa1ff7 SZ |
430 | (concat |
431 | (nnweb-definition 'address) | |
432 | "?" | |
23f87bed | 433 | (mm-url-encode-www-form-urlencoded |
95fa1ff7 | 434 | `(("q" . ,search) |
5f5475ac | 435 | ("num" . "100") |
95fa1ff7 | 436 | ("hq" . "") |
5f5475ac | 437 | ("hl" . "en") |
95fa1ff7 SZ |
438 | ("lr" . "") |
439 | ("safe" . "off") | |
46e8fe3d MB |
440 | ("sites" . "groups") |
441 | ("filter" . "0"))))) | |
95fa1ff7 SZ |
442 | t) |
443 | ||
444 | (defun nnweb-google-identity (url) | |
445 | "Return an unique identifier based on URL." | |
446 | (if (string-match "selm=\\([^ &>]+\\)" url) | |
447 | (match-string 1 url) | |
448 | url)) | |
449 | ||
23f87bed MB |
450 | ;;; |
451 | ;;; gmane.org | |
452 | ;;; | |
453 | (defun nnweb-gmane-create-mapping () | |
454 | "Perform the search and create a number-to-url alist." | |
455 | (save-excursion | |
456 | (set-buffer nnweb-buffer) | |
457 | (erase-buffer) | |
458 | (when (funcall (nnweb-definition 'search) nnweb-search) | |
459 | (let ((more t) | |
460 | (case-fold-search t) | |
461 | (active (or (cadr (assoc nnweb-group nnweb-group-alist)) | |
462 | (cons 1 0))) | |
463 | subject group url | |
464 | map) | |
465 | ;; Remove stuff from the beginning of results | |
466 | (goto-char (point-min)) | |
467 | (search-forward "Search Results</h1><ul>" nil t) | |
468 | (delete-region (point-min) (point)) | |
469 | (goto-char (point-min)) | |
470 | ;; Iterate over the actual hits | |
471 | (while (re-search-forward ".*href=\"\\([^\"]+\\)\">\\(.*\\)" nil t) | |
472 | (setq url (concat "http://gmane.org/" (match-string 1))) | |
473 | (setq subject (match-string 2)) | |
474 | (unless (nnweb-get-hashtb url) | |
475 | (push | |
476 | (list | |
477 | (incf (cdr active)) | |
478 | (make-full-mail-header | |
479 | (cdr active) (concat "(" group ") " subject) nil nil | |
480 | nil nil 0 0 url)) | |
481 | map) | |
482 | (nnweb-set-hashtb (cadar map) (car map)))) | |
483 | ;; Return the articles in the right order. | |
484 | (setq nnweb-articles | |
485 | (sort (nconc nnweb-articles map) 'car-less-than-car)))))) | |
486 | ||
487 | (defun nnweb-gmane-wash-article () | |
488 | (let ((case-fold-search t)) | |
489 | (goto-char (point-min)) | |
490 | (re-search-forward "<!--X-Head-of-Message-->" nil t) | |
491 | (delete-region (point-min) (point)) | |
492 | (goto-char (point-min)) | |
493 | (while (looking-at "^<li><em>\\([^ ]+\\)</em>.*</li>") | |
494 | (replace-match "\\1\\2" t) | |
495 | (forward-line 1)) | |
496 | (mm-url-remove-markup))) | |
497 | ||
498 | (defun nnweb-gmane-search (search) | |
499 | (mm-url-insert | |
500 | (concat | |
501 | (nnweb-definition 'address) | |
502 | "?" | |
503 | (mm-url-encode-www-form-urlencoded | |
504 | `(("query" . ,search))))) | |
505 | (setq buffer-file-name nil) | |
506 | t) | |
507 | ||
508 | ||
509 | (defun nnweb-gmane-identity (url) | |
510 | "Return a unique identifier based on URL." | |
511 | (if (string-match "group=\\(.+\\)" url) | |
512 | (match-string 1 url) | |
513 | url)) | |
514 | ||
16409b0b GM |
515 | ;;; |
516 | ;;; General web/w3 interface utility functions | |
517 | ;;; | |
518 | ||
519 | (defun nnweb-insert-html (parse) | |
520 | "Insert HTML based on a w3 parse tree." | |
521 | (if (stringp parse) | |
95fa1ff7 | 522 | (insert (nnheader-string-as-multibyte parse)) |
16409b0b GM |
523 | (insert "<" (symbol-name (car parse)) " ") |
524 | (insert (mapconcat | |
525 | (lambda (param) | |
526 | (concat (symbol-name (car param)) "=" | |
527 | (prin1-to-string | |
528 | (if (consp (cdr param)) | |
529 | (cadr param) | |
530 | (cdr param))))) | |
531 | (nth 1 parse) | |
532 | " ")) | |
533 | (insert ">\n") | |
534 | (mapcar 'nnweb-insert-html (nth 2 parse)) | |
535 | (insert "</" (symbol-name (car parse)) ">\n"))) | |
536 | ||
16409b0b GM |
537 | (defun nnweb-parse-find (type parse &optional maxdepth) |
538 | "Find the element of TYPE in PARSE." | |
539 | (catch 'found | |
540 | (nnweb-parse-find-1 type parse maxdepth))) | |
541 | ||
542 | (defun nnweb-parse-find-1 (type contents maxdepth) | |
543 | (when (or (null maxdepth) | |
544 | (not (zerop maxdepth))) | |
545 | (when (consp contents) | |
546 | (when (eq (car contents) type) | |
547 | (throw 'found contents)) | |
548 | (when (listp (cdr contents)) | |
549 | (dolist (element contents) | |
550 | (when (consp element) | |
551 | (nnweb-parse-find-1 type element | |
552 | (and maxdepth (1- maxdepth))))))))) | |
553 | ||
554 | (defun nnweb-parse-find-all (type parse) | |
555 | "Find all elements of TYPE in PARSE." | |
556 | (catch 'found | |
557 | (nnweb-parse-find-all-1 type parse))) | |
558 | ||
559 | (defun nnweb-parse-find-all-1 (type contents) | |
560 | (let (result) | |
561 | (when (consp contents) | |
562 | (if (eq (car contents) type) | |
563 | (push contents result) | |
564 | (when (listp (cdr contents)) | |
565 | (dolist (element contents) | |
566 | (when (consp element) | |
567 | (setq result | |
568 | (nconc result (nnweb-parse-find-all-1 type element)))))))) | |
569 | result)) | |
570 | ||
571 | (defvar nnweb-text) | |
572 | (defun nnweb-text (parse) | |
573 | "Return a list of text contents in PARSE." | |
574 | (let ((nnweb-text nil)) | |
575 | (nnweb-text-1 parse) | |
576 | (nreverse nnweb-text))) | |
577 | ||
578 | (defun nnweb-text-1 (contents) | |
579 | (dolist (element contents) | |
580 | (if (stringp element) | |
581 | (push element nnweb-text) | |
582 | (when (and (consp element) | |
583 | (listp (cdr element))) | |
584 | (nnweb-text-1 element))))) | |
585 | ||
eec82323 LMI |
586 | (provide 'nnweb) |
587 | ||
ab5796a9 | 588 | ;;; arch-tag: f59307eb-c90f-479f-b7d2-dbd8bf51b697 |
eec82323 | 589 | ;;; nnweb.el ends here |