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 | |
136 | dont-check) | |
137 | (nnweb-read-overview group)) | |
eec82323 LMI |
138 | (cond |
139 | ((not nnweb-articles) | |
140 | (nnheader-report 'nnweb "No matching articles")) | |
141 | (t | |
142 | (let ((active (if nnweb-ephemeral-p | |
143 | (cons (caar nnweb-articles) | |
144 | (caar (last nnweb-articles))) | |
145 | (cadr (assoc group nnweb-group-alist))))) | |
146 | (nnheader-report 'nnweb "Opened group %s" group) | |
147 | (nnheader-insert | |
148 | "211 %d %d %d %s\n" (length nnweb-articles) | |
149 | (car active) (cdr active) group))))) | |
150 | ||
151 | (deffoo nnweb-close-group (group &optional server) | |
152 | (nnweb-possibly-change-server group server) | |
153 | (when (gnus-buffer-live-p nnweb-buffer) | |
154 | (save-excursion | |
155 | (set-buffer nnweb-buffer) | |
156 | (set-buffer-modified-p nil) | |
157 | (kill-buffer nnweb-buffer))) | |
158 | t) | |
159 | ||
160 | (deffoo nnweb-request-article (article &optional group server buffer) | |
161 | (nnweb-possibly-change-server group server) | |
162 | (save-excursion | |
163 | (set-buffer (or buffer nntp-server-buffer)) | |
164 | (let* ((header (cadr (assq article nnweb-articles))) | |
165 | (url (and header (mail-header-xref header)))) | |
166 | (when (or (and url | |
16409b0b | 167 | (mm-with-unibyte-current-buffer |
23f87bed | 168 | (mm-url-insert url))) |
eec82323 LMI |
169 | (and (stringp article) |
170 | (nnweb-definition 'id t) | |
171 | (let ((fetch (nnweb-definition 'id)) | |
95fa1ff7 | 172 | art active) |
eec82323 LMI |
173 | (when (string-match "^<\\(.*\\)>$" article) |
174 | (setq art (match-string 1 article))) | |
95fa1ff7 SZ |
175 | (when (and fetch art) |
176 | (setq url (format fetch art)) | |
177 | (mm-with-unibyte-current-buffer | |
23f87bed | 178 | (mm-url-insert url)) |
95fa1ff7 SZ |
179 | (if (nnweb-definition 'reference t) |
180 | (setq article | |
181 | (funcall (nnweb-definition | |
182 | 'reference) article))))))) | |
eec82323 | 183 | (unless nnheader-callback-function |
95fa1ff7 | 184 | (funcall (nnweb-definition 'article))) |
eec82323 | 185 | (nnheader-report 'nnweb "Fetched article %s" article) |
16409b0b | 186 | (cons group (and (numberp article) article)))))) |
eec82323 LMI |
187 | |
188 | (deffoo nnweb-close-server (&optional server) | |
189 | (when (and (nnweb-server-opened server) | |
190 | (gnus-buffer-live-p nnweb-buffer)) | |
191 | (save-excursion | |
192 | (set-buffer nnweb-buffer) | |
193 | (set-buffer-modified-p nil) | |
194 | (kill-buffer nnweb-buffer))) | |
195 | (nnoo-close-server 'nnweb server)) | |
196 | ||
197 | (deffoo nnweb-request-list (&optional server) | |
198 | (nnweb-possibly-change-server nil server) | |
199 | (save-excursion | |
200 | (set-buffer nntp-server-buffer) | |
46e8fe3d | 201 | (nnmail-generate-active (list (assoc server nnweb-group-alist))) |
eec82323 LMI |
202 | t)) |
203 | ||
204 | (deffoo nnweb-request-update-info (group info &optional server) | |
16409b0b | 205 | (nnweb-possibly-change-server group server)) |
eec82323 LMI |
206 | |
207 | (deffoo nnweb-asynchronous-p () | |
23f87bed | 208 | nil) |
eec82323 LMI |
209 | |
210 | (deffoo nnweb-request-create-group (group &optional server args) | |
211 | (nnweb-possibly-change-server nil server) | |
212 | (nnweb-request-delete-group group) | |
46e8fe3d | 213 | (push `(,group ,(cons 1 0)) nnweb-group-alist) |
eec82323 LMI |
214 | (nnweb-write-active) |
215 | t) | |
216 | ||
217 | (deffoo nnweb-request-delete-group (group &optional force server) | |
218 | (nnweb-possibly-change-server group server) | |
16409b0b GM |
219 | (gnus-pull group nnweb-group-alist t) |
220 | (nnweb-write-active) | |
eec82323 LMI |
221 | (gnus-delete-file (nnweb-overview-file group)) |
222 | t) | |
223 | ||
224 | (nnoo-define-skeleton nnweb) | |
225 | ||
226 | ;;; Internal functions | |
227 | ||
228 | (defun nnweb-read-overview (group) | |
229 | "Read the overview of GROUP and build the map." | |
230 | (when (file-exists-p (nnweb-overview-file group)) | |
16409b0b | 231 | (mm-with-unibyte-buffer |
eec82323 LMI |
232 | (nnheader-insert-file-contents (nnweb-overview-file group)) |
233 | (goto-char (point-min)) | |
234 | (let (header) | |
235 | (while (not (eobp)) | |
236 | (setq header (nnheader-parse-nov)) | |
237 | (forward-line 1) | |
238 | (push (list (mail-header-number header) | |
239 | header (mail-header-xref header)) | |
240 | nnweb-articles) | |
241 | (nnweb-set-hashtb header (car nnweb-articles))))))) | |
242 | ||
243 | (defun nnweb-write-overview (group) | |
244 | "Write the overview file for GROUP." | |
16409b0b | 245 | (with-temp-file (nnweb-overview-file group) |
eec82323 LMI |
246 | (let ((articles nnweb-articles)) |
247 | (while articles | |
248 | (nnheader-insert-nov (cadr (pop articles))))))) | |
249 | ||
250 | (defun nnweb-set-hashtb (header data) | |
251 | (gnus-sethash (nnweb-identifier (mail-header-xref header)) | |
252 | data nnweb-hashtb)) | |
253 | ||
254 | (defun nnweb-get-hashtb (url) | |
255 | (gnus-gethash (nnweb-identifier url) nnweb-hashtb)) | |
256 | ||
257 | (defun nnweb-identifier (ident) | |
258 | (funcall (nnweb-definition 'identifier) ident)) | |
259 | ||
260 | (defun nnweb-overview-file (group) | |
261 | "Return the name of the overview file of GROUP." | |
262 | (nnheader-concat nnweb-directory group ".overview")) | |
263 | ||
264 | (defun nnweb-write-active () | |
265 | "Save the active file." | |
16409b0b GM |
266 | (gnus-make-directory nnweb-directory) |
267 | (with-temp-file (nnheader-concat nnweb-directory "active") | |
eec82323 LMI |
268 | (prin1 `(setq nnweb-group-alist ',nnweb-group-alist) (current-buffer)))) |
269 | ||
270 | (defun nnweb-read-active () | |
271 | "Read the active file." | |
272 | (load (nnheader-concat nnweb-directory "active") t t t)) | |
273 | ||
274 | (defun nnweb-definition (type &optional noerror) | |
275 | "Return the definition of TYPE." | |
276 | (let ((def (cdr (assq type (assq nnweb-type nnweb-type-definition))))) | |
277 | (when (and (not def) | |
278 | (not noerror)) | |
279 | (error "Undefined definition %s" type)) | |
280 | def)) | |
281 | ||
282 | (defun nnweb-possibly-change-server (&optional group server) | |
eec82323 LMI |
283 | (when server |
284 | (unless (nnweb-server-opened server) | |
46e8fe3d MB |
285 | (nnweb-open-server server)) |
286 | (nnweb-init server)) | |
eec82323 LMI |
287 | (unless nnweb-group-alist |
288 | (nnweb-read-active)) | |
95fa1ff7 SZ |
289 | (unless nnweb-hashtb |
290 | (setq nnweb-hashtb (gnus-make-hashtable 4095))) | |
eec82323 | 291 | (when group |
46e8fe3d | 292 | (setq nnweb-group group))) |
eec82323 LMI |
293 | |
294 | (defun nnweb-init (server) | |
295 | "Initialize buffers and such." | |
296 | (unless (gnus-buffer-live-p nnweb-buffer) | |
297 | (setq nnweb-buffer | |
298 | (save-excursion | |
16409b0b GM |
299 | (mm-with-unibyte |
300 | (nnheader-set-temp-buffer | |
301 | (format " *nnweb %s %s %s*" | |
302 | nnweb-type nnweb-search server)) | |
303 | (current-buffer)))))) | |
eec82323 | 304 | |
95fa1ff7 | 305 | ;;; |
4a2358e9 | 306 | ;;; groups.google.com |
95fa1ff7 SZ |
307 | ;;; |
308 | ||
309 | (defun nnweb-google-wash-article () | |
4a2358e9 MB |
310 | ;; We have Google's masked e-mail addresses here. :-/ |
311 | (let ((case-fold-search t)) | |
95fa1ff7 | 312 | (goto-char (point-min)) |
d752cf53 MB |
313 | (if (save-excursion |
314 | (or (re-search-forward "The requested message.*could not be found." | |
315 | nil t) | |
316 | (not (and (re-search-forward "^<pre>" nil t) | |
317 | (re-search-forward "^</pre>" nil t))))) | |
318 | ;; FIXME: Don't know how to indicate "not found". | |
319 | ;; Should this function throw an error? --rsteib | |
320 | (progn | |
321 | (gnus-message 3 "Requested article not found") | |
322 | (erase-buffer)) | |
323 | (delete-region (point-min) | |
324 | (1+ (re-search-forward "^<pre>" nil t))) | |
325 | (goto-char (point-min)) | |
326 | (delete-region (- (re-search-forward "^</pre>" nil t) (length "</pre>")) | |
327 | (point-max)) | |
328 | (mm-url-decode-entities)))) | |
95fa1ff7 SZ |
329 | |
330 | (defun nnweb-google-parse-1 (&optional Message-ID) | |
46e8fe3d | 331 | "Parse search result in current buffer." |
95fa1ff7 SZ |
332 | (let ((i 0) |
333 | (case-fold-search t) | |
334 | (active (cadr (assoc nnweb-group nnweb-group-alist))) | |
335 | Subject Score Date Newsgroups From | |
336 | map url mid) | |
337 | (unless active | |
46e8fe3d | 338 | (push (list nnweb-group (setq active (cons 1 0))) |
95fa1ff7 SZ |
339 | nnweb-group-alist)) |
340 | ;; Go through all the article hits on this page. | |
341 | (goto-char (point-min)) | |
46e8fe3d MB |
342 | (while |
343 | (re-search-forward | |
344 | "a +href=\"/group/\\([^>\"]+\\)/browse_thread/[^>]+#\\([0-9a-f]+\\)" | |
345 | nil t) | |
346 | (setq Newsgroups (match-string-no-properties 1) | |
347 | ;; Note: Starting with Google Groups 2, `mid' is a Google-internal | |
348 | ;; ID, not a proper Message-ID. | |
349 | mid (match-string-no-properties 2) | |
debad045 | 350 | url (format |
46e8fe3d | 351 | (nnweb-definition 'result) Newsgroups mid)) |
95fa1ff7 SZ |
352 | (narrow-to-region (search-forward ">" nil t) |
353 | (search-forward "</a>" nil t)) | |
23f87bed MB |
354 | (mm-url-remove-markup) |
355 | (mm-url-decode-entities) | |
95fa1ff7 SZ |
356 | (setq Subject (buffer-string)) |
357 | (goto-char (point-max)) | |
358 | (widen) | |
46e8fe3d MB |
359 | (narrow-to-region (point) |
360 | (search-forward "</td" nil t)) | |
361 | ||
362 | (mm-url-remove-markup) | |
363 | (mm-url-decode-entities) | |
364 | (search-backward " - ") | |
95fa1ff7 | 365 | (when (looking-at |
46e8fe3d | 366 | " - \\([a-zA-Z]+\\) \\([0-9]+\\)\\(?: \\([0-9]\\{4\\}\\)\\)?, [^\n]+by \\([^<\n]+\\)\n") |
23f87bed MB |
367 | (setq From (match-string 4) |
368 | Date (format "%s %s 00:00:00 %s" | |
46e8fe3d MB |
369 | (match-string 1) |
370 | (match-string 2) | |
371 | (or (match-string 3) | |
372 | (substring (current-time-string) -4))))) | |
373 | ||
374 | (widen) | |
95fa1ff7 SZ |
375 | (forward-line 1) |
376 | (incf i) | |
377 | (unless (nnweb-get-hashtb url) | |
378 | (push | |
379 | (list | |
380 | (incf (cdr active)) | |
381 | (make-full-mail-header | |
382 | (cdr active) (if Newsgroups | |
383 | (concat "(" Newsgroups ") " Subject) | |
384 | Subject) | |
385 | From Date (or Message-ID mid) | |
386 | nil 0 0 url)) | |
387 | map) | |
388 | (nnweb-set-hashtb (cadar map) (car map)))) | |
389 | map)) | |
390 | ||
391 | (defun nnweb-google-reference (id) | |
392 | (let ((map (nnweb-google-parse-1 id)) header) | |
393 | (setq nnweb-articles | |
394 | (nconc nnweb-articles map)) | |
395 | (when (setq header (cadar map)) | |
396 | (mm-with-unibyte-current-buffer | |
23f87bed | 397 | (mm-url-insert (mail-header-xref header))) |
95fa1ff7 SZ |
398 | (caar map)))) |
399 | ||
400 | (defun nnweb-google-create-mapping () | |
debad045 | 401 | "Perform the search and create a number-to-url alist." |
95fa1ff7 SZ |
402 | (save-excursion |
403 | (set-buffer nnweb-buffer) | |
404 | (erase-buffer) | |
405 | (when (funcall (nnweb-definition 'search) nnweb-search) | |
23f87bed MB |
406 | (let ((more t) |
407 | (i 0)) | |
95fa1ff7 SZ |
408 | (while more |
409 | (setq nnweb-articles | |
410 | (nconc nnweb-articles (nnweb-google-parse-1))) | |
23f87bed MB |
411 | ;; Check if there are more articles to fetch |
412 | (goto-char (point-min)) | |
413 | (incf i 100) | |
414 | (if (or (not (re-search-forward | |
46e8fe3d | 415 | "<td><a href=\"\n\\([^>\"]+\\)\"><img src=\"/img/nav_next" nil t)) |
23f87bed MB |
416 | (>= i nnweb-max-hits)) |
417 | (setq more nil) | |
418 | ;; Yup, there are more articles | |
5f5475ac | 419 | (setq more (concat (nnweb-definition 'base) (match-string 1))) |
23f87bed MB |
420 | (when more |
421 | (erase-buffer) | |
422 | (mm-url-insert more)))) | |
95fa1ff7 SZ |
423 | ;; Return the articles in the right order. |
424 | (setq nnweb-articles | |
425 | (sort nnweb-articles 'car-less-than-car)))))) | |
426 | ||
427 | (defun nnweb-google-search (search) | |
23f87bed | 428 | (mm-url-insert |
95fa1ff7 SZ |
429 | (concat |
430 | (nnweb-definition 'address) | |
431 | "?" | |
23f87bed | 432 | (mm-url-encode-www-form-urlencoded |
95fa1ff7 | 433 | `(("q" . ,search) |
5f5475ac | 434 | ("num" . "100") |
95fa1ff7 | 435 | ("hq" . "") |
5f5475ac | 436 | ("hl" . "en") |
95fa1ff7 SZ |
437 | ("lr" . "") |
438 | ("safe" . "off") | |
46e8fe3d MB |
439 | ("sites" . "groups") |
440 | ("filter" . "0"))))) | |
95fa1ff7 SZ |
441 | t) |
442 | ||
443 | (defun nnweb-google-identity (url) | |
444 | "Return an unique identifier based on URL." | |
445 | (if (string-match "selm=\\([^ &>]+\\)" url) | |
446 | (match-string 1 url) | |
447 | url)) | |
448 | ||
23f87bed MB |
449 | ;;; |
450 | ;;; gmane.org | |
451 | ;;; | |
452 | (defun nnweb-gmane-create-mapping () | |
453 | "Perform the search and create a number-to-url alist." | |
454 | (save-excursion | |
455 | (set-buffer nnweb-buffer) | |
456 | (erase-buffer) | |
457 | (when (funcall (nnweb-definition 'search) nnweb-search) | |
458 | (let ((more t) | |
459 | (case-fold-search t) | |
460 | (active (or (cadr (assoc nnweb-group nnweb-group-alist)) | |
461 | (cons 1 0))) | |
462 | subject group url | |
463 | map) | |
464 | ;; Remove stuff from the beginning of results | |
465 | (goto-char (point-min)) | |
466 | (search-forward "Search Results</h1><ul>" nil t) | |
467 | (delete-region (point-min) (point)) | |
468 | (goto-char (point-min)) | |
469 | ;; Iterate over the actual hits | |
470 | (while (re-search-forward ".*href=\"\\([^\"]+\\)\">\\(.*\\)" nil t) | |
471 | (setq url (concat "http://gmane.org/" (match-string 1))) | |
472 | (setq subject (match-string 2)) | |
473 | (unless (nnweb-get-hashtb url) | |
474 | (push | |
475 | (list | |
476 | (incf (cdr active)) | |
477 | (make-full-mail-header | |
478 | (cdr active) (concat "(" group ") " subject) nil nil | |
479 | nil nil 0 0 url)) | |
480 | map) | |
481 | (nnweb-set-hashtb (cadar map) (car map)))) | |
482 | ;; Return the articles in the right order. | |
483 | (setq nnweb-articles | |
484 | (sort (nconc nnweb-articles map) 'car-less-than-car)))))) | |
485 | ||
486 | (defun nnweb-gmane-wash-article () | |
487 | (let ((case-fold-search t)) | |
488 | (goto-char (point-min)) | |
489 | (re-search-forward "<!--X-Head-of-Message-->" nil t) | |
490 | (delete-region (point-min) (point)) | |
491 | (goto-char (point-min)) | |
492 | (while (looking-at "^<li><em>\\([^ ]+\\)</em>.*</li>") | |
493 | (replace-match "\\1\\2" t) | |
494 | (forward-line 1)) | |
495 | (mm-url-remove-markup))) | |
496 | ||
497 | (defun nnweb-gmane-search (search) | |
498 | (mm-url-insert | |
499 | (concat | |
500 | (nnweb-definition 'address) | |
501 | "?" | |
502 | (mm-url-encode-www-form-urlencoded | |
503 | `(("query" . ,search))))) | |
504 | (setq buffer-file-name nil) | |
505 | t) | |
506 | ||
507 | ||
508 | (defun nnweb-gmane-identity (url) | |
509 | "Return a unique identifier based on URL." | |
510 | (if (string-match "group=\\(.+\\)" url) | |
511 | (match-string 1 url) | |
512 | url)) | |
513 | ||
16409b0b GM |
514 | ;;; |
515 | ;;; General web/w3 interface utility functions | |
516 | ;;; | |
517 | ||
518 | (defun nnweb-insert-html (parse) | |
519 | "Insert HTML based on a w3 parse tree." | |
520 | (if (stringp parse) | |
95fa1ff7 | 521 | (insert (nnheader-string-as-multibyte parse)) |
16409b0b GM |
522 | (insert "<" (symbol-name (car parse)) " ") |
523 | (insert (mapconcat | |
524 | (lambda (param) | |
525 | (concat (symbol-name (car param)) "=" | |
526 | (prin1-to-string | |
527 | (if (consp (cdr param)) | |
528 | (cadr param) | |
529 | (cdr param))))) | |
530 | (nth 1 parse) | |
531 | " ")) | |
532 | (insert ">\n") | |
533 | (mapcar 'nnweb-insert-html (nth 2 parse)) | |
534 | (insert "</" (symbol-name (car parse)) ">\n"))) | |
535 | ||
16409b0b GM |
536 | (defun nnweb-parse-find (type parse &optional maxdepth) |
537 | "Find the element of TYPE in PARSE." | |
538 | (catch 'found | |
539 | (nnweb-parse-find-1 type parse maxdepth))) | |
540 | ||
541 | (defun nnweb-parse-find-1 (type contents maxdepth) | |
542 | (when (or (null maxdepth) | |
543 | (not (zerop maxdepth))) | |
544 | (when (consp contents) | |
545 | (when (eq (car contents) type) | |
546 | (throw 'found contents)) | |
547 | (when (listp (cdr contents)) | |
548 | (dolist (element contents) | |
549 | (when (consp element) | |
550 | (nnweb-parse-find-1 type element | |
551 | (and maxdepth (1- maxdepth))))))))) | |
552 | ||
553 | (defun nnweb-parse-find-all (type parse) | |
554 | "Find all elements of TYPE in PARSE." | |
555 | (catch 'found | |
556 | (nnweb-parse-find-all-1 type parse))) | |
557 | ||
558 | (defun nnweb-parse-find-all-1 (type contents) | |
559 | (let (result) | |
560 | (when (consp contents) | |
561 | (if (eq (car contents) type) | |
562 | (push contents result) | |
563 | (when (listp (cdr contents)) | |
564 | (dolist (element contents) | |
565 | (when (consp element) | |
566 | (setq result | |
567 | (nconc result (nnweb-parse-find-all-1 type element)))))))) | |
568 | result)) | |
569 | ||
570 | (defvar nnweb-text) | |
571 | (defun nnweb-text (parse) | |
572 | "Return a list of text contents in PARSE." | |
573 | (let ((nnweb-text nil)) | |
574 | (nnweb-text-1 parse) | |
575 | (nreverse nnweb-text))) | |
576 | ||
577 | (defun nnweb-text-1 (contents) | |
578 | (dolist (element contents) | |
579 | (if (stringp element) | |
580 | (push element nnweb-text) | |
581 | (when (and (consp element) | |
582 | (listp (cdr element))) | |
583 | (nnweb-text-1 element))))) | |
584 | ||
eec82323 LMI |
585 | (provide 'nnweb) |
586 | ||
ab5796a9 | 587 | ;;; arch-tag: f59307eb-c90f-479f-b7d2-dbd8bf51b697 |
eec82323 | 588 | ;;; nnweb.el ends here |