Revision: miles@gnu.org--gnu-2004/emacs--cvs-trunk--0--patch-716
[bpt/emacs.git] / lisp / gnus / spam.el
1 ;;; spam.el --- Identifying spam
2 ;; Copyright (C) 2002, 2003, 2004 Free Software Foundation, Inc.
3
4 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
5 ;; Keywords: network
6
7 ;; This file is part of GNU Emacs.
8
9 ;; GNU Emacs is free software; you can redistribute it and/or modify
10 ;; it under the terms of the GNU General Public License as published by
11 ;; the Free Software Foundation; either version 2, or (at your option)
12 ;; any later version.
13
14 ;; GNU Emacs is distributed in the hope that it will be useful,
15 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
16 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 ;; GNU General Public License for more details.
18
19 ;; You should have received a copy of the GNU General Public License
20 ;; along with GNU Emacs; see the file COPYING. If not, write to the
21 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
22 ;; Boston, MA 02111-1307, USA.
23
24 ;;; Commentary:
25
26 ;;; This module addresses a few aspects of spam control under Gnus. Page
27 ;;; breaks are used for grouping declarations and documentation relating to
28 ;;; each particular aspect.
29
30 ;;; The integration with Gnus is not yet complete. See various `FIXME'
31 ;;; comments, below, for supplementary explanations or discussions.
32
33 ;;; Several TODO items are marked as such
34
35 ;; TODO: spam scores, detection of spam in newsgroups, cross-server splitting,
36 ;; remote processing, training through files
37
38 ;;; Code:
39
40 (eval-when-compile (require 'cl))
41
42 (require 'gnus-sum)
43
44 (require 'gnus-uu) ; because of key prefix issues
45 ;;; for the definitions of group content classification and spam processors
46 (require 'gnus)
47 (require 'message) ;for the message-fetch-field functions
48
49 ;; for nnimap-split-download-body-default
50 (eval-when-compile (require 'nnimap))
51
52 ;; autoload executable-find
53 (eval-and-compile
54 ;; executable-find is not autoloaded in Emacs 20
55 (autoload 'executable-find "executable"))
56
57 ;; autoload query-dig
58 (eval-and-compile
59 (autoload 'query-dig "dig"))
60
61 ;; autoload spam-report
62 (eval-and-compile
63 (autoload 'spam-report-gmane "spam-report"))
64
65 ;; autoload gnus-registry
66 (eval-and-compile
67 (autoload 'gnus-registry-group-count "gnus-registry")
68 (autoload 'gnus-registry-add-group "gnus-registry")
69 (autoload 'gnus-registry-store-extra-entry "gnus-registry")
70 (autoload 'gnus-registry-fetch-extra "gnus-registry"))
71
72 ;; autoload query-dns
73 (eval-and-compile
74 (autoload 'query-dns "dns"))
75
76 ;;; Main parameters.
77
78 (defgroup spam nil
79 "Spam configuration."
80 :version "21.4")
81
82 (defcustom spam-directory (nnheader-concat gnus-directory "spam/")
83 "Directory for spam whitelists and blacklists."
84 :type 'directory
85 :group 'spam)
86
87 (defcustom spam-move-spam-nonspam-groups-only t
88 "Whether spam should be moved in non-spam groups only.
89 When t, only ham and unclassified groups will have their spam moved
90 to the spam-process-destination. When nil, spam will also be moved from
91 spam groups."
92 :type 'boolean
93 :group 'spam)
94
95 (defcustom spam-process-ham-in-nonham-groups nil
96 "Whether ham should be processed in non-ham groups."
97 :type 'boolean
98 :group 'spam)
99
100 (defcustom spam-log-to-registry nil
101 "Whether spam/ham processing should be logged in the registry."
102 :type 'boolean
103 :group 'spam)
104
105 (defcustom spam-split-symbolic-return nil
106 "Whether `spam-split' should work with symbols or group names."
107 :type 'boolean
108 :group 'spam)
109
110 (defcustom spam-split-symbolic-return-positive nil
111 "Whether `spam-split' should ALWAYS work with symbols or group names.
112 Do not set this if you use `spam-split' in a fancy split
113 method."
114 :type 'boolean
115 :group 'spam)
116
117 (defcustom spam-process-ham-in-spam-groups nil
118 "Whether ham should be processed in spam groups."
119 :type 'boolean
120 :group 'spam)
121
122 (defcustom spam-mark-only-unseen-as-spam t
123 "Whether only unseen articles should be marked as spam in spam groups.
124 When nil, all unread articles in a spam group are marked as
125 spam. Set this if you want to leave an article unread in a spam group
126 without losing it to the automatic spam-marking process."
127 :type 'boolean
128 :group 'spam)
129
130 (defcustom spam-mark-ham-unread-before-move-from-spam-group nil
131 "Whether ham should be marked unread before it's moved.
132 The article is moved out of a spam group according to ham-process-destination.
133 This variable is an official entry in the international Longest Variable Name
134 Competition."
135 :type 'boolean
136 :group 'spam)
137
138 (defcustom spam-disable-spam-split-during-ham-respool nil
139 "Whether `spam-split' should be ignored while resplitting ham in a process
140 destination. This is useful to prevent ham from ending up in the same spam
141 group after the resplit. Don't set this to t if you have spam-split as the
142 last rule in your split configuration."
143 :type 'boolean
144 :group 'spam)
145
146 (defcustom spam-autodetect-recheck-messages nil
147 "Should spam.el recheck all meessages when autodetecting?
148 Normally this is nil, so only unseen messages will be checked."
149 :type 'boolean
150 :group 'spam)
151
152 (defcustom spam-whitelist (expand-file-name "whitelist" spam-directory)
153 "The location of the whitelist.
154 The file format is one regular expression per line.
155 The regular expression is matched against the address."
156 :type 'file
157 :group 'spam)
158
159 (defcustom spam-blacklist (expand-file-name "blacklist" spam-directory)
160 "The location of the blacklist.
161 The file format is one regular expression per line.
162 The regular expression is matched against the address."
163 :type 'file
164 :group 'spam)
165
166 (defcustom spam-use-dig t
167 "Whether `query-dig' should be used instead of `query-dns'."
168 :type 'boolean
169 :group 'spam)
170
171 (defcustom spam-use-blacklist nil
172 "Whether the blacklist should be used by `spam-split'."
173 :type 'boolean
174 :group 'spam)
175
176 (defcustom spam-blacklist-ignored-regexes nil
177 "Regular expressions that the blacklist should ignore."
178 :type '(repeat (regexp :tag "Regular expression to ignore when blacklisting"))
179 :group 'spam)
180
181 (defcustom spam-use-whitelist nil
182 "Whether the whitelist should be used by `spam-split'."
183 :type 'boolean
184 :group 'spam)
185
186 (defcustom spam-use-whitelist-exclusive nil
187 "Whether whitelist-exclusive should be used by `spam-split'.
188 Exclusive whitelisting means that all messages from senders not in the whitelist
189 are considered spam."
190 :type 'boolean
191 :group 'spam)
192
193 (defcustom spam-use-blackholes nil
194 "Whether blackholes should be used by `spam-split'."
195 :type 'boolean
196 :group 'spam)
197
198 (defcustom spam-use-hashcash nil
199 "Whether hashcash payments should be detected by `spam-split'."
200 :type 'boolean
201 :group 'spam)
202
203 (defcustom spam-use-regex-headers nil
204 "Whether a header regular expression match should be used by `spam-split'.
205 Also see the variables `spam-regex-headers-spam' and `spam-regex-headers-ham'."
206 :type 'boolean
207 :group 'spam)
208
209 (defcustom spam-use-regex-body nil
210 "Whether a body regular expression match should be used by `spam-split'.
211 Also see the variables `spam-regex-body-spam' and `spam-regex-body-ham'."
212 :type 'boolean
213 :group 'spam)
214
215 (defcustom spam-use-bogofilter-headers nil
216 "Whether bogofilter headers should be used by `spam-split'.
217 Enable this if you pre-process messages with Bogofilter BEFORE Gnus sees them."
218 :type 'boolean
219 :group 'spam)
220
221 (defcustom spam-use-bogofilter nil
222 "Whether bogofilter should be invoked by `spam-split'.
223 Enable this if you want Gnus to invoke Bogofilter on new messages."
224 :type 'boolean
225 :group 'spam)
226
227 (defcustom spam-use-BBDB nil
228 "Whether BBDB should be used by `spam-split'."
229 :type 'boolean
230 :group 'spam)
231
232 (defcustom spam-use-BBDB-exclusive nil
233 "Whether BBDB-exclusive should be used by `spam-split'.
234 Exclusive BBDB means that all messages from senders not in the BBDB are
235 considered spam."
236 :type 'boolean
237 :group 'spam)
238
239 (defcustom spam-use-ifile nil
240 "Whether ifile should be used by `spam-split'."
241 :type 'boolean
242 :group 'spam)
243
244 (defcustom spam-use-stat nil
245 "Whether `spam-stat' should be used by `spam-split'."
246 :type 'boolean
247 :group 'spam)
248
249 (defcustom spam-use-spamoracle nil
250 "Whether spamoracle should be used by `spam-split'."
251 :type 'boolean
252 :group 'spam)
253
254 (defcustom spam-install-hooks (or
255 spam-use-dig
256 spam-use-blacklist
257 spam-use-whitelist
258 spam-use-whitelist-exclusive
259 spam-use-blackholes
260 spam-use-hashcash
261 spam-use-regex-headers
262 spam-use-regex-body
263 spam-use-bogofilter-headers
264 spam-use-bogofilter
265 spam-use-BBDB
266 spam-use-BBDB-exclusive
267 spam-use-ifile
268 spam-use-stat
269 spam-use-spamoracle)
270 "Whether the spam hooks should be installed.
271 Default to t if one of the spam-use-* variables is set."
272 :group 'spam
273 :type 'boolean)
274
275 (defcustom spam-split-group "spam"
276 "Group name where incoming spam should be put by `spam-split'."
277 :type 'string
278 :group 'spam)
279
280 ;;; TODO: deprecate this variable, it's confusing since it's a list of strings,
281 ;;; not regular expressions
282 (defcustom spam-junk-mailgroups (cons
283 spam-split-group
284 '("mail.junk" "poste.pourriel"))
285 "Mailgroups with spam contents.
286 All unmarked article in such group receive the spam mark on group entry."
287 :type '(repeat (string :tag "Group"))
288 :group 'spam)
289
290 (defcustom spam-blackhole-servers '("bl.spamcop.net" "relays.ordb.org"
291 "dev.null.dk" "relays.visi.com")
292 "List of blackhole servers."
293 :type '(repeat (string :tag "Server"))
294 :group 'spam)
295
296 (defcustom spam-blackhole-good-server-regex nil
297 "String matching IP addresses that should not be checked in the blackholes."
298 :type '(radio (const nil) regexp)
299 :group 'spam)
300
301 (defcustom spam-face 'gnus-splash-face
302 "Face for spam-marked articles."
303 :type 'face
304 :group 'spam)
305
306 (defcustom spam-regex-headers-spam '("^X-Spam-Flag: YES")
307 "Regular expression for positive header spam matches."
308 :type '(repeat (regexp :tag "Regular expression to match spam header"))
309 :group 'spam)
310
311 (defcustom spam-regex-headers-ham '("^X-Spam-Flag: NO")
312 "Regular expression for positive header ham matches."
313 :type '(repeat (regexp :tag "Regular expression to match ham header"))
314 :group 'spam)
315
316 (defcustom spam-regex-body-spam '()
317 "Regular expression for positive body spam matches."
318 :type '(repeat (regexp :tag "Regular expression to match spam body"))
319 :group 'spam)
320
321 (defcustom spam-regex-body-ham '()
322 "Regular expression for positive body ham matches."
323 :type '(repeat (regexp :tag "Regular expression to match ham body"))
324 :group 'spam)
325
326 (defgroup spam-ifile nil
327 "Spam ifile configuration."
328 :group 'spam)
329
330 (defcustom spam-ifile-path (executable-find "ifile")
331 "File path of the ifile executable program."
332 :type '(choice (file :tag "Location of ifile")
333 (const :tag "ifile is not installed"))
334 :group 'spam-ifile)
335
336 (defcustom spam-ifile-database-path nil
337 "File path of the ifile database."
338 :type '(choice (file :tag "Location of the ifile database")
339 (const :tag "Use the default"))
340 :group 'spam-ifile)
341
342 (defcustom spam-ifile-spam-category "spam"
343 "Name of the spam ifile category."
344 :type 'string
345 :group 'spam-ifile)
346
347 (defcustom spam-ifile-ham-category nil
348 "Name of the ham ifile category.
349 If nil, the current group name will be used."
350 :type '(choice (string :tag "Use a fixed category")
351 (const :tag "Use the current group name"))
352 :group 'spam-ifile)
353
354 (defcustom spam-ifile-all-categories nil
355 "Whether the ifile check will return all categories, or just spam.
356 Set this to t if you want to use the `spam-split' invocation of ifile as
357 your main source of newsgroup names."
358 :type 'boolean
359 :group 'spam-ifile)
360
361 (defgroup spam-bogofilter nil
362 "Spam bogofilter configuration."
363 :group 'spam)
364
365 (defcustom spam-bogofilter-path (executable-find "bogofilter")
366 "File path of the Bogofilter executable program."
367 :type '(choice (file :tag "Location of bogofilter")
368 (const :tag "Bogofilter is not installed"))
369 :group 'spam-bogofilter)
370
371 (defcustom spam-bogofilter-header "X-Bogosity"
372 "The header that Bogofilter inserts in messages."
373 :type 'string
374 :group 'spam-bogofilter)
375
376 (defcustom spam-bogofilter-spam-switch "-s"
377 "The switch that Bogofilter uses to register spam messages."
378 :type 'string
379 :group 'spam-bogofilter)
380
381 (defcustom spam-bogofilter-ham-switch "-n"
382 "The switch that Bogofilter uses to register ham messages."
383 :type 'string
384 :group 'spam-bogofilter)
385
386 (defcustom spam-bogofilter-spam-strong-switch "-S"
387 "The switch that Bogofilter uses to unregister ham messages."
388 :type 'string
389 :group 'spam-bogofilter)
390
391 (defcustom spam-bogofilter-ham-strong-switch "-N"
392 "The switch that Bogofilter uses to unregister spam messages."
393 :type 'string
394 :group 'spam-bogofilter)
395
396 (defcustom spam-bogofilter-bogosity-positive-spam-header "^\\(Yes\\|Spam\\)"
397 "The regex on `spam-bogofilter-header' for positive spam identification."
398 :type 'regexp
399 :group 'spam-bogofilter)
400
401 (defcustom spam-bogofilter-database-directory nil
402 "Directory path of the Bogofilter databases."
403 :type '(choice (directory
404 :tag "Location of the Bogofilter database directory")
405 (const :tag "Use the default"))
406 :group 'spam-bogofilter)
407
408 (defgroup spam-spamoracle nil
409 "Spam spamoracle configuration."
410 :group 'spam)
411
412 (defcustom spam-spamoracle-database nil
413 "Location of spamoracle database file. When nil, use the default
414 spamoracle database."
415 :type '(choice (directory :tag "Location of spamoracle database file.")
416 (const :tag "Use the default"))
417 :group 'spam-spamoracle)
418
419 (defcustom spam-spamoracle-binary (executable-find "spamoracle")
420 "Location of the spamoracle binary."
421 :type '(choice (directory :tag "Location of the spamoracle binary")
422 (const :tag "Use the default"))
423 :group 'spam-spamoracle)
424
425 ;;; Key bindings for spam control.
426
427 (gnus-define-keys gnus-summary-mode-map
428 "St" spam-bogofilter-score
429 "Sx" gnus-summary-mark-as-spam
430 "Mst" spam-bogofilter-score
431 "Msx" gnus-summary-mark-as-spam
432 "\M-d" gnus-summary-mark-as-spam)
433
434 (defvar spam-old-ham-articles nil
435 "List of old ham articles, generated when a group is entered.")
436
437 (defvar spam-old-spam-articles nil
438 "List of old spam articles, generated when a group is entered.")
439
440 (defvar spam-split-disabled nil
441 "If non-nil, `spam-split' is disabled, and always returns nil.")
442
443 (defvar spam-split-last-successful-check nil
444 "`spam-split' will set this to nil or a spam-use-XYZ check if it
445 finds ham or spam.")
446
447 ;; convenience functions
448 (defun spam-xor (a b)
449 "Logical exclusive `or'."
450 (and (or a b) (not (and a b))))
451
452 (defun spam-group-ham-mark-p (group mark &optional spam)
453 (when (stringp group)
454 (let* ((marks (spam-group-ham-marks group spam))
455 (marks (if (symbolp mark)
456 marks
457 (mapcar 'symbol-value marks))))
458 (memq mark marks))))
459
460 (defun spam-group-spam-mark-p (group mark)
461 (spam-group-ham-mark-p group mark t))
462
463 (defun spam-group-ham-marks (group &optional spam)
464 (when (stringp group)
465 (let* ((marks (if spam
466 (gnus-parameter-spam-marks group)
467 (gnus-parameter-ham-marks group)))
468 (marks (car marks))
469 (marks (if (listp (car marks)) (car marks) marks)))
470 marks)))
471
472 (defun spam-group-spam-marks (group)
473 (spam-group-ham-marks group t))
474
475 (defun spam-group-spam-contents-p (group)
476 (if (stringp group)
477 (or (member group spam-junk-mailgroups)
478 (memq 'gnus-group-spam-classification-spam
479 (gnus-parameter-spam-contents group)))
480 nil))
481
482 (defun spam-group-ham-contents-p (group)
483 (if (stringp group)
484 (memq 'gnus-group-spam-classification-ham
485 (gnus-parameter-spam-contents group))
486 nil))
487
488 (defvar spam-list-of-processors
489 '((gnus-group-spam-exit-processor-report-gmane spam spam-use-gmane)
490 (gnus-group-spam-exit-processor-bogofilter spam spam-use-bogofilter)
491 (gnus-group-spam-exit-processor-blacklist spam spam-use-blacklist)
492 (gnus-group-spam-exit-processor-ifile spam spam-use-ifile)
493 (gnus-group-spam-exit-processor-stat spam spam-use-stat)
494 (gnus-group-spam-exit-processor-spamoracle spam spam-use-spamoracle)
495 (gnus-group-ham-exit-processor-ifile ham spam-use-ifile)
496 (gnus-group-ham-exit-processor-bogofilter ham spam-use-bogofilter)
497 (gnus-group-ham-exit-processor-stat ham spam-use-stat)
498 (gnus-group-ham-exit-processor-whitelist ham spam-use-whitelist)
499 (gnus-group-ham-exit-processor-BBDB ham spam-use-BBDB)
500 (gnus-group-ham-exit-processor-copy ham spam-use-ham-copy)
501 (gnus-group-ham-exit-processor-spamoracle ham spam-use-spamoracle))
502 "The spam-list-of-processors list contains pairs associating a
503 ham/spam exit processor variable with a classification and a
504 spam-use-* variable.")
505
506 (defun spam-group-processor-p (group processor)
507 (if (and (stringp group)
508 (symbolp processor))
509 (or (member processor (nth 0 (gnus-parameter-spam-process group)))
510 (spam-group-processor-multiple-p
511 group
512 (cdr-safe (assoc processor spam-list-of-processors))))
513 nil))
514
515 (defun spam-group-processor-multiple-p (group processor-info)
516 (let* ((classification (nth 0 processor-info))
517 (check (nth 1 processor-info))
518 (parameters (nth 0 (gnus-parameter-spam-process group)))
519 found)
520 (dolist (parameter parameters)
521 (when (and (null found)
522 (listp parameter)
523 (eq classification (nth 0 parameter))
524 (eq check (nth 1 parameter)))
525 (setq found t)))
526 found))
527
528 (defun spam-group-spam-processor-report-gmane-p (group)
529 (spam-group-processor-p group 'gnus-group-spam-exit-processor-report-gmane))
530
531 (defun spam-group-spam-processor-bogofilter-p (group)
532 (spam-group-processor-p group 'gnus-group-spam-exit-processor-bogofilter))
533
534 (defun spam-group-spam-processor-blacklist-p (group)
535 (spam-group-processor-p group 'gnus-group-spam-exit-processor-blacklist))
536
537 (defun spam-group-spam-processor-ifile-p (group)
538 (spam-group-processor-p group 'gnus-group-spam-exit-processor-ifile))
539
540 (defun spam-group-ham-processor-ifile-p (group)
541 (spam-group-processor-p group 'gnus-group-ham-exit-processor-ifile))
542
543 (defun spam-group-spam-processor-spamoracle-p (group)
544 (spam-group-processor-p group 'gnus-group-spam-exit-processor-spamoracle))
545
546 (defun spam-group-ham-processor-bogofilter-p (group)
547 (spam-group-processor-p group 'gnus-group-ham-exit-processor-bogofilter))
548
549 (defun spam-group-spam-processor-stat-p (group)
550 (spam-group-processor-p group 'gnus-group-spam-exit-processor-stat))
551
552 (defun spam-group-ham-processor-stat-p (group)
553 (spam-group-processor-p group 'gnus-group-ham-exit-processor-stat))
554
555 (defun spam-group-ham-processor-whitelist-p (group)
556 (spam-group-processor-p group 'gnus-group-ham-exit-processor-whitelist))
557
558 (defun spam-group-ham-processor-BBDB-p (group)
559 (spam-group-processor-p group 'gnus-group-ham-exit-processor-BBDB))
560
561 (defun spam-group-ham-processor-copy-p (group)
562 (spam-group-processor-p group 'gnus-group-ham-exit-processor-copy))
563
564 (defun spam-group-ham-processor-spamoracle-p (group)
565 (spam-group-processor-p group 'gnus-group-ham-exit-processor-spamoracle))
566
567 ;;; Summary entry and exit processing.
568
569 (defun spam-summary-prepare ()
570 (setq spam-old-ham-articles
571 (spam-list-articles gnus-newsgroup-articles 'ham))
572 (setq spam-old-spam-articles
573 (spam-list-articles gnus-newsgroup-articles 'spam))
574 (spam-mark-junk-as-spam-routine))
575
576 ;; The spam processors are invoked for any group, spam or ham or neither
577 (defun spam-summary-prepare-exit ()
578 (unless gnus-group-is-exiting-without-update-p
579 (gnus-message 6 "Exiting summary buffer and applying spam rules")
580
581 ;; first of all, unregister any articles that are no longer ham or spam
582 ;; we have to iterate over the processors, or else we'll be too slow
583 (dolist (classification '(spam ham))
584 (let* ((old-articles (if (eq classification 'spam)
585 spam-old-spam-articles
586 spam-old-ham-articles))
587 (new-articles (spam-list-articles
588 gnus-newsgroup-articles
589 classification))
590 (changed-articles (gnus-set-difference old-articles new-articles)))
591 ;; now that we have the changed articles, we go through the processors
592 (dolist (processor-param spam-list-of-processors)
593 (let ((processor (nth 0 processor-param))
594 (processor-classification (nth 1 processor-param))
595 (check (nth 2 processor-param))
596 unregister-list)
597 (dolist (article changed-articles)
598 (let ((id (spam-fetch-field-message-id-fast article)))
599 (when (spam-log-unregistration-needed-p
600 id 'process classification check)
601 (push article unregister-list))))
602 ;; call spam-register-routine with specific articles to unregister,
603 ;; when there are articles to unregister and the check is enabled
604 (when (and unregister-list (symbol-value check))
605 (spam-register-routine classification check t unregister-list))))))
606
607 ;; find all the spam processors applicable to this group
608 (dolist (processor-param spam-list-of-processors)
609 (let ((processor (nth 0 processor-param))
610 (classification (nth 1 processor-param))
611 (check (nth 2 processor-param)))
612 (when (and (eq 'spam classification)
613 (spam-group-processor-p gnus-newsgroup-name processor))
614 (spam-register-routine classification check))))
615
616 (if spam-move-spam-nonspam-groups-only
617 (when (not (spam-group-spam-contents-p gnus-newsgroup-name))
618 (spam-mark-spam-as-expired-and-move-routine
619 (gnus-parameter-spam-process-destination gnus-newsgroup-name)))
620 (gnus-message 5 "Marking spam as expired and moving it to %s"
621 gnus-newsgroup-name)
622 (spam-mark-spam-as-expired-and-move-routine
623 (gnus-parameter-spam-process-destination gnus-newsgroup-name)))
624
625 ;; now we redo spam-mark-spam-as-expired-and-move-routine to only
626 ;; expire spam, in case the above did not expire them
627 (gnus-message 5 "Marking spam as expired without moving it")
628 (spam-mark-spam-as-expired-and-move-routine nil)
629
630 (when (or (spam-group-ham-contents-p gnus-newsgroup-name)
631 (and (spam-group-spam-contents-p gnus-newsgroup-name)
632 spam-process-ham-in-spam-groups)
633 spam-process-ham-in-nonham-groups)
634 ;; find all the ham processors applicable to this group
635 (dolist (processor-param spam-list-of-processors)
636 (let ((processor (nth 0 processor-param))
637 (classification (nth 1 processor-param))
638 (check (nth 2 processor-param)))
639 (when (and (eq 'ham classification)
640 (spam-group-processor-p gnus-newsgroup-name processor))
641 (spam-register-routine classification check)))))
642
643 (when (spam-group-ham-processor-copy-p gnus-newsgroup-name)
644 (gnus-message 5 "Copying ham")
645 (spam-ham-copy-routine
646 (gnus-parameter-ham-process-destination gnus-newsgroup-name)))
647
648 ;; now move all ham articles out of spam groups
649 (when (spam-group-spam-contents-p gnus-newsgroup-name)
650 (gnus-message 5 "Moving ham messages from spam group")
651 (spam-ham-move-routine
652 (gnus-parameter-ham-process-destination gnus-newsgroup-name))))
653
654 (setq spam-old-ham-articles nil)
655 (setq spam-old-spam-articles nil))
656
657 (defun spam-mark-junk-as-spam-routine ()
658 ;; check the global list of group names spam-junk-mailgroups and the
659 ;; group parameters
660 (when (spam-group-spam-contents-p gnus-newsgroup-name)
661 (gnus-message 5 "Marking %s articles as spam"
662 (if spam-mark-only-unseen-as-spam
663 "unseen"
664 "unread"))
665 (let ((articles (if spam-mark-only-unseen-as-spam
666 gnus-newsgroup-unseen
667 gnus-newsgroup-unreads)))
668 (dolist (article articles)
669 (gnus-summary-mark-article article gnus-spam-mark)))))
670
671 (defun spam-mark-spam-as-expired-and-move-routine (&rest groups)
672 (if (and (car-safe groups) (listp (car-safe groups)))
673 (apply 'spam-mark-spam-as-expired-and-move-routine (car groups))
674 (gnus-summary-kill-process-mark)
675 (let ((articles gnus-newsgroup-articles)
676 (backend-supports-deletions
677 (gnus-check-backend-function
678 'request-move-article gnus-newsgroup-name))
679 article tomove deletep)
680 (dolist (article articles)
681 (when (eq (gnus-summary-article-mark article) gnus-spam-mark)
682 (gnus-summary-mark-article article gnus-expirable-mark)
683 (push article tomove)))
684
685 ;; now do the actual copies
686 (dolist (group groups)
687 (when (and tomove
688 (stringp group))
689 (dolist (article tomove)
690 (gnus-summary-set-process-mark article))
691 (when tomove
692 (if (or (not backend-supports-deletions)
693 (> (length groups) 1))
694 (progn
695 (gnus-summary-copy-article nil group)
696 (setq deletep t))
697 (gnus-summary-move-article nil group)))))
698
699 ;; now delete the articles, if there was a copy done, and the
700 ;; backend allows it
701 (when (and deletep backend-supports-deletions)
702 (dolist (article tomove)
703 (gnus-summary-set-process-mark article))
704 (when tomove
705 (let ((gnus-novice-user nil)) ; don't ask me if I'm sure
706 (gnus-summary-delete-article nil))))
707
708 (gnus-summary-yank-process-mark))))
709
710 (defun spam-ham-copy-or-move-routine (copy groups)
711 (gnus-summary-kill-process-mark)
712 (let ((todo (spam-list-articles gnus-newsgroup-articles 'ham))
713 (backend-supports-deletions
714 (gnus-check-backend-function
715 'request-move-article gnus-newsgroup-name))
716 (respool-method (gnus-find-method-for-group gnus-newsgroup-name))
717 article mark todo deletep respool)
718
719 (when (member 'respool groups)
720 (setq respool t) ; boolean for later
721 (setq groups '("fake"))) ; when respooling, groups are dynamic so fake it
722
723 ;; now do the actual move
724 (dolist (group groups)
725 (when (and todo (stringp group))
726 (dolist (article todo)
727 (when spam-mark-ham-unread-before-move-from-spam-group
728 (gnus-summary-mark-article article gnus-unread-mark))
729 (gnus-summary-set-process-mark article))
730
731 (if respool ; respooling is with a "fake" group
732 (let ((spam-split-disabled
733 (or spam-split-disabled
734 spam-disable-spam-split-during-ham-respool)))
735 (gnus-summary-respool-article nil respool-method))
736 (if (or (not backend-supports-deletions) ; else, we are not respooling
737 (> (length groups) 1))
738 (progn ; if copying, copy and set deletep
739 (gnus-summary-copy-article nil group)
740 (setq deletep t))
741 (gnus-summary-move-article nil group))))) ; else move articles
742
743 ;; now delete the articles, unless a) copy is t, and there was a copy done
744 ;; b) a move was done to a single group
745 ;; c) backend-supports-deletions is nil
746 (unless copy
747 (when (and deletep backend-supports-deletions)
748 (dolist (article todo)
749 (gnus-summary-set-process-mark article))
750 (when todo
751 (let ((gnus-novice-user nil)) ; don't ask me if I'm sure
752 (gnus-summary-delete-article nil))))))
753
754 (gnus-summary-yank-process-mark))
755
756 (defun spam-ham-copy-routine (&rest groups)
757 (if (and (car-safe groups) (listp (car-safe groups)))
758 (apply 'spam-ham-copy-routine (car groups))
759 (spam-ham-copy-or-move-routine t groups)))
760
761 (defun spam-ham-move-routine (&rest groups)
762 (if (and (car-safe groups) (listp (car-safe groups)))
763 (apply 'spam-ham-move-routine (car groups))
764 (spam-ham-copy-or-move-routine nil groups)))
765
766 (eval-and-compile
767 (defalias 'spam-point-at-eol (if (fboundp 'point-at-eol)
768 'point-at-eol
769 'line-end-position)))
770
771 (defun spam-get-article-as-string (article)
772 (let ((article-buffer (spam-get-article-as-buffer article))
773 article-string)
774 (when article-buffer
775 (save-window-excursion
776 (set-buffer article-buffer)
777 (setq article-string (buffer-string))))
778 article-string))
779
780 (defun spam-get-article-as-buffer (article)
781 (let ((article-buffer))
782 (when (numberp article)
783 (save-window-excursion
784 (gnus-summary-goto-subject article)
785 (gnus-summary-show-article t)
786 (setq article-buffer (get-buffer gnus-article-buffer))))
787 article-buffer))
788
789 ;; disabled for now
790 ;; (defun spam-get-article-as-filename (article)
791 ;; (let ((article-filename))
792 ;; (when (numberp article)
793 ;; (nnml-possibly-change-directory
794 ;; (gnus-group-real-name gnus-newsgroup-name))
795 ;; (setq article-filename (expand-file-name
796 ;; (int-to-string article) nnml-current-directory)))
797 ;; (if (file-exists-p article-filename)
798 ;; article-filename
799 ;; nil)))
800
801 (defun spam-fetch-field-from-fast (article)
802 "Fetch the `from' field quickly, using the internal gnus-data-list function"
803 (if (and (numberp article)
804 (assoc article (gnus-data-list nil)))
805 (mail-header-from
806 (gnus-data-header (assoc article (gnus-data-list nil))))
807 nil))
808
809 (defun spam-fetch-field-subject-fast (article)
810 "Fetch the `subject' field quickly, using the internal
811 gnus-data-list function"
812 (if (and (numberp article)
813 (assoc article (gnus-data-list nil)))
814 (mail-header-subject
815 (gnus-data-header (assoc article (gnus-data-list nil))))
816 nil))
817
818 (defun spam-fetch-field-message-id-fast (article)
819 "Fetch the `Message-ID' field quickly, using the internal
820 gnus-data-list function"
821 (if (and (numberp article)
822 (assoc article (gnus-data-list nil)))
823 (mail-header-message-id
824 (gnus-data-header (assoc article (gnus-data-list nil))))
825 nil))
826
827 \f
828 ;;;; Spam determination.
829
830 (defvar spam-list-of-checks
831 '((spam-use-blacklist . spam-check-blacklist)
832 (spam-use-regex-headers . spam-check-regex-headers)
833 (spam-use-regex-body . spam-check-regex-body)
834 (spam-use-whitelist . spam-check-whitelist)
835 (spam-use-BBDB . spam-check-BBDB)
836 (spam-use-ifile . spam-check-ifile)
837 (spam-use-spamoracle . spam-check-spamoracle)
838 (spam-use-stat . spam-check-stat)
839 (spam-use-blackholes . spam-check-blackholes)
840 (spam-use-hashcash . spam-check-hashcash)
841 (spam-use-bogofilter-headers . spam-check-bogofilter-headers)
842 (spam-use-bogofilter . spam-check-bogofilter))
843 "The spam-list-of-checks list contains pairs associating a
844 parameter variable with a spam checking function. If the
845 parameter variable is true, then the checking function is called,
846 and its value decides what happens. Each individual check may
847 return nil, t, or a mailgroup name. The value nil means that the
848 check does not yield a decision, and so, that further checks are
849 needed. The value t means that the message is definitely not
850 spam, and that further spam checks should be inhibited.
851 Otherwise, a mailgroup name or the symbol 'spam (depending on
852 spam-split-symbolic-return) is returned where the mail should go,
853 and further checks are also inhibited. The usual mailgroup name
854 is the value of `spam-split-group', meaning that the message is
855 definitely a spam.")
856
857 (defvar spam-list-of-statistical-checks
858 '(spam-use-ifile
859 spam-use-regex-body
860 spam-use-stat
861 spam-use-bogofilter
862 spam-use-spamoracle)
863 "The spam-list-of-statistical-checks list contains all the mail
864 splitters that need to have the full message body available.")
865
866 ;;;TODO: modify to invoke self with each check if invoked without specifics
867 (defun spam-split (&rest specific-checks)
868 "Split this message into the `spam' group if it is spam.
869 This function can be used as an entry in the variable `nnmail-split-fancy',
870 for example like this: (: spam-split). It can take checks as
871 parameters. A string as a parameter will set the
872 spam-split-group to that string.
873
874 See the Info node `(gnus)Fancy Mail Splitting' for more details."
875 (interactive)
876 (setq spam-split-last-successful-check nil)
877 (unless spam-split-disabled
878 (let ((spam-split-group-choice spam-split-group))
879 (dolist (check specific-checks)
880 (when (stringp check)
881 (setq spam-split-group-choice check)
882 (setq specific-checks (delq check specific-checks))))
883
884 (let ((spam-split-group spam-split-group-choice))
885 (save-excursion
886 (save-restriction
887 (dolist (check spam-list-of-statistical-checks)
888 (when (and (symbolp check) (symbol-value check))
889 (widen)
890 (gnus-message 8 "spam-split: widening the buffer (%s requires it)"
891 (symbol-name check))
892 (return)))
893 ;; (progn (widen) (debug (buffer-string)))
894 (let ((list-of-checks spam-list-of-checks)
895 decision)
896 (while (and list-of-checks (not decision))
897 (let ((pair (pop list-of-checks)))
898 (when (and (symbol-value (car pair))
899 (or (null specific-checks)
900 (memq (car pair) specific-checks)))
901 (gnus-message 5 "spam-split: calling the %s function"
902 (symbol-name (cdr pair)))
903 (setq decision (funcall (cdr pair)))
904 ;; if we got a decision at all, save the current check
905 (when decision
906 (setq spam-split-last-successful-check (car pair)))
907
908 (when (eq decision 'spam)
909 (if spam-split-symbolic-return
910 (setq decision spam-split-group)
911 (gnus-error
912 5
913 (format "spam-split got %s but %s is nil"
914 (symbol-name decision)
915 (symbol-name spam-split-symbolic-return))))))))
916 (if (eq decision t)
917 (if spam-split-symbolic-return-positive 'ham nil)
918 decision))))))))
919
920 (defun spam-find-spam ()
921 "This function will detect spam in the current newsgroup using spam-split."
922 (interactive)
923
924 (let* ((group gnus-newsgroup-name)
925 (autodetect (gnus-parameter-spam-autodetect group))
926 (methods (gnus-parameter-spam-autodetect-methods group))
927 (first-method (nth 0 methods)))
928 (when (and autodetect
929 (not (equal first-method 'none)))
930 (mapcar
931 (lambda (article)
932 (let ((id (spam-fetch-field-message-id-fast article))
933 (subject (spam-fetch-field-subject-fast article))
934 (sender (spam-fetch-field-from-fast article)))
935 (unless (and spam-log-to-registry
936 (spam-log-registered-p id 'incoming))
937 (let* ((spam-split-symbolic-return t)
938 (spam-split-symbolic-return-positive t)
939 (split-return
940 (with-temp-buffer
941 (gnus-request-article-this-buffer
942 article
943 group)
944 (if (or (null first-method)
945 (equal first-method 'default))
946 (spam-split)
947 (apply 'spam-split methods)))))
948 (if (equal split-return 'spam)
949 (gnus-summary-mark-article article gnus-spam-mark))
950
951 (when (and split-return spam-log-to-registry)
952 (when (zerop (gnus-registry-group-count id))
953 (gnus-registry-add-group
954 id group subject sender))
955
956 (spam-log-processing-to-registry
957 id
958 'incoming
959 split-return
960 spam-split-last-successful-check
961 group))))))
962 (if spam-autodetect-recheck-messages
963 gnus-newsgroup-articles
964 gnus-newsgroup-unseen)))))
965
966 (defvar spam-registration-functions
967 ;; first the ham register, second the spam register function
968 ;; third the ham unregister, fourth the spam unregister function
969 '((spam-use-blacklist nil
970 spam-blacklist-register-routine
971 nil
972 spam-blacklist-unregister-routine)
973 (spam-use-whitelist spam-whitelist-register-routine
974 nil
975 spam-whitelist-unregister-routine
976 nil)
977 (spam-use-BBDB spam-BBDB-register-routine
978 nil
979 spam-BBDB-unregister-routine
980 nil)
981 (spam-use-ifile spam-ifile-register-ham-routine
982 spam-ifile-register-spam-routine
983 spam-ifile-unregister-ham-routine
984 spam-ifile-unregister-spam-routine)
985 (spam-use-spamoracle spam-spamoracle-learn-ham
986 spam-spamoracle-learn-spam
987 spam-spamoracle-unlearn-ham
988 spam-spamoracle-unlearn-spam)
989 (spam-use-stat spam-stat-register-ham-routine
990 spam-stat-register-spam-routine
991 spam-stat-unregister-ham-routine
992 spam-stat-unregister-spam-routine)
993 ;; note that spam-use-gmane is not a legitimate check
994 (spam-use-gmane nil
995 spam-report-gmane-register-routine
996 ;; does Gmane support unregistration?
997 nil
998 nil)
999 (spam-use-bogofilter spam-bogofilter-register-ham-routine
1000 spam-bogofilter-register-spam-routine
1001 spam-bogofilter-unregister-ham-routine
1002 spam-bogofilter-unregister-spam-routine))
1003 "The spam-registration-functions list contains pairs
1004 associating a parameter variable with the ham and spam
1005 registration functions, and the ham and spam unregistration
1006 functions")
1007
1008 (defun spam-classification-valid-p (classification)
1009 (or (eq classification 'spam)
1010 (eq classification 'ham)))
1011
1012 (defun spam-process-type-valid-p (process-type)
1013 (or (eq process-type 'incoming)
1014 (eq process-type 'process)))
1015
1016 (defun spam-registration-check-valid-p (check)
1017 (assoc check spam-registration-functions))
1018
1019 (defun spam-unregistration-check-valid-p (check)
1020 (assoc check spam-registration-functions))
1021
1022 (defun spam-registration-function (classification check)
1023 (let ((flist (cdr-safe (assoc check spam-registration-functions))))
1024 (if (eq classification 'spam)
1025 (nth 1 flist)
1026 (nth 0 flist))))
1027
1028 (defun spam-unregistration-function (classification check)
1029 (let ((flist (cdr-safe (assoc check spam-registration-functions))))
1030 (if (eq classification 'spam)
1031 (nth 3 flist)
1032 (nth 2 flist))))
1033
1034 (defun spam-list-articles (articles classification)
1035 (let ((mark-check (if (eq classification 'spam)
1036 'spam-group-spam-mark-p
1037 'spam-group-ham-mark-p))
1038 list mark-cache-yes mark-cache-no)
1039 (dolist (article articles)
1040 (let ((mark (gnus-summary-article-mark article)))
1041 (unless (memq mark mark-cache-no)
1042 (if (memq mark mark-cache-yes)
1043 (push article list)
1044 ;; else, we have to actually check the mark
1045 (if (funcall mark-check
1046 gnus-newsgroup-name
1047 mark)
1048 (progn
1049 (push article list)
1050 (push mark mark-cache-yes))
1051 (push mark mark-cache-no))))))
1052 list))
1053
1054 (defun spam-register-routine (classification
1055 check
1056 &optional unregister
1057 specific-articles)
1058 (when (and (spam-classification-valid-p classification)
1059 (spam-registration-check-valid-p check))
1060 (let* ((register-function
1061 (spam-registration-function classification check))
1062 (unregister-function
1063 (spam-unregistration-function classification check))
1064 (run-function (if unregister
1065 unregister-function
1066 register-function))
1067 (log-function (if unregister
1068 'spam-log-undo-registration
1069 'spam-log-processing-to-registry))
1070 article articles)
1071
1072 (when run-function
1073 ;; make list of articles, using specific-articles if given
1074 (setq articles (or specific-articles
1075 (spam-list-articles
1076 gnus-newsgroup-articles
1077 classification)))
1078 ;; process them
1079 (gnus-message 5 "%s %d %s articles with classification %s, check %s"
1080 (if unregister "Unregistering" "Registering")
1081 (length articles)
1082 (if specific-articles "specific" "")
1083 (symbol-name classification)
1084 (symbol-name check))
1085 (funcall run-function articles)
1086 ;; now log all the registrations (or undo them, depending on unregister)
1087 (dolist (article articles)
1088 (funcall log-function
1089 (spam-fetch-field-message-id-fast article)
1090 'process
1091 classification
1092 check
1093 gnus-newsgroup-name))))))
1094
1095 ;;; log a ham- or spam-processor invocation to the registry
1096 (defun spam-log-processing-to-registry (id type classification check group)
1097 (when spam-log-to-registry
1098 (if (and (stringp id)
1099 (stringp group)
1100 (spam-process-type-valid-p type)
1101 (spam-classification-valid-p classification)
1102 (spam-registration-check-valid-p check))
1103 (let ((cell-list (cdr-safe (gnus-registry-fetch-extra id type)))
1104 (cell (list classification check group)))
1105 (push cell cell-list)
1106 (gnus-registry-store-extra-entry
1107 id
1108 type
1109 cell-list))
1110
1111 (gnus-message 5 (format "%s called with bad ID, type, classification, check, or group"
1112 "spam-log-processing-to-registry")))))
1113
1114 ;;; check if a ham- or spam-processor registration has been done
1115 (defun spam-log-registered-p (id type)
1116 (when spam-log-to-registry
1117 (if (and (stringp id)
1118 (spam-process-type-valid-p type))
1119 (cdr-safe (gnus-registry-fetch-extra id type))
1120 (progn
1121 (gnus-message 5 (format "%s called with bad ID, type, classification, or check"
1122 "spam-log-registered-p"))
1123 nil))))
1124
1125 ;;; check if a ham- or spam-processor registration needs to be undone
1126 (defun spam-log-unregistration-needed-p (id type classification check)
1127 (when spam-log-to-registry
1128 (if (and (stringp id)
1129 (spam-process-type-valid-p type)
1130 (spam-classification-valid-p classification)
1131 (spam-registration-check-valid-p check))
1132 (let ((cell-list (cdr-safe (gnus-registry-fetch-extra id type)))
1133 found)
1134 (dolist (cell cell-list)
1135 (unless found
1136 (when (and (eq classification (nth 0 cell))
1137 (eq check (nth 1 cell)))
1138 (setq found t))))
1139 found)
1140 (progn
1141 (gnus-message 5 (format "%s called with bad ID, type, classification, or check"
1142 "spam-log-unregistration-needed-p"))
1143 nil))))
1144
1145
1146 ;;; undo a ham- or spam-processor registration (the group is not used)
1147 (defun spam-log-undo-registration (id type classification check &optional group)
1148 (when (and spam-log-to-registry
1149 (spam-log-unregistration-needed-p id type classification check))
1150 (if (and (stringp id)
1151 (spam-process-type-valid-p type)
1152 (spam-classification-valid-p classification)
1153 (spam-registration-check-valid-p check))
1154 (let ((cell-list (cdr-safe (gnus-registry-fetch-extra id type)))
1155 new-cell-list found)
1156 (dolist (cell cell-list)
1157 (unless (and (eq classification (nth 0 cell))
1158 (eq check (nth 1 cell)))
1159 (push cell new-cell-list)))
1160 (gnus-registry-store-extra-entry
1161 id
1162 type
1163 new-cell-list))
1164 (progn
1165 (gnus-message 5 (format "%s called with bad ID, type, check, or group"
1166 "spam-log-undo-registration"))
1167 nil))))
1168
1169 ;;; set up IMAP widening if it's necessary
1170 (defun spam-setup-widening ()
1171 (dolist (check spam-list-of-statistical-checks)
1172 (when (symbol-value check)
1173 (setq nnimap-split-download-body-default t))))
1174
1175 \f
1176 ;;;; Regex body
1177
1178 (defun spam-check-regex-body ()
1179 (let ((spam-regex-headers-ham spam-regex-body-ham)
1180 (spam-regex-headers-spam spam-regex-body-spam))
1181 (spam-check-regex-headers t)))
1182
1183 \f
1184 ;;;; Regex headers
1185
1186 (defun spam-check-regex-headers (&optional body)
1187 (let ((type (if body "body" "header"))
1188 (spam-split-group (if spam-split-symbolic-return
1189 'spam
1190 spam-split-group))
1191 ret found)
1192 (dolist (h-regex spam-regex-headers-ham)
1193 (unless found
1194 (goto-char (point-min))
1195 (when (re-search-forward h-regex nil t)
1196 (message "Ham regex %s search positive." type)
1197 (setq found t))))
1198 (dolist (s-regex spam-regex-headers-spam)
1199 (unless found
1200 (goto-char (point-min))
1201 (when (re-search-forward s-regex nil t)
1202 (message "Spam regex %s search positive." type)
1203 (setq found t)
1204 (setq ret spam-split-group))))
1205 ret))
1206
1207 \f
1208 ;;;; Blackholes.
1209
1210 (defun spam-reverse-ip-string (ip)
1211 (when (stringp ip)
1212 (mapconcat 'identity
1213 (nreverse (split-string ip "\\."))
1214 ".")))
1215
1216 (defun spam-check-blackholes ()
1217 "Check the Received headers for blackholed relays."
1218 (let ((headers (nnmail-fetch-field "received"))
1219 (spam-split-group (if spam-split-symbolic-return
1220 'spam
1221 spam-split-group))
1222 ips matches)
1223 (when headers
1224 (with-temp-buffer
1225 (insert headers)
1226 (goto-char (point-min))
1227 (gnus-message 5 "Checking headers for relay addresses")
1228 (while (re-search-forward
1229 "\\([0-9]+\\.[0-9]+\\.[0-9]+\\.[0-9]+\\)" nil t)
1230 (gnus-message 9 "Blackhole search found host IP %s." (match-string 1))
1231 (push (spam-reverse-ip-string (match-string 1))
1232 ips)))
1233 (dolist (server spam-blackhole-servers)
1234 (dolist (ip ips)
1235 (unless (and spam-blackhole-good-server-regex
1236 ;; match the good-server-regex against the reversed (again) IP string
1237 (string-match
1238 spam-blackhole-good-server-regex
1239 (spam-reverse-ip-string ip)))
1240 (unless matches
1241 (let ((query-string (concat ip "." server)))
1242 (if spam-use-dig
1243 (let ((query-result (query-dig query-string)))
1244 (when query-result
1245 (gnus-message 5 "(DIG): positive blackhole check '%s'"
1246 query-result)
1247 (push (list ip server query-result)
1248 matches)))
1249 ;; else, if not using dig.el
1250 (when (query-dns query-string)
1251 (gnus-message 5 "positive blackhole check")
1252 (push (list ip server (query-dns query-string 'TXT))
1253 matches)))))))))
1254 (when matches
1255 spam-split-group)))
1256 \f
1257 ;;;; Hashcash.
1258
1259 (eval-when-compile
1260 (autoload 'mail-check-payment "hashcash"))
1261
1262 (condition-case nil
1263 (progn
1264 (require 'hashcash)
1265
1266 (defun spam-check-hashcash ()
1267 "Check the headers for hashcash payments."
1268 (mail-check-payment))) ;mail-check-payment returns a boolean
1269
1270 (file-error))
1271 \f
1272 ;;;; BBDB
1273
1274 ;;; original idea for spam-check-BBDB from Alexander Kotelnikov
1275 ;;; <sacha@giotto.sj.ru>
1276
1277 ;; all this is done inside a condition-case to trap errors
1278
1279 (eval-when-compile
1280 (autoload 'bbdb-buffer "bbdb")
1281 (autoload 'bbdb-create-internal "bbdb")
1282 (autoload 'bbdb-search-simple "bbdb"))
1283
1284 (eval-and-compile
1285 (when (condition-case nil
1286 (progn
1287 (require 'bbdb)
1288 (require 'bbdb-com))
1289 (file-error
1290 (defalias 'spam-BBDB-register-routine 'ignore)
1291 (defalias 'spam-enter-ham-BBDB 'ignore)
1292 nil))
1293
1294 (defun spam-enter-ham-BBDB (addresses &optional remove)
1295 "Enter an address into the BBDB; implies ham (non-spam) sender"
1296 (dolist (from addresses)
1297 (when (stringp from)
1298 (let* ((parsed-address (gnus-extract-address-components from))
1299 (name (or (nth 0 parsed-address) "Ham Sender"))
1300 (remove-function (if remove
1301 'bbdb-delete-record-internal
1302 'ignore))
1303 (net-address (nth 1 parsed-address))
1304 (record (and net-address
1305 (bbdb-search-simple nil net-address))))
1306 (when net-address
1307 (gnus-message 5 "%s address %s %s BBDB"
1308 (if remove "Deleting" "Adding")
1309 from
1310 (if remove "from" "to"))
1311 (if record
1312 (funcall remove-function record)
1313 (bbdb-create-internal name nil net-address nil nil
1314 "ham sender added by spam.el")))))))
1315
1316 (defun spam-BBDB-register-routine (articles &optional unregister)
1317 (let (addresses)
1318 (dolist (article articles)
1319 (when (stringp (spam-fetch-field-from-fast article))
1320 (push (spam-fetch-field-from-fast article) addresses)))
1321 ;; now do the register/unregister action
1322 (spam-enter-ham-BBDB addresses unregister)))
1323
1324 (defun spam-BBDB-unregister-routine (articles)
1325 (spam-BBDB-register-routine articles t))
1326
1327 (defun spam-check-BBDB ()
1328 "Mail from people in the BBDB is classified as ham or non-spam"
1329 (let ((who (nnmail-fetch-field "from"))
1330 (spam-split-group (if spam-split-symbolic-return
1331 'spam
1332 spam-split-group)))
1333 (when who
1334 (setq who (nth 1 (gnus-extract-address-components who)))
1335 (if (bbdb-search-simple nil who)
1336 t
1337 (if spam-use-BBDB-exclusive
1338 spam-split-group
1339 nil)))))))
1340
1341 \f
1342 ;;;; ifile
1343
1344 ;;; check the ifile backend; return nil if the mail was NOT classified
1345 ;;; as spam
1346
1347 (defun spam-get-ifile-database-parameter ()
1348 "Get the command-line parameter for ifile's database from
1349 spam-ifile-database-path."
1350 (if spam-ifile-database-path
1351 (format "--db-file=%s" spam-ifile-database-path)
1352 nil))
1353
1354 (defun spam-check-ifile ()
1355 "Check the ifile backend for the classification of this message."
1356 (let ((article-buffer-name (buffer-name))
1357 (spam-split-group (if spam-split-symbolic-return
1358 'spam
1359 spam-split-group))
1360 category return)
1361 (with-temp-buffer
1362 (let ((temp-buffer-name (buffer-name))
1363 (db-param (spam-get-ifile-database-parameter)))
1364 (save-excursion
1365 (set-buffer article-buffer-name)
1366 (apply 'call-process-region
1367 (point-min) (point-max) spam-ifile-path
1368 nil temp-buffer-name nil "-c"
1369 (if db-param `(,db-param "-q") `("-q"))))
1370 ;; check the return now (we're back in the temp buffer)
1371 (goto-char (point-min))
1372 (if (not (eobp))
1373 (setq category (buffer-substring (point) (spam-point-at-eol))))
1374 (when (not (zerop (length category))) ; we need a category here
1375 (if spam-ifile-all-categories
1376 (setq return category)
1377 ;; else, if spam-ifile-all-categories is not set...
1378 (when (string-equal spam-ifile-spam-category category)
1379 (setq return spam-split-group)))))) ; note return is nil otherwise
1380 return))
1381
1382 (defun spam-ifile-register-with-ifile (articles category &optional unregister)
1383 "Register an article, given as a string, with a category.
1384 Uses `gnus-newsgroup-name' if category is nil (for ham registration)."
1385 (let ((category (or category gnus-newsgroup-name))
1386 (add-or-delete-option (if unregister "-d" "-i"))
1387 (db (spam-get-ifile-database-parameter))
1388 parameters)
1389 (with-temp-buffer
1390 (dolist (article articles)
1391 (let ((article-string (spam-get-article-as-string article)))
1392 (when (stringp article-string)
1393 (insert article-string))))
1394 (apply 'call-process-region
1395 (point-min) (point-max) spam-ifile-path
1396 nil nil nil
1397 add-or-delete-option category
1398 (if db `(,db "-h") `("-h"))))))
1399
1400 (defun spam-ifile-register-spam-routine (articles &optional unregister)
1401 (spam-ifile-register-with-ifile articles spam-ifile-spam-category unregister))
1402
1403 (defun spam-ifile-unregister-spam-routine (articles)
1404 (spam-ifile-register-spam-routine articles t))
1405
1406 (defun spam-ifile-register-ham-routine (articles &optional unregister)
1407 (spam-ifile-register-with-ifile articles spam-ifile-ham-category unregister))
1408
1409 (defun spam-ifile-unregister-ham-routine (articles)
1410 (spam-ifile-register-ham-routine articles t))
1411
1412 \f
1413 ;;;; spam-stat
1414
1415 (eval-when-compile
1416 (autoload 'spam-stat-buffer-change-to-non-spam "spam-stat")
1417 (autoload 'spam-stat-buffer-change-to-spam "spam-stat")
1418 (autoload 'spam-stat-buffer-is-non-spam "spam-stat")
1419 (autoload 'spam-stat-buffer-is-spam "spam-stat")
1420 (autoload 'spam-stat-load "spam-stat")
1421 (autoload 'spam-stat-save "spam-stat")
1422 (autoload 'spam-stat-split-fancy "spam-stat"))
1423
1424 (eval-and-compile
1425 (when (condition-case nil
1426 (let ((spam-stat-install-hooks nil))
1427 (require 'spam-stat))
1428 (file-error
1429 (defalias 'spam-stat-register-ham-routine 'ignore)
1430 (defalias 'spam-stat-register-spam-routine 'ignore)
1431 nil))
1432
1433 (defun spam-check-stat ()
1434 "Check the spam-stat backend for the classification of this message"
1435 (let ((spam-split-group (if spam-split-symbolic-return
1436 'spam
1437 spam-split-group))
1438 (spam-stat-split-fancy-spam-group spam-split-group) ; override
1439 (spam-stat-buffer (buffer-name)) ; stat the current buffer
1440 category return)
1441 (spam-stat-split-fancy)))
1442
1443 (defun spam-stat-register-spam-routine (articles &optional unregister)
1444 (dolist (article articles)
1445 (let ((article-string (spam-get-article-as-string article)))
1446 (with-temp-buffer
1447 (insert article-string)
1448 (if unregister
1449 (spam-stat-buffer-change-to-non-spam)
1450 (spam-stat-buffer-is-spam))))))
1451
1452 (defun spam-stat-unregister-spam-routine (articles)
1453 (spam-stat-register-spam-routine articles t))
1454
1455 (defun spam-stat-register-ham-routine (articles &optional unregister)
1456 (dolist (article articles)
1457 (let ((article-string (spam-get-article-as-string article)))
1458 (with-temp-buffer
1459 (insert article-string)
1460 (if unregister
1461 (spam-stat-buffer-change-to-spam)
1462 (spam-stat-buffer-is-non-spam))))))
1463
1464 (defun spam-stat-unregister-ham-routine (articles)
1465 (spam-stat-register-ham-routine articles t))
1466
1467 (defun spam-maybe-spam-stat-load ()
1468 (when spam-use-stat (spam-stat-load)))
1469
1470 (defun spam-maybe-spam-stat-save ()
1471 (when spam-use-stat (spam-stat-save)))))
1472
1473 \f
1474
1475 ;;;; Blacklists and whitelists.
1476
1477 (defvar spam-whitelist-cache nil)
1478 (defvar spam-blacklist-cache nil)
1479
1480 (defun spam-kill-whole-line ()
1481 (beginning-of-line)
1482 (let ((kill-whole-line t))
1483 (kill-line)))
1484
1485 ;;; address can be a list, too
1486 (defun spam-enter-whitelist (address &optional remove)
1487 "Enter ADDRESS (list or single) into the whitelist.
1488 With a non-nil REMOVE, remove them."
1489 (interactive "sAddress: ")
1490 (spam-enter-list address spam-whitelist remove)
1491 (setq spam-whitelist-cache nil))
1492
1493 ;;; address can be a list, too
1494 (defun spam-enter-blacklist (address &optional remove)
1495 "Enter ADDRESS (list or single) into the blacklist.
1496 With a non-nil REMOVE, remove them."
1497 (interactive "sAddress: ")
1498 (spam-enter-list address spam-blacklist remove)
1499 (setq spam-blacklist-cache nil))
1500
1501 (defun spam-enter-list (addresses file &optional remove)
1502 "Enter ADDRESSES into the given FILE.
1503 Either the whitelist or the blacklist files can be used. With
1504 REMOVE not nil, remove the ADDRESSES."
1505 (if (stringp addresses)
1506 (spam-enter-list (list addresses) file remove)
1507 ;; else, we have a list of addresses here
1508 (unless (file-exists-p (file-name-directory file))
1509 (make-directory (file-name-directory file) t))
1510 (save-excursion
1511 (set-buffer
1512 (find-file-noselect file))
1513 (dolist (a addresses)
1514 (when (stringp a)
1515 (goto-char (point-min))
1516 (if (re-search-forward (regexp-quote a) nil t)
1517 ;; found the address
1518 (when remove
1519 (spam-kill-whole-line))
1520 ;; else, the address was not found
1521 (unless remove
1522 (goto-char (point-max))
1523 (unless (bobp)
1524 (insert "\n"))
1525 (insert a "\n")))))
1526 (save-buffer))))
1527
1528 ;;; returns t if the sender is in the whitelist, nil or
1529 ;;; spam-split-group otherwise
1530 (defun spam-check-whitelist ()
1531 ;; FIXME! Should it detect when file timestamps change?
1532 (let ((spam-split-group (if spam-split-symbolic-return
1533 'spam
1534 spam-split-group)))
1535 (unless spam-whitelist-cache
1536 (setq spam-whitelist-cache (spam-parse-list spam-whitelist)))
1537 (if (spam-from-listed-p spam-whitelist-cache)
1538 t
1539 (if spam-use-whitelist-exclusive
1540 spam-split-group
1541 nil))))
1542
1543 (defun spam-check-blacklist ()
1544 ;; FIXME! Should it detect when file timestamps change?
1545 (let ((spam-split-group (if spam-split-symbolic-return
1546 'spam
1547 spam-split-group)))
1548 (unless spam-blacklist-cache
1549 (setq spam-blacklist-cache (spam-parse-list spam-blacklist)))
1550 (and (spam-from-listed-p spam-blacklist-cache) spam-split-group)))
1551
1552 (defun spam-parse-list (file)
1553 (when (file-readable-p file)
1554 (let (contents address)
1555 (with-temp-buffer
1556 (insert-file-contents file)
1557 (while (not (eobp))
1558 (setq address (buffer-substring (point) (spam-point-at-eol)))
1559 (forward-line 1)
1560 ;; insert the e-mail address if detected, otherwise the raw data
1561 (unless (zerop (length address))
1562 (let ((pure-address (nth 1 (gnus-extract-address-components address))))
1563 (push (or pure-address address) contents)))))
1564 (nreverse contents))))
1565
1566 (defun spam-from-listed-p (cache)
1567 (let ((from (nnmail-fetch-field "from"))
1568 found)
1569 (while cache
1570 (let ((address (pop cache)))
1571 (unless (zerop (length address)) ; 0 for a nil address too
1572 (setq address (regexp-quote address))
1573 ;; fix regexp-quote's treatment of user-intended regexes
1574 (while (string-match "\\\\\\*" address)
1575 (setq address (replace-match ".*" t t address))))
1576 (when (and address (string-match address from))
1577 (setq found t
1578 cache nil))))
1579 found))
1580
1581 (defun spam-filelist-register-routine (articles blacklist &optional unregister)
1582 (let ((de-symbol (if blacklist 'spam-use-whitelist 'spam-use-blacklist))
1583 (declassification (if blacklist 'ham 'spam))
1584 (enter-function
1585 (if blacklist 'spam-enter-blacklist 'spam-enter-whitelist))
1586 (remove-function
1587 (if blacklist 'spam-enter-whitelist 'spam-enter-blacklist))
1588 from addresses unregister-list)
1589 (dolist (article articles)
1590 (let ((from (spam-fetch-field-from-fast article))
1591 (id (spam-fetch-field-message-id-fast article))
1592 sender-ignored)
1593 (when (stringp from)
1594 (dolist (ignore-regex spam-blacklist-ignored-regexes)
1595 (when (and (not sender-ignored)
1596 (stringp ignore-regex)
1597 (string-match ignore-regex from))
1598 (setq sender-ignored t)))
1599 ;; remember the messages we need to unregister, unless remove is set
1600 (when (and
1601 (null unregister)
1602 (spam-log-unregistration-needed-p
1603 id 'process declassification de-symbol))
1604 (push from unregister-list))
1605 (unless sender-ignored
1606 (push from addresses)))))
1607
1608 (if unregister
1609 (funcall enter-function addresses t) ; unregister all these addresses
1610 ;; else, register normally and unregister what we need to
1611 (funcall remove-function unregister-list t)
1612 (dolist (article unregister-list)
1613 (spam-log-undo-registration
1614 (spam-fetch-field-message-id-fast article)
1615 'process
1616 declassification
1617 de-symbol))
1618 (funcall enter-function addresses nil))))
1619
1620 (defun spam-blacklist-unregister-routine (articles)
1621 (spam-blacklist-register-routine articles t))
1622
1623 (defun spam-blacklist-register-routine (articles &optional unregister)
1624 (spam-filelist-register-routine articles t unregister))
1625
1626 (defun spam-whitelist-unregister-routine (articles)
1627 (spam-whitelist-register-routine articles t))
1628
1629 (defun spam-whitelist-register-routine (articles &optional unregister)
1630 (spam-filelist-register-routine articles nil unregister))
1631
1632 \f
1633 ;;;; Spam-report glue
1634 (defun spam-report-gmane-register-routine (articles)
1635 (when articles
1636 (apply 'spam-report-gmane articles)))
1637
1638 \f
1639 ;;;; Bogofilter
1640 (defun spam-check-bogofilter-headers (&optional score)
1641 (let ((header (nnmail-fetch-field spam-bogofilter-header))
1642 (spam-split-group (if spam-split-symbolic-return
1643 'spam
1644 spam-split-group)))
1645 (when header ; return nil when no header
1646 (if score ; scoring mode
1647 (if (string-match "spamicity=\\([0-9.]+\\)" header)
1648 (match-string 1 header)
1649 "0")
1650 ;; spam detection mode
1651 (when (string-match spam-bogofilter-bogosity-positive-spam-header
1652 header)
1653 spam-split-group)))))
1654
1655 ;; return something sensible if the score can't be determined
1656 (defun spam-bogofilter-score ()
1657 "Get the Bogofilter spamicity score"
1658 (interactive)
1659 (save-window-excursion
1660 (gnus-summary-show-article t)
1661 (set-buffer gnus-article-buffer)
1662 (let ((score (or (spam-check-bogofilter-headers t)
1663 (spam-check-bogofilter t))))
1664 (message "Spamicity score %s" score)
1665 (or score "0"))
1666 (gnus-summary-show-article)))
1667
1668 (defun spam-check-bogofilter (&optional score)
1669 "Check the Bogofilter backend for the classification of this message"
1670 (let ((article-buffer-name (buffer-name))
1671 (db spam-bogofilter-database-directory)
1672 return)
1673 (with-temp-buffer
1674 (let ((temp-buffer-name (buffer-name)))
1675 (save-excursion
1676 (set-buffer article-buffer-name)
1677 (apply 'call-process-region
1678 (point-min) (point-max)
1679 spam-bogofilter-path
1680 nil temp-buffer-name nil
1681 (if db `("-d" ,db "-v") `("-v"))))
1682 (setq return (spam-check-bogofilter-headers score))))
1683 return))
1684
1685 (defun spam-bogofilter-register-with-bogofilter (articles
1686 spam
1687 &optional unregister)
1688 "Register an article, given as a string, as spam or non-spam."
1689 (dolist (article articles)
1690 (let ((article-string (spam-get-article-as-string article))
1691 (db spam-bogofilter-database-directory)
1692 (switch (if unregister
1693 (if spam
1694 spam-bogofilter-spam-strong-switch
1695 spam-bogofilter-ham-strong-switch)
1696 (if spam
1697 spam-bogofilter-spam-switch
1698 spam-bogofilter-ham-switch))))
1699 (when (stringp article-string)
1700 (with-temp-buffer
1701 (insert article-string)
1702
1703 (apply 'call-process-region
1704 (point-min) (point-max)
1705 spam-bogofilter-path
1706 nil nil nil switch
1707 (if db `("-d" ,db "-v") `("-v"))))))))
1708
1709 (defun spam-bogofilter-register-spam-routine (articles &optional unregister)
1710 (spam-bogofilter-register-with-bogofilter articles t unregister))
1711
1712 (defun spam-bogofilter-unregister-spam-routine (articles)
1713 (spam-bogofilter-register-spam-routine articles t))
1714
1715 (defun spam-bogofilter-register-ham-routine (articles &optional unregister)
1716 (spam-bogofilter-register-with-bogofilter articles nil unregister))
1717
1718 (defun spam-bogofilter-unregister-ham-routine (articles)
1719 (spam-bogofilter-register-ham-routine articles t))
1720
1721
1722 \f
1723 ;;;; spamoracle
1724 (defun spam-check-spamoracle ()
1725 "Run spamoracle on an article to determine whether it's spam."
1726 (let ((article-buffer-name (buffer-name))
1727 (spam-split-group (if spam-split-symbolic-return
1728 'spam
1729 spam-split-group)))
1730 (with-temp-buffer
1731 (let ((temp-buffer-name (buffer-name)))
1732 (save-excursion
1733 (set-buffer article-buffer-name)
1734 (let ((status
1735 (apply 'call-process-region
1736 (point-min) (point-max)
1737 spam-spamoracle-binary
1738 nil temp-buffer-name nil
1739 (if spam-spamoracle-database
1740 `("-f" ,spam-spamoracle-database "mark")
1741 '("mark")))))
1742 (if (eq 0 status)
1743 (progn
1744 (set-buffer temp-buffer-name)
1745 (goto-char (point-min))
1746 (when (re-search-forward "^X-Spam: yes;" nil t)
1747 spam-split-group))
1748 (error "Error running spamoracle: %s" status))))))))
1749
1750 (defun spam-spamoracle-learn (articles article-is-spam-p &optional unregister)
1751 "Run spamoracle in training mode."
1752 (with-temp-buffer
1753 (let ((temp-buffer-name (buffer-name)))
1754 (save-excursion
1755 (goto-char (point-min))
1756 (dolist (article articles)
1757 (insert (spam-get-article-as-string article)))
1758 (let* ((arg (if (spam-xor unregister article-is-spam-p)
1759 "-spam"
1760 "-good"))
1761 (status
1762 (apply 'call-process-region
1763 (point-min) (point-max)
1764 spam-spamoracle-binary
1765 nil temp-buffer-name nil
1766 (if spam-spamoracle-database
1767 `("-f" ,spam-spamoracle-database
1768 "add" ,arg)
1769 `("add" ,arg)))))
1770 (unless (eq 0 status)
1771 (error "Error running spamoracle: %s" status)))))))
1772
1773 (defun spam-spamoracle-learn-ham (articles &optional unregister)
1774 (spam-spamoracle-learn articles nil unregister))
1775
1776 (defun spam-spamoracle-unlearn-ham (articles &optional unregister)
1777 (spam-spamoracle-learn-ham articles t))
1778
1779 (defun spam-spamoracle-learn-spam (articles &optional unregister)
1780 (spam-spamoracle-learn articles t unregister))
1781
1782 (defun spam-spamoracle-unlearn-spam (articles &optional unregister)
1783 (spam-spamoracle-learn-spam articles t))
1784
1785 \f
1786 ;;;; Hooks
1787
1788 ;;;###autoload
1789 (defun spam-initialize ()
1790 "Install the spam.el hooks and do other initialization"
1791 (interactive)
1792 (setq spam-install-hooks t)
1793 ;; TODO: How do we redo this every time spam-face is customized?
1794 (push '((eq mark gnus-spam-mark) . spam-face)
1795 gnus-summary-highlight)
1796 ;; Add hooks for loading and saving the spam stats
1797 (add-hook 'gnus-save-newsrc-hook 'spam-maybe-spam-stat-save)
1798 (add-hook 'gnus-get-top-new-news-hook 'spam-maybe-spam-stat-load)
1799 (add-hook 'gnus-startup-hook 'spam-maybe-spam-stat-load)
1800 (add-hook 'gnus-summary-prepare-exit-hook 'spam-summary-prepare-exit)
1801 (add-hook 'gnus-summary-prepare-hook 'spam-summary-prepare)
1802 (add-hook 'gnus-get-new-news-hook 'spam-setup-widening)
1803 (add-hook 'gnus-summary-prepare-hook 'spam-find-spam))
1804
1805 (defun spam-unload-hook ()
1806 "Uninstall the spam.el hooks"
1807 (interactive)
1808 (remove-hook 'gnus-save-newsrc-hook 'spam-maybe-spam-stat-save)
1809 (remove-hook 'gnus-get-top-new-news-hook 'spam-maybe-spam-stat-load)
1810 (remove-hook 'gnus-startup-hook 'spam-maybe-spam-stat-load)
1811 (remove-hook 'gnus-summary-prepare-exit-hook 'spam-summary-prepare-exit)
1812 (remove-hook 'gnus-summary-prepare-hook 'spam-summary-prepare)
1813 (remove-hook 'gnus-get-new-news-hook 'spam-setup-widening)
1814 (remove-hook 'gnus-summary-prepare-hook 'spam-find-spam))
1815
1816 (add-hook 'spam-unload-hook 'spam-unload-hook)
1817
1818 (when spam-install-hooks
1819 (spam-initialize))
1820
1821 (provide 'spam)
1822
1823 ;;; arch-tag: 07e6e0ca-ab0a-4412-b445-1f6c72a4f27f
1824 ;;; spam.el ends here