| blob | 7419cedac5f1f05955a67144b8e21f5c57af077e |
1 ;;; gnus-cite.el --- parse citations in articles for Gnus
3 ;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
4 ;; 2004, 2005, 2006, 2007, 2008, 2009, 2010 Free Software Foundation, Inc.
6 ;; Author: Per Abhiddenware
8 ;; This file is part of GNU Emacs.
10 ;; GNU Emacs is free software: you can redistribute it and/or modify
11 ;; it under the terms of the GNU General Public License as published by
12 ;; the Free Software Foundation, either version 3 of the License, or
13 ;; (at your option) any later version.
15 ;; GNU Emacs is distributed in the hope that it will be useful,
16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 ;; GNU General Public License for more details.
20 ;; You should have received a copy of the GNU General Public License
21 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
23 ;;; Commentary:
25 ;;; Code:
37 ;;; Customization:
40 "Citation."
46 "Format of opened cited text buttons."
51 "Format of closed cited text buttons."
56 "The number of lines of hidden cited text to remain visible.
57 Or a pair (cons) of numbers which are the number of lines at the top
58 and bottom of the text, respectively, to remain visible."
61 integer
65 "Maximum article size (in bytes) where parsing citations is allowed.
66 Set it to nil to parse all articles."
69 integer))
72 "Maximum possible length for a citation prefix."
79 "*Regexp matching normal Supercite attribution lines.
80 The first grouping must match prefixes added by other packages."
85 "Regexp matching mangled Supercite attribution lines.
86 The first regexp group should match the Supercite attribution."
91 "Minimum number of identical prefixes before we believe it's a citation."
95 ;; Some Microsoft products put in a citation that extends to the
96 ;; remainder of the message:
97 ;;
98 ;; -----Original Message-----
99 ;; From: ...
100 ;; To: ...
101 ;; Sent: ... [date, in non-RFC-2822 format]
102 ;; Subject: ...
103 ;;
104 ;; Cited message, with no prefixes
105 ;;
106 ;; The four headers are always the same. But note they are prone to
107 ;; folding without additional indentation.
108 ;;
109 ;; Others use "----- Original Message -----" instead, and properly quote
110 ;; the body using "> ". This style is handled without special cases.
113 "In article\\|in <\\|On \\(Mon\\|Tue\\|Wed\\|Thu\\|Fri\\|Sat\\|Sun\\),\\|----- ?Original Message ?-----"
114 "*Regexp matching the beginning of an attribution line."
119 "\\(\\(wrote\\|writes\\|said\\|says\\|>\\)\\(:\\|\\.\\.\\.\\)\\|----- ?Original Message ?-----\\)[ \t]*$"
120 "*Regexp matching the end of an attribution line.
121 The text matching the first grouping will be used as a button."
127 "Regexp matching Microsoft-type rest-of-message citations."
134 Those lines may have been quoted by MTAs in order not to mix up with
135 the envelope From line."
141 "Face used for attribution lines."
143 ;; backward-compatibility alias
148 "Face used for attribution lines.
149 It is merged with the face for the cited text belonging to the attribution."
162 "Citation face."
164 ;; backward-compatibility alias
176 "Citation face."
178 ;; backward-compatibility alias
190 "Citation face."
192 ;; backward-compatibility alias
204 "Citation face."
206 ;; backward-compatibility alias
218 "Citation face."
220 ;; backward-compatibility alias
232 "Citation face."
234 ;; backward-compatibility alias
246 "Citation face."
248 ;; backward-compatibility alias
260 "Citation face."
262 ;; backward-compatibility alias
274 "Citation face."
276 ;; backward-compatibility alias
288 "Citation face."
290 ;; backward-compatibility alias
302 "Citation face."
304 ;; backward-compatibility alias
310 gnus-cite-7 gnus-cite-8 gnus-cite-9 gnus-cite-10 gnus-cite-11)
311 "*List of faces used for highlighting citations.
313 When there are citations from multiple articles in the same message,
314 Gnus will try to give each citation from each article its own face.
315 This should make it easier to see who wrote what."
330 list)
334 "Only hide excess citation if above this percentage of the body."
339 "Only hide excess citation if above this number of lines in the body."
344 "If non-nil, put a blank line between the citation header and the button."
348 ;; This has to go here because its default value depends on
349 ;; gnus-cite-face-list.
351 "List of faces that are not worth reading.
352 If an article has more pages below the one you are looking at, but
353 nothing on those pages is a word of at least three letters that is not
354 in a boring face, then the pages will be skipped."
358 ;;; Internal Variables:
364 ;; Alist of citation prefixes.
365 ;; The cdr is a list of lines with that prefix.
368 ;; Alist of attribution lines.
369 ;; The car is a line number.
370 ;; The cdr is the prefix for the citation started by that line.
373 ;; Alist of citation prefixes that have no matching attribution.
374 ;; The cdr is a list of lines with that prefix.
377 ;; Alist of attribution lines that have no matching citation.
378 ;; Each member has the form (WROTE IN PREFIX TAG), where
379 ;; WROTE: is the attribution line number
380 ;; IN: is the line number of the previous line if part of the same attribution,
381 ;; PREFIX: Is the citation prefix of the attribution line(s), and
382 ;; TAG: Is a Supercite tag, if any.
391 gnus-cited-opened-text-button-line-format-alist)
395 ;;; Commands:
398 "Highlight cited text.
399 Each citation in the article will be highlighted with a different face.
400 The faces are taken from `gnus-cite-face-list'.
401 Attribution lines are highlighted with the same face as the
402 corresponding citation merged with the face `gnus-cite-attribution'.
404 Text is considered cited if at least `gnus-cite-minimum-match-count'
405 lines matches `message-cite-prefix-regexp' with the same prefix.
407 Lines matching `gnus-cite-attribution-suffix' and perhaps
408 `gnus-cite-attribution-prefix' are considered attribution lines."
416 face entry prefix skip numbers number face-alist)
417 ;; Loop through citation prefixes.
432 ;; Loop through attribution lines.
441 ;; Add attribution button.
446 t)
449 ;; Highlight attribution line.
452 ;; Loop through attribution lines.
462 "Dissect the article buffer looking for cited text."
466 prefix numbers number marks m)
467 ;; Loop through citation prefixes.
482 ;; Skip to the beginning of the body.
485 ;; Find the end of the body.
489 ;; Sort the marks.
517 marks))))
520 "Do word wrapping in the current article.
521 If WIDTH (the numerical prefix), use that text width when filling."
538 use-hard-newlines)
544 ;; All this information is now incorrect.
546 gnus-cite-attribution-alist nil
547 gnus-cite-loose-prefix-alist nil
548 gnus-cite-loose-attribution-alist nil
549 gnus-cite-article nil)))))
552 "Return true if the current line is long, and it's natural text."
556 ;; The line is long.
559 ;; It doesn't start with spaces.
561 ;; Not cited text.
563 citep)
570 "Toggle hiding of all cited text except attribution lines.
571 See the documentation for `gnus-article-highlight-citation'.
572 If given a negative prefix, always show; if given a positive prefix,
573 always hide."
579 marks
582 gnus-hidden-properties))
584 found beg end start)
587 'gnus-callback
599 end nil)
608 ;; Skip past lines we want to leave visible.
613 gnus-cited-lines-visible))
625 ;; We use markers for the end-points to facilitate later
626 ;; wrapping and mangling of text.
641 `gnus-article-toggle-cited-text
648 "Toggle hiding the text in REGION.
649 ARG can be nil or a number. Positive means hide, negative
650 means show, nil means toggle."
658 buffer-read-only)
665 ;; Can't remove 'cite from g-a-wash-types here because
666 ;; multiple citations may be hidden -jas
670 gnus-hidden-properties))))
673 'article-type nil beg end
675 gnus-hidden-properties))))
688 gnus-cited-opened-text-button-line-format-spec
689 gnus-cited-closed-text-button-line-format-spec))
691 `gnus-article-toggle-cited-text
692 args)
697 "Toggle hiding of cited text that has an attribution line.
698 If given a negative prefix, always show; if given a positive prefix,
699 always hide.
700 This will do nothing unless at least `gnus-cite-hide-percentage'
701 percent and at least `gnus-cite-hide-absolute' lines of the body is
702 cited text with attributions. When called interactively, these two
703 variables are ignored.
704 See also the documentation for `gnus-article-highlight-citation'."
717 total)
723 gnus-cite-prefix-alist))))
742 gnus-hidden-properties)))))))))
746 "Hide cited text in non-root articles."
754 ;;; Internal functions:
757 "Always parse the buffer."
759 ;;Reset parser information.
761 gnus-cite-attribution-alist nil
762 gnus-cite-loose-prefix-alist nil
763 gnus-cite-loose-attribution-alist nil)
766 ;; Parse if not too large.
769 ()
785 ;; Wrap chopped gnus-cite-parse.
796 ;; Parse and connect citation prefixes and attribution lines.
798 ;; Parse current buffer searching for citation prefixes.
806 alist entry start begin end numbers prefix guess-limit)
807 ;; Get all potential prefixes in `alist'.
809 ;; Each line.
813 start end)
815 ;; Ignore standard Supercite attribution prefix.
821 ;; Ignore very long prefixes.
824 ;; Ignore quoted envelope From_.
832 ;; Each prefix.
843 ;; Horrible special case for some Microsoft mailers.
847 ;; Assume this search ends up at the beginning of a line.
860 ;; Need to do 1+ because we're in the bol.
864 ;; We got all the potential prefixes. Now create
865 ;; `gnus-cite-prefix-alist' containing the oldest prefix for each
866 ;; line that appears at least `gnus-cite-minimum-match-count'
867 ;; times. First sort them by length. Longer is older.
876 ;; No lines with this prefix that wasn't also part of
877 ;; a longer prefix.
878 )
880 ;; Too few lines with this prefix. We keep it a bit
881 ;; longer in case it is an exact match for an attribution
882 ;; line, but we don't remove the line from other
883 ;; prefixes.
887 gnus-cite-prefix-alist)
888 ;; Remove articles from other prefixes.
890 current)
899 ;; Parse attributions
905 ;; Check previous line for an attribution leader.
917 t)
919 start t))
924 ;; don't add duplicates
927 end)))
930 gnus-cite-loose-attribution-alist)
934 ;; Connect attributions to citations
936 ;; No citations have been connected to attribution lines yet.
939 ;; Parse current buffer searching for attribution lines.
940 ;; Find exact supercite citations.
947 ;; Find loose supercite citations after attributions.
954 ;; Find loose supercite citations anywhere.
961 ;; Find nested citations after attributions.
965 ;; Find nested citations anywhere.
969 ;; Remove loose prefixes with too few lines.
971 entry)
978 gnus-cite-loose-prefix-alist
980 ;; Find flat attributions.
982 ;; Find any attributions (are we getting desperate yet?).
986 ;; Match all loose attributions and citations (SORT AFTER FUN) .
987 ;;
988 ;; If SORT is `small', the citation with the shortest prefix will be
989 ;; used, if it is `first' the first prefix will be used, if it is
990 ;; `small-if-unique' the shortest prefix will be used if the
991 ;; attribution line does not share its own prefix with other
992 ;; loose attribution lines, otherwise the first prefix will be used.
993 ;;
994 ;; If AFTER is non-nil, only citations after the attribution line
995 ;; will be considered.
996 ;;
997 ;; If FUN is non-nil, it will be called with the arguments (WROTE
998 ;; PREFIX TAG) and expected to return a regular expression. Only
999 ;; citations whose prefix matches the regular expression will be
1000 ;; considered.
1001 ;;
1002 ;; WROTE is the attribution line number.
1003 ;; PREFIX is the attribution line prefix.
1004 ;; TAG is the Supercite tag on the attribution line.
1007 att wrote in prefix tag regexp limit smallest best size)
1020 smallest 1000000
1021 best nil)
1023 cite candidate numbers first compare)
1032 regexp
1036 smallest compare))))
1038 ()
1047 current)
1054 ()
1062 ;; Return a list of loose attribution lines prefixed by PREFIX.
1064 att line lines)
1071 lines))
1074 ;; At line NUMBER, ignore PREFIX and add FACE to the rest of the line.
1077 from to overlay)
1088 gnus-cite-overlay-list)
1098 number)
1105 ;; Can't remove 'cite from g-a-wash-types here because
1106 ;; multiple citations may be hidden -jas
1108 gnus-hidden-properties))
1115 gnus-hidden-properties))))
1117 gnus-article-mime-handle-alist))
1121 ;; Return citation prefix for LINE.
1124 entry)
1130 prefix))
1133 "Make the citation variables local to the article buffer."
1135 gnus-cite-overlay-list gnus-cite-prefix-alist
1136 gnus-cite-attribution-alist gnus-cite-loose-prefix-alist
1137 gnus-cite-loose-attribution-alist)))
1142 "Say whether the current line is a cited line."
1148 prefix)
1150 found)))
1153 ;; Highlighting of different citation levels in message-mode.
1154 ;; - message-cite-prefix will be overridden if this is enabled.
1164 "Search for a cited line and set match data accordingly.
1165 Returns nil if there is no such line before LIMIT, t otherwise."
1171 gnus-message-max-citation-depth))
1180 t))
1183 ;; eval-when-compile ;; This breaks in XEmacs
1187 ;; (require 'gnus-cite)
1202 "Toggle `gnus-message-citation-mode' in current buffer.
1203 This buffer local minor mode provides additional font-lock support for
1204 nested citations.
1205 With prefix ARG, turn `gnus-message-citation-mode' on if and only if ARG
1206 is positive.
1207 Automatically turn `font-lock-mode' on when `gnus-message-citation-mode'
1208 is turned on."
1209 nil ;; init-value
1211 nil ;; keymap
1215 font-lock-defaults)))
1216 default keywords)
1221 defaults
1224 ;; `gnus-message-citation-keywords' should be the last
1225 ;; elements of the keywords because the others are unlikely
1226 ;; to have the OVERRIDE flags -- XEmacs applies a keyword
1227 ;; having no OVERRIDE flag to matched text even if it has
1228 ;; already other faces, while Emacs doesn't.
1231 gnus-message-citation-keywords))
1233 ;; Force `font-lock-set-defaults' to update `font-lock-keywords'.
1238 font-lock-keywords nil))
1247 "Turn on `gnus-message-citation-mode'."
1250 "Turn off `gnus-message-citation-mode'."
1257 ;; Local Variables:
1258 ;; coding: iso-8859-1
1259 ;; End:
1261 ;;; gnus-cite.el ends here