| blob | 5a6d6f8f2433762d56dd406d010ce3c9876be9fa |
1 ;;; gnus-cite.el --- parse citations in articles for Gnus
3 ;; Copyright (C) 1995-2014 Free Software Foundation, Inc.
5 ;; Author: Per Abhiddenware
7 ;; This file is part of GNU Emacs.
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 3 of the License, or
12 ;; (at your option) any later version.
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.
19 ;; You should have received a copy of the GNU General Public License
20 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
22 ;;; Commentary:
24 ;;; Code:
36 ;;; Customization:
39 "Citation."
45 "Format of opened cited text buttons."
50 "Format of closed cited text buttons."
55 "The number of lines of hidden cited text to remain visible.
56 Or a pair (cons) of numbers which are the number of lines at the top
57 and bottom of the text, respectively, to remain visible."
60 integer
64 "Maximum article size (in bytes) where parsing citations is allowed.
65 Set it to nil to parse all articles."
68 integer))
71 "Maximum possible length for a citation prefix."
78 "*Regexp matching normal Supercite attribution lines.
79 The first grouping must match prefixes added by other packages."
84 "Regexp matching mangled Supercite attribution lines.
85 The first regexp group should match the Supercite attribution."
90 "Minimum number of identical prefixes before we believe it's a citation."
94 ;; Some Microsoft products put in a citation that extends to the
95 ;; remainder of the message:
96 ;;
97 ;; -----Original Message-----
98 ;; From: ...
99 ;; To: ...
100 ;; Sent: ... [date, in non-RFC-2822 format]
101 ;; Subject: ...
102 ;;
103 ;; Cited message, with no prefixes
104 ;;
105 ;; The four headers are always the same. But note they are prone to
106 ;; folding without additional indentation.
107 ;;
108 ;; Others use "----- Original Message -----" instead, and properly quote
109 ;; the body using "> ". This style is handled without special cases.
112 "In article\\|in <\\|On \\(Mon\\|Tue\\|Wed\\|Thu\\|Fri\\|Sat\\|Sun\\),\\|----- ?Original Message ?-----"
113 "*Regexp matching the beginning of an attribution line."
118 "\\(\\(wrote\\|writes\\|said\\|says\\|>\\)\\(:\\|\\.\\.\\.\\)\\|----- ?Original Message ?-----\\)[ \t]*$"
119 "*Regexp matching the end of an attribution line.
120 The text matching the first grouping will be used as a button."
126 "Regexp matching Microsoft-type rest-of-message citations."
133 Those lines may have been quoted by MTAs in order not to mix up with
134 the envelope From line."
140 "Face used for attribution lines."
142 ;; backward-compatibility alias
147 "Face used for attribution lines.
148 It is merged with the face for the cited text belonging to the attribution."
161 "Citation face."
163 ;; backward-compatibility alias
175 "Citation face."
177 ;; backward-compatibility alias
189 "Citation face."
191 ;; backward-compatibility alias
203 "Citation face."
205 ;; backward-compatibility alias
217 "Citation face."
219 ;; backward-compatibility alias
231 "Citation face."
233 ;; backward-compatibility alias
245 "Citation face."
247 ;; backward-compatibility alias
259 "Citation face."
261 ;; backward-compatibility alias
273 "Citation face."
275 ;; backward-compatibility alias
287 "Citation face."
289 ;; backward-compatibility alias
301 "Citation face."
303 ;; backward-compatibility alias
309 gnus-cite-7 gnus-cite-8 gnus-cite-9 gnus-cite-10 gnus-cite-11)
310 "*List of faces used for highlighting citations.
312 When there are citations from multiple articles in the same message,
313 Gnus will try to give each citation from each article its own face.
314 This should make it easier to see who wrote what."
329 list)
333 "Only hide excess citation if above this percentage of the body."
338 "Only hide excess citation if above this number of lines in the body."
343 "If non-nil, put a blank line between the citation header and the button."
347 ;; This has to go here because its default value depends on
348 ;; gnus-cite-face-list.
350 "List of faces that are not worth reading.
351 If an article has more pages below the one you are looking at, but
352 nothing on those pages is a word of at least three letters that is not
353 in a boring face, then the pages will be skipped."
357 ;;; Internal Variables:
363 ;; Alist of citation prefixes.
364 ;; The cdr is a list of lines with that prefix.
367 ;; Alist of attribution lines.
368 ;; The car is a line number.
369 ;; The cdr is the prefix for the citation started by that line.
372 ;; Alist of citation prefixes that have no matching attribution.
373 ;; The cdr is a list of lines with that prefix.
376 ;; Alist of attribution lines that have no matching citation.
377 ;; Each member has the form (WROTE IN PREFIX TAG), where
378 ;; WROTE: is the attribution line number
379 ;; IN: is the line number of the previous line if part of the same attribution,
380 ;; PREFIX: Is the citation prefix of the attribution line(s), and
381 ;; TAG: Is a Supercite tag, if any.
390 gnus-cited-opened-text-button-line-format-alist)
394 ;;; Commands:
397 "Highlight cited text.
398 Each citation in the article will be highlighted with a different face.
399 The faces are taken from `gnus-cite-face-list'.
400 Attribution lines are highlighted with the same face as the
401 corresponding citation merged with the face `gnus-cite-attribution'.
403 Text is considered cited if at least `gnus-cite-minimum-match-count'
404 lines matches `message-cite-prefix-regexp' with the same prefix.
406 Lines matching `gnus-cite-attribution-suffix' and perhaps
407 `gnus-cite-attribution-prefix' are considered attribution lines."
415 face entry prefix skip numbers number face-alist)
416 ;; Loop through citation prefixes.
431 ;; Loop through attribution lines.
440 ;; Add attribution button.
445 t)
448 ;; Highlight attribution line.
451 ;; Loop through attribution lines.
461 "Dissect the article buffer looking for cited text."
465 prefix numbers number marks m)
466 ;; Loop through citation prefixes.
481 ;; Skip to the beginning of the body.
484 ;; Find the end of the body.
488 ;; Sort the marks.
517 marks))))
523 "Do word wrapping in the current article.
524 If WIDTH (the numerical prefix), use that text width when
525 filling. If LONG-LINES, only fill sections that have lines
526 longer than the frame width."
544 use-hard-newlines)
547 ;; Note: the XEmacs version of `fill-region' inserts a newline
548 ;; unless the region ends with a newline.
565 ;; All this information is now incorrect.
567 gnus-cite-attribution-alist nil
568 gnus-cite-loose-prefix-alist nil
569 gnus-cite-loose-attribution-alist nil
570 gnus-cite-article nil)))))
574 columns)
591 ;; We know know that there are long lines here, but does this look
592 ;; like code? Check for ragged edges on the left.
596 "Toggle hiding of all cited text except attribution lines.
597 See the documentation for `gnus-article-highlight-citation'.
598 If given a negative prefix, always show; if given a positive prefix,
599 always hide."
605 marks
608 gnus-hidden-properties))
610 found beg end start)
613 'gnus-callback
625 end nil)
634 ;; Skip past lines we want to leave visible.
639 gnus-cited-lines-visible))
651 ;; We use markers for the end-points to facilitate later
652 ;; wrapping and mangling of text.
667 `gnus-article-toggle-cited-text
674 "Toggle hiding the text in REGION.
675 ARG can be nil or a number. Positive means hide, negative
676 means show, nil means toggle."
684 buffer-read-only)
691 ;; Can't remove 'cite from g-a-wash-types here because
692 ;; multiple citations may be hidden -jas
696 gnus-hidden-properties))))
699 'article-type nil beg end
701 gnus-hidden-properties))))
714 gnus-cited-opened-text-button-line-format-spec
715 gnus-cited-closed-text-button-line-format-spec))
717 `gnus-article-toggle-cited-text
718 args)
723 "Toggle hiding of cited text that has an attribution line.
724 If given a negative prefix, always show; if given a positive prefix,
725 always hide.
726 This will do nothing unless at least `gnus-cite-hide-percentage'
727 percent and at least `gnus-cite-hide-absolute' lines of the body is
728 cited text with attributions. When called interactively, these two
729 variables are ignored.
730 See also the documentation for `gnus-article-highlight-citation'."
743 total)
750 gnus-cite-prefix-alist))))
758 "Hide cited text in non-root articles."
766 ;;; Internal functions:
769 "Always parse the buffer."
771 ;;Reset parser information.
773 gnus-cite-attribution-alist nil
774 gnus-cite-loose-prefix-alist nil
775 gnus-cite-loose-attribution-alist nil)
778 ;; Parse if not too large.
781 ()
797 ;; Wrap chopped gnus-cite-parse.
808 ;; Parse and connect citation prefixes and attribution lines.
810 ;; Parse current buffer searching for citation prefixes.
818 alist entry start begin end numbers prefix guess-limit)
819 ;; Get all potential prefixes in `alist'.
821 ;; Each line.
825 start end)
827 ;; Ignore standard Supercite attribution prefix.
833 ;; Ignore very long prefixes.
836 ;; Ignore quoted envelope From_.
844 ;; Each prefix.
855 ;; Horrible special case for some Microsoft mailers.
859 ;; Assume this search ends up at the beginning of a line.
872 ;; Need to do 1+ because we're in the bol.
876 ;; We got all the potential prefixes. Now create
877 ;; `gnus-cite-prefix-alist' containing the oldest prefix for each
878 ;; line that appears at least `gnus-cite-minimum-match-count'
879 ;; times. First sort them by length. Longer is older.
888 ;; No lines with this prefix that wasn't also part of
889 ;; a longer prefix.
890 )
892 ;; Too few lines with this prefix. We keep it a bit
893 ;; longer in case it is an exact match for an attribution
894 ;; line, but we don't remove the line from other
895 ;; prefixes.
899 gnus-cite-prefix-alist)
900 ;; Remove articles from other prefixes.
902 current)
911 ;; Parse attributions
917 ;; Check previous line for an attribution leader.
929 t)
931 start t))
936 ;; don't add duplicates
939 end)))
942 gnus-cite-loose-attribution-alist)
946 ;; Connect attributions to citations
948 ;; No citations have been connected to attribution lines yet.
951 ;; Parse current buffer searching for attribution lines.
952 ;; Find exact supercite citations.
959 ;; Find loose supercite citations after attributions.
966 ;; Find loose supercite citations anywhere.
973 ;; Find nested citations after attributions.
977 ;; Find nested citations anywhere.
981 ;; Remove loose prefixes with too few lines.
983 entry)
990 gnus-cite-loose-prefix-alist
992 ;; Find flat attributions.
994 ;; Find any attributions (are we getting desperate yet?).
998 ;; Match all loose attributions and citations (SORT AFTER FUN) .
999 ;;
1000 ;; If SORT is `small', the citation with the shortest prefix will be
1001 ;; used, if it is `first' the first prefix will be used, if it is
1002 ;; `small-if-unique' the shortest prefix will be used if the
1003 ;; attribution line does not share its own prefix with other
1004 ;; loose attribution lines, otherwise the first prefix will be used.
1005 ;;
1006 ;; If AFTER is non-nil, only citations after the attribution line
1007 ;; will be considered.
1008 ;;
1009 ;; If FUN is non-nil, it will be called with the arguments (WROTE
1010 ;; PREFIX TAG) and expected to return a regular expression. Only
1011 ;; citations whose prefix matches the regular expression will be
1012 ;; considered.
1013 ;;
1014 ;; WROTE is the attribution line number.
1015 ;; PREFIX is the attribution line prefix.
1016 ;; TAG is the Supercite tag on the attribution line.
1019 att wrote in prefix tag regexp limit smallest best size)
1032 smallest 1000000
1033 best nil)
1035 cite candidate numbers first compare)
1044 regexp
1048 smallest compare))))
1050 ()
1059 current)
1066 ()
1074 ;; Return a list of loose attribution lines prefixed by PREFIX.
1076 att line lines)
1083 lines))
1086 ;; At line NUMBER, ignore PREFIX and add FACE to the rest of the line.
1089 from to overlay)
1100 gnus-cite-overlay-list)
1110 number)
1117 ;; Can't remove 'cite from g-a-wash-types here because
1118 ;; multiple citations may be hidden -jas
1120 gnus-hidden-properties))
1127 gnus-hidden-properties))))
1129 gnus-article-mime-handle-alist))
1133 ;; Return citation prefix for LINE.
1136 entry)
1142 prefix))
1145 "Make the citation variables local to the article buffer."
1147 gnus-cite-overlay-list gnus-cite-prefix-alist
1148 gnus-cite-attribution-alist gnus-cite-loose-prefix-alist
1149 gnus-cite-loose-attribution-alist)))
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 "Minor mode providing more font-lock support for nested citations.
1203 When enabled, it automatically turns on `font-lock-mode'."
1204 nil ;; init-value
1206 nil ;; keymap
1208 ;; FIXME: Use font-lock-add-keywords!
1211 font-lock-defaults)))
1212 default keywords)
1217 defaults
1220 ;; `gnus-message-citation-keywords' should be the last
1221 ;; elements of the keywords because the others are unlikely
1222 ;; to have the OVERRIDE flags -- XEmacs applies a keyword
1223 ;; having no OVERRIDE flag to matched text even if it has
1224 ;; already other faces, while Emacs doesn't.
1227 gnus-message-citation-keywords))
1229 ;; Force `font-lock-set-defaults' to update `font-lock-keywords'.
1234 font-lock-keywords nil))
1245 "Turn on `gnus-message-citation-mode'."
1248 "Turn off `gnus-message-citation-mode'."
1255 ;; Local Variables:
1256 ;; coding: utf-8
1257 ;; End:
1259 ;;; gnus-cite.el ends here