| blob | 3194e966f0f465a64d3c66f8818ffe81594feeb9 |
1 ;;; gnus-cite.el --- parse citations in articles for Gnus
3 ;; Copyright (C) 1995-2017 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:
33 ;;; Customization:
36 "Citation."
42 "Format of opened cited text buttons."
47 "Format of closed cited text buttons."
52 "The number of lines of hidden cited text to remain visible.
53 Or a pair (cons) of numbers which are the number of lines at the top
54 and bottom of the text, respectively, to remain visible."
57 integer
61 "Maximum article size (in bytes) where parsing citations is allowed.
62 Set it to nil to parse all articles."
65 integer))
68 "Maximum possible length for a citation prefix."
75 "Regexp matching normal Supercite attribution lines.
76 The first grouping must match prefixes added by other packages."
81 "Regexp matching mangled Supercite attribution lines.
82 The first regexp group should match the Supercite attribution."
87 "Minimum number of identical prefixes before we believe it's a citation."
91 ;; Some Microsoft products put in a citation that extends to the
92 ;; remainder of the message:
93 ;;
94 ;; -----Original Message-----
95 ;; From: ...
96 ;; To: ...
97 ;; Sent: ... [date, in non-RFC-2822 format]
98 ;; Subject: ...
99 ;;
100 ;; Cited message, with no prefixes
101 ;;
102 ;; The four headers are always the same. But note they are prone to
103 ;; folding without additional indentation.
104 ;;
105 ;; Others use "----- Original Message -----" instead, and properly quote
106 ;; the body using "> ". This style is handled without special cases.
109 "In article\\|in <\\|On \\(Mon\\|Tue\\|Wed\\|Thu\\|Fri\\|Sat\\|Sun\\),\\|----- ?Original Message ?-----"
110 "Regexp matching the beginning of an attribution line."
115 "\\(\\(wrote\\|writes\\|said\\|says\\|>\\)\\(:\\|\\.\\.\\.\\)\\|----- ?Original Message ?-----\\)[ \t]*$"
116 "Regexp matching the end of an attribution line.
117 The text matching the first grouping will be used as a button."
123 "Regexp matching Microsoft-type rest-of-message citations."
130 Those lines may have been quoted by MTAs in order not to mix up with
131 the envelope From line."
137 "Face used for attribution lines."
139 ;; backward-compatibility alias
144 "Face used for attribution lines.
145 It is merged with the face for the cited text belonging to the attribution."
158 "Citation face."
160 ;; backward-compatibility alias
172 "Citation face."
174 ;; backward-compatibility alias
186 "Citation face."
188 ;; backward-compatibility alias
200 "Citation face."
202 ;; backward-compatibility alias
214 "Citation face."
216 ;; backward-compatibility alias
228 "Citation face."
230 ;; backward-compatibility alias
242 "Citation face."
244 ;; backward-compatibility alias
256 "Citation face."
258 ;; backward-compatibility alias
270 "Citation face."
272 ;; backward-compatibility alias
284 "Citation face."
286 ;; backward-compatibility alias
298 "Citation face."
300 ;; backward-compatibility alias
306 gnus-cite-7 gnus-cite-8 gnus-cite-9 gnus-cite-10 gnus-cite-11)
307 "List of faces used for highlighting citations.
309 When there are citations from multiple articles in the same message,
310 Gnus will try to give each citation from each article its own face.
311 This should make it easier to see who wrote what."
326 list)
330 "Only hide excess citation if above this percentage of the body."
335 "Only hide excess citation if above this number of lines in the body."
340 "If non-nil, put a blank line between the citation header and the button."
344 ;; This has to go here because its default value depends on
345 ;; gnus-cite-face-list.
347 "List of faces that are not worth reading.
348 If an article has more pages below the one you are looking at, but
349 nothing on those pages is a word of at least three letters that is not
350 in a boring face, then the pages will be skipped."
354 ;;; Internal Variables:
360 ;; Alist of citation prefixes.
361 ;; The cdr is a list of lines with that prefix.
364 ;; Alist of attribution lines.
365 ;; The car is a line number.
366 ;; The cdr is the prefix for the citation started by that line.
369 ;; Alist of citation prefixes that have no matching attribution.
370 ;; The cdr is a list of lines with that prefix.
373 ;; Alist of attribution lines that have no matching citation.
374 ;; Each member has the form (WROTE IN PREFIX TAG), where
375 ;; WROTE: is the attribution line number
376 ;; IN: is the line number of the previous line if part of the same attribution,
377 ;; PREFIX: Is the citation prefix of the attribution line(s), and
378 ;; TAG: Is a Supercite tag, if any.
387 gnus-cited-opened-text-button-line-format-alist)
391 ;;; Commands:
394 "Highlight cited text.
395 Each citation in the article will be highlighted with a different face.
396 The faces are taken from `gnus-cite-face-list'.
397 Attribution lines are highlighted with the same face as the
398 corresponding citation merged with the face `gnus-cite-attribution'.
400 Text is considered cited if at least `gnus-cite-minimum-match-count'
401 lines matches `message-cite-prefix-regexp' with the same prefix.
403 Lines matching `gnus-cite-attribution-suffix' and perhaps
404 `gnus-cite-attribution-prefix' are considered attribution lines."
412 face entry prefix skip numbers number face-alist)
413 ;; Loop through citation prefixes.
428 ;; Loop through attribution lines.
437 ;; Add attribution button.
442 t)
445 ;; Highlight attribution line.
448 ;; Loop through attribution lines.
458 "Dissect the article buffer looking for cited text."
462 prefix numbers number marks m)
463 ;; Loop through citation prefixes.
478 ;; Skip to the beginning of the body.
481 ;; Find the end of the body.
485 ;; Sort the marks.
514 marks))))
520 "Do word wrapping in the current article.
521 If WIDTH (the numerical prefix), use that text width when
522 filling. If LONG-LINES, only fill sections that have lines
523 longer than the frame width."
540 use-hard-newlines)
543 ;; Note: the XEmacs version of `fill-region' inserts a newline
544 ;; unless the region ends with a newline.
561 ;; All this information is now incorrect.
563 gnus-cite-attribution-alist nil
564 gnus-cite-loose-prefix-alist nil
565 gnus-cite-loose-attribution-alist nil
566 gnus-cite-article nil)))))
570 columns)
587 ;; We know know that there are long lines here, but does this look
588 ;; like code? Check for ragged edges on the left.
592 "Toggle hiding of all cited text except attribution lines.
593 See the documentation for `gnus-article-highlight-citation'.
594 If given a negative prefix, always show; if given a positive prefix,
595 always hide."
601 marks
604 gnus-hidden-properties))
606 found beg end start)
609 'gnus-callback
621 end nil)
630 ;; Skip past lines we want to leave visible.
635 gnus-cited-lines-visible))
647 ;; We use markers for the end-points to facilitate later
648 ;; wrapping and mangling of text.
663 `gnus-article-toggle-cited-text
670 "Toggle hiding the text in REGION.
671 ARG can be nil or a number. Positive means hide, negative
672 means show, nil means toggle."
680 buffer-read-only)
687 ;; Can't remove 'cite from g-a-wash-types here because
688 ;; multiple citations may be hidden -jas
692 gnus-hidden-properties))))
695 'article-type nil beg end
697 gnus-hidden-properties))))
710 gnus-cited-opened-text-button-line-format-spec
711 gnus-cited-closed-text-button-line-format-spec))
713 `gnus-article-toggle-cited-text
714 args)
719 "Toggle hiding of cited text that has an attribution line.
720 If given a negative prefix, always show; if given a positive prefix,
721 always hide.
722 This will do nothing unless at least `gnus-cite-hide-percentage'
723 percent and at least `gnus-cite-hide-absolute' lines of the body is
724 cited text with attributions. When called interactively, these two
725 variables are ignored.
726 See also the documentation for `gnus-article-highlight-citation'."
739 total)
746 gnus-cite-prefix-alist))))
754 "Hide cited text in non-root articles."
762 ;;; Internal functions:
765 "Always parse the buffer."
767 ;;Reset parser information.
769 gnus-cite-attribution-alist nil
770 gnus-cite-loose-prefix-alist nil
771 gnus-cite-loose-attribution-alist nil)
774 ;; Parse if not too large.
777 ()
793 ;; Wrap chopped gnus-cite-parse.
804 ;; Parse and connect citation prefixes and attribution lines.
806 ;; Parse current buffer searching for citation prefixes.
814 alist entry start begin end numbers prefix guess-limit)
815 ;; Get all potential prefixes in `alist'.
817 ;; Each line.
821 start end)
823 ;; Ignore standard Supercite attribution prefix.
829 ;; Ignore very long prefixes.
832 ;; Ignore quoted envelope From_.
840 ;; Each prefix.
851 ;; Horrible special case for some Microsoft mailers.
855 ;; Assume this search ends up at the beginning of a line.
868 ;; Need to do 1+ because we're in the bol.
872 ;; We got all the potential prefixes. Now create
873 ;; `gnus-cite-prefix-alist' containing the oldest prefix for each
874 ;; line that appears at least `gnus-cite-minimum-match-count'
875 ;; times. First sort them by length. Longer is older.
884 ;; No lines with this prefix that wasn't also part of
885 ;; a longer prefix.
886 )
888 ;; Too few lines with this prefix. We keep it a bit
889 ;; longer in case it is an exact match for an attribution
890 ;; line, but we don't remove the line from other
891 ;; prefixes.
895 gnus-cite-prefix-alist)
896 ;; Remove articles from other prefixes.
898 current)
907 ;; Parse attributions
913 ;; Check previous line for an attribution leader.
925 t)
927 start t))
932 ;; don't add duplicates
935 end)))
938 gnus-cite-loose-attribution-alist)
942 ;; Connect attributions to citations
944 ;; No citations have been connected to attribution lines yet.
947 ;; Parse current buffer searching for attribution lines.
948 ;; Find exact supercite citations.
955 ;; Find loose supercite citations after attributions.
962 ;; Find loose supercite citations anywhere.
969 ;; Find nested citations after attributions.
973 ;; Find nested citations anywhere.
977 ;; Remove loose prefixes with too few lines.
979 entry)
986 gnus-cite-loose-prefix-alist
988 ;; Find flat attributions.
990 ;; Find any attributions (are we getting desperate yet?).
994 ;; Match all loose attributions and citations (SORT AFTER FUN) .
995 ;;
996 ;; If SORT is `small', the citation with the shortest prefix will be
997 ;; used, if it is `first' the first prefix will be used, if it is
998 ;; `small-if-unique' the shortest prefix will be used if the
999 ;; attribution line does not share its own prefix with other
1000 ;; loose attribution lines, otherwise the first prefix will be used.
1001 ;;
1002 ;; If AFTER is non-nil, only citations after the attribution line
1003 ;; will be considered.
1004 ;;
1005 ;; If FUN is non-nil, it will be called with the arguments (WROTE
1006 ;; PREFIX TAG) and expected to return a regular expression. Only
1007 ;; citations whose prefix matches the regular expression will be
1008 ;; considered.
1009 ;;
1010 ;; WROTE is the attribution line number.
1011 ;; PREFIX is the attribution line prefix.
1012 ;; TAG is the Supercite tag on the attribution line.
1015 att wrote in prefix tag regexp limit smallest best size)
1028 smallest 1000000
1029 best nil)
1031 cite candidate numbers first compare)
1040 regexp
1044 smallest compare))))
1046 ()
1055 current)
1062 ()
1070 ;; Return a list of loose attribution lines prefixed by PREFIX.
1072 att line lines)
1079 lines))
1082 ;; At line NUMBER, ignore PREFIX and add FACE to the rest of the line.
1085 from to overlay)
1096 gnus-cite-overlay-list)
1106 number)
1113 ;; Can't remove 'cite from g-a-wash-types here because
1114 ;; multiple citations may be hidden -jas
1116 gnus-hidden-properties))
1123 gnus-hidden-properties))))
1125 gnus-article-mime-handle-alist))
1129 ;; Return citation prefix for LINE.
1132 entry)
1138 prefix))
1141 "Make the citation variables local to the article buffer."
1143 gnus-cite-overlay-list gnus-cite-prefix-alist
1144 gnus-cite-attribution-alist gnus-cite-loose-prefix-alist
1145 gnus-cite-loose-attribution-alist)))
1149 ;; Highlighting of different citation levels in message-mode.
1150 ;; - message-cite-prefix will be overridden if this is enabled.
1160 "Search for a cited line and set match data accordingly.
1161 Returns nil if there is no such line before LIMIT, t otherwise."
1167 gnus-message-max-citation-depth))
1176 t))
1179 ;; eval-when-compile ;; This breaks in XEmacs
1183 ;; (require 'gnus-cite)
1196 "Minor mode providing more font-lock support for nested citations.
1197 When enabled, it automatically turns on `font-lock-mode'."
1198 nil ;; init-value
1200 nil ;; keymap
1202 ;; FIXME: Use font-lock-add-keywords!
1204 default keywords)
1209 defaults
1212 ;; `gnus-message-citation-keywords' should be the last
1213 ;; elements of the keywords because the others are unlikely
1214 ;; to have the OVERRIDE flags -- XEmacs applies a keyword
1215 ;; having no OVERRIDE flag to matched text even if it has
1216 ;; already other faces, while Emacs doesn't.
1219 gnus-message-citation-keywords))
1221 ;; Force `font-lock-set-defaults' to update `font-lock-keywords'.
1229 "Turn on `gnus-message-citation-mode'."
1232 "Turn off `gnus-message-citation-mode'."
1237 ;; Local Variables:
1238 ;; coding: utf-8
1239 ;; End:
1241 ;;; gnus-cite.el ends here