| blob | bbf9e527db717c1b86b28a07a4df534f611f60ee |
1 ;;; gnus-cite.el --- parse citations in articles for Gnus
3 ;; Copyright (C) 1995-2018 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 <https://www.gnu.org/licenses/>.
22 ;;; Commentary:
24 ;;; Code:
31 ;;; Customization:
34 "Citation."
40 "Format of opened cited text buttons."
45 "Format of closed cited text buttons."
50 "The number of lines of hidden cited text to remain visible.
51 Or a pair (cons) of numbers which are the number of lines at the top
52 and bottom of the text, respectively, to remain visible."
55 integer
59 "Maximum article size (in bytes) where parsing citations is allowed.
60 Set it to nil to parse all articles."
63 integer))
66 "Maximum possible length for a citation prefix."
73 "Regexp matching normal Supercite attribution lines.
74 The first grouping must match prefixes added by other packages."
79 "Regexp matching mangled Supercite attribution lines.
80 The first regexp group should match the Supercite attribution."
85 "Minimum number of identical prefixes before we believe it's a citation."
89 ;; Some Microsoft products put in a citation that extends to the
90 ;; remainder of the message:
91 ;;
92 ;; -----Original Message-----
93 ;; From: ...
94 ;; To: ...
95 ;; Sent: ... [date, in non-RFC-2822 format]
96 ;; Subject: ...
97 ;;
98 ;; Cited message, with no prefixes
99 ;;
100 ;; The four headers are always the same. But note they are prone to
101 ;; folding without additional indentation.
102 ;;
103 ;; Others use "----- Original Message -----" instead, and properly quote
104 ;; the body using "> ". This style is handled without special cases.
107 "In article\\|in <\\|On \\(Mon\\|Tue\\|Wed\\|Thu\\|Fri\\|Sat\\|Sun\\),\\|----- ?Original Message ?-----"
108 "Regexp matching the beginning of an attribution line."
113 "\\(\\(wrote\\|writes\\|said\\|says\\|>\\)\\(:\\|\\.\\.\\.\\)\\|----- ?Original Message ?-----\\)[ \t]*$"
114 "Regexp matching the end of an attribution line.
115 The text matching the first grouping will be used as a button."
121 "Regexp matching Microsoft-type rest-of-message citations."
128 Those lines may have been quoted by MTAs in order not to mix up with
129 the envelope From line."
135 "Face used for attribution lines."
139 "Face used for attribution lines.
140 It is merged with the face for the cited text belonging to the attribution."
153 "Citation face."
164 "Citation face."
175 "Citation face."
186 "Citation face."
197 "Citation face."
208 "Citation face."
219 "Citation face."
230 "Citation face."
241 "Citation face."
252 "Citation face."
263 "Citation face."
268 gnus-cite-7 gnus-cite-8 gnus-cite-9 gnus-cite-10 gnus-cite-11)
269 "List of faces used for highlighting citations.
271 When there are citations from multiple articles in the same message,
272 Gnus will try to give each citation from each article its own face.
273 This should make it easier to see who wrote what."
288 list)
292 "Only hide excess citation if above this percentage of the body."
297 "Only hide excess citation if above this number of lines in the body."
302 "If non-nil, put a blank line between the citation header and the button."
306 ;; This has to go here because its default value depends on
307 ;; gnus-cite-face-list.
309 "List of faces that are not worth reading.
310 If an article has more pages below the one you are looking at, but
311 nothing on those pages is a word of at least three letters that is not
312 in a boring face, then the pages will be skipped."
316 ;;; Internal Variables:
322 ;; Alist of citation prefixes.
323 ;; The cdr is a list of lines with that prefix.
326 ;; Alist of attribution lines.
327 ;; The car is a line number.
328 ;; The cdr is the prefix for the citation started by that line.
331 ;; Alist of citation prefixes that have no matching attribution.
332 ;; The cdr is a list of lines with that prefix.
335 ;; Alist of attribution lines that have no matching citation.
336 ;; Each member has the form (WROTE IN PREFIX TAG), where
337 ;; WROTE: is the attribution line number
338 ;; IN: is the line number of the previous line if part of the same attribution,
339 ;; PREFIX: Is the citation prefix of the attribution line(s), and
340 ;; TAG: Is a Supercite tag, if any.
349 gnus-cited-opened-text-button-line-format-alist)
353 ;;; Commands:
356 "Highlight cited text.
357 Each citation in the article will be highlighted with a different face.
358 The faces are taken from `gnus-cite-face-list'.
359 Attribution lines are highlighted with the same face as the
360 corresponding citation merged with the face `gnus-cite-attribution'.
362 Text is considered cited if at least `gnus-cite-minimum-match-count'
363 lines matches `message-cite-prefix-regexp' with the same prefix.
365 Lines matching `gnus-cite-attribution-suffix' and perhaps
366 `gnus-cite-attribution-prefix' are considered attribution lines."
374 face entry prefix skip numbers number face-alist)
375 ;; Loop through citation prefixes.
390 ;; Loop through attribution lines.
399 ;; Add attribution button.
404 t)
407 ;; Highlight attribution line.
410 ;; Loop through attribution lines.
420 "Dissect the article buffer looking for cited text."
424 prefix numbers number marks m)
425 ;; Loop through citation prefixes.
440 ;; Skip to the beginning of the body.
443 ;; Find the end of the body.
447 ;; Sort the marks.
476 marks))))
482 "Do word wrapping in the current article.
483 If WIDTH (the numerical prefix), use that text width when
484 filling.
486 If LONG-LINES, only fill sections that have lines longer than the
487 frame width.
489 Sections that are heuristically interpreted as not being
490 text (i.e., computer code and the like) will not be folded."
507 use-hard-newlines)
526 ;; All this information is now incorrect.
528 gnus-cite-attribution-alist nil
529 gnus-cite-loose-prefix-alist nil
530 gnus-cite-loose-attribution-alist nil
531 gnus-cite-article nil)))))
535 columns)
552 ;; We know know that there are long lines here, but does this look
553 ;; like code? Check for ragged edges on the left.
557 "Toggle hiding of all cited text except attribution lines.
558 See the documentation for `gnus-article-highlight-citation'.
559 If given a negative prefix, always show; if given a positive prefix,
560 always hide."
566 marks
569 gnus-hidden-properties))
571 found beg end start)
574 'gnus-callback
586 end nil)
595 ;; Skip past lines we want to leave visible.
600 gnus-cited-lines-visible))
612 ;; We use markers for the end-points to facilitate later
613 ;; wrapping and mangling of text.
628 `gnus-article-toggle-cited-text
635 "Toggle hiding the text in REGION.
636 ARG can be nil or a number. Positive means hide, negative
637 means show, nil means toggle."
645 buffer-read-only)
652 ;; Can't remove 'cite from g-a-wash-types here because
653 ;; multiple citations may be hidden -jas
657 gnus-hidden-properties))))
660 'article-type nil beg end
662 gnus-hidden-properties))))
675 gnus-cited-opened-text-button-line-format-spec
676 gnus-cited-closed-text-button-line-format-spec))
678 `gnus-article-toggle-cited-text
679 args)
684 "Toggle hiding of cited text that has an attribution line.
685 If given a negative prefix, always show; if given a positive prefix,
686 always hide.
687 This will do nothing unless at least `gnus-cite-hide-percentage'
688 percent and at least `gnus-cite-hide-absolute' lines of the body is
689 cited text with attributions. When called interactively, these two
690 variables are ignored.
691 See also the documentation for `gnus-article-highlight-citation'."
704 total)
711 gnus-cite-prefix-alist))))
719 "Hide cited text in non-root articles."
727 ;;; Internal functions:
730 "Always parse the buffer."
732 ;;Reset parser information.
734 gnus-cite-attribution-alist nil
735 gnus-cite-loose-prefix-alist nil
736 gnus-cite-loose-attribution-alist nil)
739 ;; Parse if not too large.
742 ()
758 ;; Wrap chopped gnus-cite-parse.
769 ;; Parse and connect citation prefixes and attribution lines.
771 ;; Parse current buffer searching for citation prefixes.
779 alist entry start begin end numbers prefix guess-limit)
780 ;; Get all potential prefixes in `alist'.
782 ;; Each line.
786 start end)
788 ;; Ignore standard Supercite attribution prefix.
794 ;; Ignore very long prefixes.
797 ;; Ignore quoted envelope From_.
805 ;; Each prefix.
816 ;; Horrible special case for some Microsoft mailers.
820 ;; Assume this search ends up at the beginning of a line.
833 ;; Need to do 1+ because we're in the bol.
837 ;; We got all the potential prefixes. Now create
838 ;; `gnus-cite-prefix-alist' containing the oldest prefix for each
839 ;; line that appears at least `gnus-cite-minimum-match-count'
840 ;; times. First sort them by length. Longer is older.
849 ;; No lines with this prefix that wasn't also part of
850 ;; a longer prefix.
851 )
853 ;; Too few lines with this prefix. We keep it a bit
854 ;; longer in case it is an exact match for an attribution
855 ;; line, but we don't remove the line from other
856 ;; prefixes.
860 gnus-cite-prefix-alist)
861 ;; Remove articles from other prefixes.
863 current)
872 ;; Parse attributions
878 ;; Check previous line for an attribution leader.
890 t)
892 start t))
897 ;; don't add duplicates
900 end)))
903 gnus-cite-loose-attribution-alist)
907 ;; Connect attributions to citations
909 ;; No citations have been connected to attribution lines yet.
912 ;; Parse current buffer searching for attribution lines.
913 ;; Find exact supercite citations.
920 ;; Find loose supercite citations after attributions.
927 ;; Find loose supercite citations anywhere.
934 ;; Find nested citations after attributions.
938 ;; Find nested citations anywhere.
942 ;; Remove loose prefixes with too few lines.
944 entry)
951 gnus-cite-loose-prefix-alist
953 ;; Find flat attributions.
955 ;; Find any attributions (are we getting desperate yet?).
959 ;; Match all loose attributions and citations (SORT AFTER FUN) .
960 ;;
961 ;; If SORT is `small', the citation with the shortest prefix will be
962 ;; used, if it is `first' the first prefix will be used, if it is
963 ;; `small-if-unique' the shortest prefix will be used if the
964 ;; attribution line does not share its own prefix with other
965 ;; loose attribution lines, otherwise the first prefix will be used.
966 ;;
967 ;; If AFTER is non-nil, only citations after the attribution line
968 ;; will be considered.
969 ;;
970 ;; If FUN is non-nil, it will be called with the arguments (WROTE
971 ;; PREFIX TAG) and expected to return a regular expression. Only
972 ;; citations whose prefix matches the regular expression will be
973 ;; considered.
974 ;;
975 ;; WROTE is the attribution line number.
976 ;; PREFIX is the attribution line prefix.
977 ;; TAG is the Supercite tag on the attribution line.
980 att wrote in prefix tag regexp limit smallest best size)
993 smallest 1000000
994 best nil)
996 cite candidate numbers first compare)
1005 regexp
1009 smallest compare))))
1011 ()
1020 current)
1027 ()
1035 ;; Return a list of loose attribution lines prefixed by PREFIX.
1037 att line lines)
1044 lines))
1047 ;; At line NUMBER, ignore PREFIX and add FACE to the rest of the line.
1050 from to overlay)
1061 gnus-cite-overlay-list)
1071 number)
1078 ;; Can't remove 'cite from g-a-wash-types here because
1079 ;; multiple citations may be hidden -jas
1081 gnus-hidden-properties))
1088 gnus-hidden-properties))))
1090 gnus-article-mime-handle-alist))
1094 ;; Return citation prefix for LINE.
1097 entry)
1103 prefix))
1106 "Make the citation variables local to the article buffer."
1108 gnus-cite-overlay-list gnus-cite-prefix-alist
1109 gnus-cite-attribution-alist gnus-cite-loose-prefix-alist
1110 gnus-cite-loose-attribution-alist)))
1114 ;; Highlighting of different citation levels in message-mode.
1115 ;; - message-cite-prefix will be overridden if this is enabled.
1125 "Search for a cited line and set match data accordingly.
1126 Returns nil if there is no such line before LIMIT, t otherwise."
1132 gnus-message-max-citation-depth))
1141 t))
1144 ;; eval-when-compile ;; This breaks in XEmacs
1148 ;; (require 'gnus-cite)
1161 "Minor mode providing more font-lock support for nested citations.
1162 When enabled, it automatically turns on `font-lock-mode'."
1163 nil ;; init-value
1165 nil ;; keymap
1167 ;; FIXME: Use font-lock-add-keywords!
1169 default keywords)
1174 defaults
1177 ;; `gnus-message-citation-keywords' should be the last
1178 ;; elements of the keywords because the others are unlikely
1179 ;; to have the OVERRIDE flags -- XEmacs applies a keyword
1180 ;; having no OVERRIDE flag to matched text even if it has
1181 ;; already other faces, while Emacs doesn't.
1184 gnus-message-citation-keywords))
1186 ;; Force `font-lock-set-defaults' to update `font-lock-keywords'.
1194 "Turn on `gnus-message-citation-mode'."
1197 "Turn off `gnus-message-citation-mode'."
1202 ;; Local Variables:
1203 ;; coding: utf-8
1204 ;; End:
1206 ;;; gnus-cite.el ends here