1 ;;; paragraphs.el --- paragraph and sentence parsing
3 ;; Copyright (C) 1985-1987, 1991, 1994-1997, 1999-2015 Free Software
6 ;; Maintainer: emacs-devel@gnu.org
10 ;; This file is part of GNU Emacs.
12 ;; GNU Emacs is free software: you can redistribute it and/or modify
13 ;; it under the terms of the GNU General Public License as published by
14 ;; the Free Software Foundation, either version 3 of the License, or
15 ;; (at your option) any later version.
17 ;; GNU Emacs is distributed in the hope that it will be useful,
18 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
19 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 ;; GNU General Public License for more details.
22 ;; You should have received a copy of the GNU General Public License
23 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
27 ;; This package provides the paragraph-oriented commands documented in the
32 (defgroup paragraphs nil
33 "Paragraph and sentence parsing."
36 (put 'use-hard-newlines
'permanent-local t
)
37 (define-minor-mode use-hard-newlines
38 "Toggle distinguishing between hard and soft newlines.
39 With a prefix argument ARG, enable the feature if ARG is
40 positive, and disable it otherwise. If called from Lisp, enable
41 it if ARG is omitted or nil.
43 When enabled, the functions `newline' and `open-line' add the
44 text-property `hard' to newlines that they insert, and a line is
45 only considered as a candidate to match `paragraph-start' or
46 `paragraph-separate' if it follows a hard newline.
48 When enabling, if there are newlines in the buffer but no hard
49 newlines, ask the user whether to mark as hard any newlines
50 preceding a `paragraph-start' line. From a program, second arg
51 INSERT specifies whether to do this; it can be `never' to change
52 nothing, t or `always' to force marking, `guess' to try to do the
53 right thing with no questions, nil or anything else to ask the
56 Newlines not marked hard are called \"soft\", and are always internal
57 to paragraphs. The fill functions insert and delete only soft newlines."
60 (when use-hard-newlines
62 ;; Intuit hard newlines --
63 ;; mark as hard any newlines preceding a paragraph-start line.
64 (if (or (eq insert t
) (eq insert
'always
)
65 (and (not (eq 'never insert
))
66 (not (text-property-any (point-min) (point-max) 'hard t
))
68 (goto-char (point-min))
69 (search-forward "\n" nil t
))
70 (or (eq insert
'guess
)
71 (y-or-n-p "Make newlines between paragraphs hard? "))))
73 (goto-char (point-min))
74 (while (search-forward "\n" nil t
)
77 (when (looking-at paragraph-start
)
78 (set-hard-newline-properties (1- pos
) pos
))
79 ;; If paragraph-separate, newline after it is hard too.
80 (when (looking-at paragraph-separate
)
81 (set-hard-newline-properties (1- pos
) pos
)
84 (set-hard-newline-properties (point) (1+ (point)))))))))))
86 (defcustom paragraph-start
"\f\\|[ \t]*$" "\
87 Regexp for beginning of a line that starts OR separates paragraphs.
88 This regexp should match lines that separate paragraphs
89 and should also match lines that start a paragraph
90 \(and are part of that paragraph).
92 This is matched against the text at the left margin, which is not necessarily
93 the beginning of the line, so it should never use \"^\" as an anchor. This
94 ensures that the paragraph functions will work equally well within a region
95 of text indented by a margin setting.
97 The variable `paragraph-separate' specifies how to distinguish
98 lines that start paragraphs from lines that separate them.
100 If the variable `use-hard-newlines' is non-nil, then only lines following a
101 hard newline are considered to match."
104 (put 'paragraph-start
'safe-local-variable
'stringp
)
106 ;; paragraph-start requires a hard newline, but paragraph-separate does not:
107 ;; It is assumed that paragraph-separate is distinctive enough to be believed
108 ;; whenever it occurs, while it is reasonable to set paragraph-start to
109 ;; something very minimal, even including "." (which makes every hard newline
110 ;; start a new paragraph).
112 (defcustom paragraph-separate
"[ \t\f]*$"
113 "Regexp for beginning of a line that separates paragraphs.
114 If you change this, you may have to change `paragraph-start' also.
116 This is matched against the text at the left margin, which is not necessarily
117 the beginning of the line, so it should not use \"^\" as an anchor. This
118 ensures that the paragraph functions will work equally within a region of
119 text indented by a margin setting."
122 (put 'paragraph-separate
'safe-local-variable
'stringp
)
124 (defcustom sentence-end-double-space t
125 "Non-nil means a single space does not end a sentence.
126 This is relevant for filling. See also `sentence-end-without-period'
127 and `colon-double-space'.
129 This value is used by the function `sentence-end' to construct the
130 regexp describing the end of a sentence, when the value of the variable
131 `sentence-end' is nil. See Info node `(elisp)Standard Regexps'."
134 (put 'sentence-end-double-space
'safe-local-variable
'booleanp
)
136 (defcustom sentence-end-without-period nil
137 "Non-nil means a sentence will end without a period.
138 For example, a sentence in Thai text ends with double space but
141 This value is used by the function `sentence-end' to construct the
142 regexp describing the end of a sentence, when the value of the variable
143 `sentence-end' is nil. See Info node `(elisp)Standard Regexps'."
146 (put 'sentence-end-without-period
'safe-local-variable
'booleanp
)
148 (defcustom sentence-end-without-space
150 "String of characters that end sentence without following spaces.
152 This value is used by the function `sentence-end' to construct the
153 regexp describing the end of a sentence, when the value of the variable
154 `sentence-end' is nil. See Info node `(elisp)Standard Regexps'."
157 (put 'sentence-end-without-space
'safe-local-variable
'stringp
)
159 (defcustom sentence-end nil
160 "Regexp describing the end of a sentence.
161 The value includes the whitespace following the sentence.
162 All paragraph boundaries also end sentences, regardless.
164 The value nil means to use the default value defined by the
165 function `sentence-end'. You should always use this function
166 to obtain the value of this variable."
168 :type
'(choice regexp
(const :tag
"Use default value" nil
)))
169 (put 'sentence-end
'safe-local-variable
'string-or-null-p
)
171 (defcustom sentence-end-base
"[.?!…‽][]\"'”’)}]*"
172 "Regexp matching the basic end of a sentence, not including following space."
176 (put 'sentence-end-base
'safe-local-variable
'stringp
)
178 (defun sentence-end ()
179 "Return the regexp describing the end of a sentence.
181 This function returns either the value of the variable `sentence-end'
182 if it is non-nil, or the default value constructed from the
183 variables `sentence-end-base', `sentence-end-double-space',
184 `sentence-end-without-period' and `sentence-end-without-space'.
186 The default value specifies that in order to be recognized as the
187 end of a sentence, the ending period, question mark, or exclamation point
188 must be followed by two spaces, with perhaps some closing delimiters
189 in between. See Info node `(elisp)Standard Regexps'."
191 ;; We accept non-break space along with space.
192 (concat (if sentence-end-without-period
"\\w[ \u00a0][ \u00a0]\\|")
195 (if sentence-end-double-space
196 "\\($\\|[ \u00a0]$\\|\t\\|[ \u00a0][ \u00a0]\\)" "\\($\\|[\t \u00a0]\\)")
197 "\\|[" sentence-end-without-space
"]+"
201 (defcustom page-delimiter
"^\014"
202 "Regexp describing line-beginnings that separate pages."
205 (put 'page-delimiter
'safe-local-variable
'stringp
)
207 (defcustom paragraph-ignore-fill-prefix nil
208 "Non-nil means the paragraph commands are not affected by `fill-prefix'.
209 This is desirable in modes where blank lines are the paragraph delimiters."
212 (put 'paragraph-ignore-fill-prefix
'safe-local-variable
'booleanp
)
214 (defun forward-paragraph (&optional arg
)
215 "Move forward to end of paragraph.
216 With argument ARG, do it ARG times;
217 a negative argument ARG = -N means move backward N paragraphs.
219 A line which `paragraph-start' matches either separates paragraphs
220 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
221 A paragraph end is the beginning of a line which is not part of the paragraph
222 to which the end of the previous line belongs, or the end of the buffer.
223 Returns the count of paragraphs left to move."
225 (or arg
(setq arg
1))
226 (let* ((opoint (point))
228 (and fill-prefix
(not (equal fill-prefix
""))
229 (not paragraph-ignore-fill-prefix
)
230 (regexp-quote fill-prefix
)))
231 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
232 ;; These regexps shouldn't be anchored, because we look for them
233 ;; starting at the left-margin. This allows paragraph commands to
234 ;; work normally with indented text.
235 ;; This hack will not find problem cases like "whatever\\|^something".
236 (parstart (if (and (not (equal "" paragraph-start
))
237 (equal ?^
(aref paragraph-start
0)))
238 (substring paragraph-start
1)
240 (parsep (if (and (not (equal "" paragraph-separate
))
241 (equal ?^
(aref paragraph-separate
0)))
242 (substring paragraph-separate
1)
245 (if fill-prefix-regexp
247 fill-prefix-regexp
"[ \t]*$")
249 ;; This is used for searching.
250 (sp-parstart (concat "^[ \t]*\\(?:" parstart
"\\|" parsep
"\\)"))
252 (while (and (< arg
0) (not (bobp)))
253 (if (and (not (looking-at parsep
))
254 (re-search-backward "^\n" (max (1- (point)) (point-min)) t
)
258 ;; Move back over paragraph-separating lines.
259 (forward-char -
1) (beginning-of-line)
260 (while (and (not (bobp))
261 (progn (move-to-left-margin)
262 (looking-at parsep
)))
267 ;; Go to end of the previous (non-separating) line.
269 ;; Search back for line that starts or separates paragraphs.
270 (if (if fill-prefix-regexp
271 ;; There is a fill prefix; it overrides parstart.
272 (let (multiple-lines)
273 (while (and (progn (beginning-of-line) (not (bobp)))
274 (progn (move-to-left-margin)
275 (not (looking-at parsep
)))
276 (looking-at fill-prefix-regexp
))
277 (unless (= (point) start
)
278 (setq multiple-lines t
))
280 (move-to-left-margin)
281 ;; This deleted code caused a long hanging-indent line
282 ;; not to be filled together with the following lines.
283 ;; ;; Don't move back over a line before the paragraph
284 ;; ;; which doesn't start with fill-prefix
285 ;; ;; unless that is the only line we've moved over.
286 ;; (and (not (looking-at fill-prefix-regexp))
290 (while (and (re-search-backward sp-parstart nil
1)
292 ;; Found a candidate, but need to check if it is a
294 (progn (setq start
(point))
295 (move-to-left-margin)
296 (not (looking-at parsep
)))
297 (not (and (looking-at parstart
)
298 (or (not use-hard-newlines
)
301 (1- start
) 'hard
)))))
302 (setq found-start nil
)
307 ;; Move forward over paragraph separators.
308 ;; We know this cannot reach the place we started
309 ;; because we know we moved back over a non-separator.
310 (while (and (not (eobp))
311 (progn (move-to-left-margin)
312 (looking-at parsep
)))
314 ;; If line before paragraph is just margin, back up to there.
316 (if (> (current-column) (current-left-margin))
318 (skip-chars-backward " \t")
321 ;; No starter or separator line => use buffer beg.
322 (goto-char (point-min))))))
324 (while (and (> arg
0) (not (eobp)))
325 ;; Move forward over separator lines...
326 (while (and (not (eobp))
327 (progn (move-to-left-margin) (not (eobp)))
330 (unless (eobp) (setq arg
(1- arg
)))
331 ;; ... and one more line.
333 (if fill-prefix-regexp
334 ;; There is a fill prefix; it overrides parstart.
335 (while (and (not (eobp))
336 (progn (move-to-left-margin) (not (eobp)))
337 (not (looking-at parsep
))
338 (looking-at fill-prefix-regexp
))
340 (while (and (re-search-forward sp-parstart nil
1)
341 (progn (setq start
(match-beginning 0))
344 (progn (move-to-left-margin)
345 (not (looking-at parsep
)))
346 (or (not (looking-at parstart
))
347 (and use-hard-newlines
348 (not (get-text-property (1- start
) 'hard
)))))
350 (if (< (point) (point-max))
352 (constrain-to-field nil opoint t
)
353 ;; Return the number of steps that could not be done.
356 (defun backward-paragraph (&optional arg
)
357 "Move backward to start of paragraph.
358 With argument ARG, do it ARG times;
359 a negative argument ARG = -N means move forward N paragraphs.
361 A paragraph start is the beginning of a line which is a
362 `paragraph-start' or which is ordinary text and follows a
363 `paragraph-separate'ing line; except: if the first real line of a
364 paragraph is preceded by a blank line, the paragraph starts at that
367 See `forward-paragraph' for more information."
369 (or arg
(setq arg
1))
370 (forward-paragraph (- arg
)))
372 (defun mark-paragraph (&optional arg allow-extend
)
373 "Put point at beginning of this paragraph, mark at end.
374 The paragraph marked is the one that contains point or follows point.
376 With argument ARG, puts mark at end of a following paragraph, so that
377 the number of paragraphs marked equals ARG.
379 If ARG is negative, point is put at end of this paragraph, mark is put
380 at beginning of this or a previous paragraph.
382 Interactively (or if ALLOW-EXTEND is non-nil), if this command is
383 repeated or (in Transient Mark mode) if the mark is active,
384 it marks the next ARG paragraphs after the ones already marked."
386 (unless arg
(setq arg
1))
388 (error "Cannot mark zero paragraphs"))
389 (cond ((and allow-extend
390 (or (and (eq last-command this-command
) (mark t
))
391 (and transient-mark-mode mark-active
)))
395 (forward-paragraph arg
)
398 (forward-paragraph arg
)
400 (backward-paragraph arg
))))
402 (defun kill-paragraph (arg)
403 "Kill forward to end of paragraph.
404 With arg N, kill forward to Nth end of paragraph;
405 negative arg -N means kill backward to Nth start of paragraph."
407 (kill-region (point) (progn (forward-paragraph arg
) (point))))
409 (defun backward-kill-paragraph (arg)
410 "Kill back to start of paragraph.
411 With arg N, kill back to Nth start of paragraph;
412 negative arg -N means kill forward to Nth end of paragraph."
414 (kill-region (point) (progn (backward-paragraph arg
) (point))))
416 (defun transpose-paragraphs (arg)
417 "Interchange the current paragraph with the next one.
418 With prefix argument ARG a non-zero integer, moves the current
419 paragraph past ARG paragraphs, leaving point after the current paragraph.
420 If ARG is positive, moves the current paragraph forwards, if
421 ARG is negative moves it backwards. If ARG is zero, exchanges
422 the current paragraph with the one containing the mark."
424 (transpose-subr 'forward-paragraph arg
))
426 (defun start-of-paragraph-text ()
427 (let ((opoint (point)) npoint
)
428 (forward-paragraph -
1)
429 (setq npoint
(point))
430 (skip-chars-forward " \t\n")
431 ;; If the range of blank lines found spans the original start point,
432 ;; try again from the beginning of it.
433 ;; Must be careful to avoid infinite loop
434 ;; when following a single return at start of buffer.
435 (if (and (>= (point) opoint
) (< npoint opoint
))
438 (if (> npoint
(point-min))
439 (start-of-paragraph-text))))))
441 (defun end-of-paragraph-text ()
442 (let ((opoint (point)))
443 (forward-paragraph 1)
444 (if (eq (preceding-char) ?
\n) (forward-char -
1))
445 (if (<= (point) opoint
)
448 (if (< (point) (point-max))
449 (end-of-paragraph-text))))))
451 (defun forward-sentence (&optional arg
)
452 "Move forward to next end of sentence. With argument, repeat.
453 With negative argument, move backward repeatedly to start of sentence.
455 The variable `sentence-end' is a regular expression that matches ends of
456 sentences. Also, every paragraph boundary terminates sentences as well."
458 (or arg
(setq arg
1))
459 (let ((opoint (point))
460 (sentence-end (sentence-end)))
463 par-beg par-text-beg
)
465 (start-of-paragraph-text)
466 ;; Start of real text in the paragraph.
467 ;; We move back to here if we don't see a sentence-end.
468 (setq par-text-beg
(point))
469 ;; Start of the first line of the paragraph.
470 ;; We use this as the search limit
471 ;; to allow sentence-end to match if it is anchored at
472 ;; BOL and the paragraph starts indented.
474 (setq par-beg
(point)))
475 (if (and (re-search-backward sentence-end par-beg t
)
476 (or (< (match-end 0) pos
)
477 (re-search-backward sentence-end par-beg t
)))
478 (goto-char (match-end 0))
479 (goto-char par-text-beg
)))
482 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
483 (if (re-search-forward sentence-end par-end t
)
484 (skip-chars-backward " \t\n")
485 (goto-char par-end
)))
487 (constrain-to-field nil opoint t
)))
489 (defun repunctuate-sentences ()
490 "Put two spaces at the end of sentences from point to the end of buffer.
491 It works using `query-replace-regexp'."
493 (query-replace-regexp "\\([]\"')]?\\)\\([.?!]\\)\\([]\"')]?\\) +"
497 (defun backward-sentence (&optional arg
)
498 "Move backward to start of sentence. With arg, do it arg times.
499 See `forward-sentence' for more information."
501 (or arg
(setq arg
1))
502 (forward-sentence (- arg
)))
504 (defun kill-sentence (&optional arg
)
505 "Kill from point to end of sentence.
506 With arg, repeat; negative arg -N means kill back to Nth start of sentence."
508 (kill-region (point) (progn (forward-sentence arg
) (point))))
510 (defun backward-kill-sentence (&optional arg
)
511 "Kill back from point to start of sentence.
512 With arg, repeat, or kill forward to Nth end of sentence if negative arg -N."
514 (kill-region (point) (progn (backward-sentence arg
) (point))))
516 (defun mark-end-of-sentence (arg)
517 "Put mark at end of sentence. Arg works as in `forward-sentence'.
518 If this command is repeated, it marks the next ARG sentences after the
519 ones already marked."
523 (if (and (eq last-command this-command
) (mark t
))
525 (forward-sentence arg
)
529 (defun transpose-sentences (arg)
530 "Interchange the current sentence with the next one.
531 With prefix argument ARG a non-zero integer, moves the current
532 sentence past ARG sentences, leaving point after the current sentence.
533 If ARG is positive, moves the current sentence forwards, if
534 ARG is negative moves it backwards. If ARG is zero, exchanges
535 the current sentence with the one containing the mark."
537 (transpose-subr 'forward-sentence arg
))
543 ;;; paragraphs.el ends here