1 ;;; character-fold.el --- match unicode to similar ASCII -*- lexical-binding: t; -*-
3 ;; Copyright (C) 2015-2016 Free Software Foundation, Inc.
5 ;; Maintainer: emacs-devel@gnu.org
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/>.
25 (eval-and-compile (put 'character-fold-table
'char-table-extra-slots
1))
27 (defconst character-fold-table
29 (let ((equiv (make-char-table 'character-fold-table
))
30 (equiv-multi (make-char-table 'character-fold-table
))
31 (table (unicode-property-table-internal 'decomposition
)))
32 (set-char-table-extra-slot equiv
0 equiv-multi
)
34 ;; Ensure the table is populated.
35 (let ((func (char-table-extra-slot table
1)))
36 (map-char-table (lambda (char v
)
38 (funcall func
(car char
) v table
)))
41 ;; Compile a list of all complex characters that each simple
42 ;; character should match.
43 ;; In summary this loop does 3 things:
44 ;; - A complex character might be allowed to match its decomp.
45 ;; - The decomp is allowed to match the complex character.
46 ;; - A single char of the decomp might be allowed to match the
48 ;; Some examples in the comments below.
52 ;; Skip trivial cases like ?a decomposing to (?a).
53 (unless (and (not (cdr decomp
))
54 (eq char
(car decomp
)))
55 (if (symbolp (car decomp
))
56 ;; Discard a possible formatting tag.
57 (setq decomp
(cdr decomp
))
58 ;; If there's no formatting tag, ensure that char matches
59 ;; its decomp exactly. This is because we want 'ä' to
60 ;; match 'ä', but we don't want '¹' to match '1'.
62 (cons (apply #'string decomp
)
65 ;; Allow the entire decomp to match char. If decomp has
66 ;; multiple characters, this is done by adding an entry
67 ;; to the alist of the first character in decomp. This
68 ;; allows 'ff' to match 'ff', 'ä' to match 'ä', and '1' to
70 (let ((make-decomp-match-char
73 (aset equiv-multi
(car decomp
)
74 (cons (cons (apply #'string
(cdr decomp
))
75 (regexp-quote (string char
)))
76 (aref equiv-multi
(car decomp
))))
77 (aset equiv
(car decomp
)
78 (cons (char-to-string char
)
79 (aref equiv
(car decomp
))))))))
80 (funcall make-decomp-match-char decomp char
)
81 ;; Do it again, without the non-spacing characters.
82 ;; This allows 'a' to match 'ä'.
83 (let ((simpler-decomp nil
)
86 (if (> (get-char-code-property c
'canonical-combining-class
) 0)
88 (push c simpler-decomp
)))
89 (when (and simpler-decomp found-one
)
90 (funcall make-decomp-match-char simpler-decomp char
)
91 ;; Finally, if the decomp only had one spacing
92 ;; character, we allow this character to match the
93 ;; decomp. This is to let 'a' match 'ä'.
94 (unless (cdr simpler-decomp
)
95 (aset equiv
(car simpler-decomp
)
96 (cons (apply #'string decomp
)
97 (aref equiv
(car simpler-decomp
)))))))))))
100 ;; Add some manual entries.
101 (dolist (it '((?
\" """ "“" "”" "”" "„" "⹂" "〞" "‟" "‟" "❞" "❝" "❠" "“" "„" "〝" "〟" "🙷" "🙶" "🙸" "«" "»")
102 (?
' "❟" "❛" "❜" "‘" "’" "‚" "‛" "‚" "" "❮" "❯" "‹" "›")
103 (?
` "❛" "‘" "‛" "" "❮" "‹")))
106 (aset equiv idx
(append chars
(aref equiv idx
)))))
108 ;; Convert the lists of characters we compiled into regexps.
110 (lambda (char dec-list
)
111 (let ((re (regexp-opt (cons (char-to-string char
) dec-list
))))
113 (set-char-table-range equiv char re
)
114 (aset equiv char re
))))
117 "Used for folding characters of the same group during search.
118 This is a char-table with the `character-fold-table' subtype.
120 Let us refer to the character in question by char-x.
121 Each entry is either nil (meaning char-x only matches literally)
122 or a regexp. This regexp should match anything that char-x can
123 match by itself \(including char-x). For instance, the default
124 regexp for the ?+ character is \"[+⁺₊﬩﹢+]\".
126 This table also has one extra slot which is also a char-table.
127 Each entry in the extra slot is an alist used for multi-character
128 matching (which may be nil). The elements of the alist should
129 have the form (SUFFIX . OTHER-REGEXP). If the characters after
130 char-x are equal to SUFFIX, then this combination of char-x +
131 SUFFIX is allowed to match OTHER-REGEXP. This is in addition to
132 char-x being allowed to match REGEXP.
133 For instance, the default alist for ?f includes:
134 \((\"fl\" . \"ffl\") (\"fi\" . \"ffi\")
135 (\"i\" . \"fi\") (\"f\" . \"ff\"))
137 Exceptionally for the space character (32), ALIST is ignored.")
139 (defun character-fold--make-space-string (n)
140 "Return a string that matches N spaces."
141 (format "\\(?:%s\\|%s\\)"
144 (make-list n
(or (aref character-fold-table ?\s
) " ")))))
147 (defun character-fold-to-regexp (string &optional _lax from
)
148 "Return a regexp matching anything that character-folds into STRING.
149 Any character in STRING that has an entry in
150 `character-fold-table' is replaced with that entry (which is a
151 regexp) and other characters are `regexp-quote'd.
153 If the resulting regexp would be too long for Emacs to handle,
154 just return the result of calling `regexp-quote' on STRING.
156 FROM is for internal use. It specifies an index in the STRING
157 from which to start."
159 (multi-char-table (char-table-extra-slot character-fold-table
0))
161 (end (length string
))
163 ;; When the user types a space, we want to match the table entry
164 ;; for ?\s, which is generally a regexp like "[ ...]". However,
165 ;; the `search-spaces-regexp' variable doesn't "see" spaces inside
166 ;; these regexp constructs, so we need to use "\\( \\|[ ...]\\)"
167 ;; instead (to manually expose a space). Furthermore, the lax
168 ;; search engine acts on a bunch of spaces, not on individual
169 ;; spaces, so if the string contains sequential spaces like " ", we
170 ;; need to keep them grouped together like this: "\\( \\|[ ...][ ...]\\)".
172 (pcase (aref string i
)
173 (`?\s
(setq spaces
(1+ spaces
)))
174 (c (when (> spaces
0)
175 (push (character-fold--make-space-string spaces
) out
)
177 (let ((regexp (or (aref character-fold-table c
)
178 (regexp-quote (string c
))))
179 ;; Long string. The regexp would probably be too long.
180 (alist (unless (> end
50)
181 (aref multi-char-table c
))))
182 (push (let ((matched-entries nil
)
184 (dolist (entry alist
)
185 (let* ((suffix (car entry
))
186 (len-suf (length suffix
)))
187 (when (eq (compare-strings suffix
0 nil
188 string
(1+ i
) (+ i
1 len-suf
)
191 (push (cons len-suf
(cdr entry
)) matched-entries
)
192 (setq max-length
(max max-length len-suf
)))))
193 ;; If no suffixes matched, just go on.
194 (if (not matched-entries
)
196 ;;; If N suffixes match, we "branch" out into N+1 executions for the
197 ;;; length of the longest match. This means "fix" will match "fix" but
198 ;;; not "fⅸ", but it's necessary to keep the regexp size from scaling
199 ;;; exponentially. See https://lists.gnu.org/archive/html/emacs-devel/2015-11/msg02562.html
200 (let ((subs (substring string
(1+ i
) (+ i
1 max-length
))))
201 ;; `i' is still going to inc by 1 below.
202 (setq i
(+ i max-length
))
205 (mapconcat (lambda (entry)
206 (let ((length (car entry
))
207 (suffix-regexp (cdr entry
)))
208 (concat suffix-regexp
209 (character-fold-to-regexp subs nil length
))))
210 `((0 .
,regexp
) .
,matched-entries
) "\\|")
215 (push (character-fold--make-space-string spaces
) out
))
216 (let ((regexp (apply #'concat
(nreverse out
))))
217 ;; Limited by `MAX_BUF_SIZE' in `regex.c'.
218 (if (> (length regexp
) 5000)
219 (regexp-quote string
)
223 ;;; Commands provided for completeness.
224 (defun character-fold-search-forward (string &optional bound noerror count
)
225 "Search forward for a character-folded version of STRING.
226 STRING is converted to a regexp with `character-fold-to-regexp',
227 which is searched for with `re-search-forward'.
228 BOUND NOERROR COUNT are passed to `re-search-forward'."
229 (interactive "sSearch: ")
230 (re-search-forward (character-fold-to-regexp string
) bound noerror count
))
232 (defun character-fold-search-backward (string &optional bound noerror count
)
233 "Search backward for a character-folded version of STRING.
234 STRING is converted to a regexp with `character-fold-to-regexp',
235 which is searched for with `re-search-backward'.
236 BOUND NOERROR COUNT are passed to `re-search-backward'."
237 (interactive "sSearch: ")
238 (re-search-backward (character-fold-to-regexp string
) bound noerror count
))
240 (provide 'character-fold
)
242 ;;; character-fold.el ends here