1 /* Iterating through multibyte strings: macros for multi-byte encodings.
2 Copyright (C) 2001, 2005, 2007, 2009-2023 Free Software Foundation, Inc.
3
4 This file is free software: you can redistribute it and/or modify
5 it under the terms of the GNU Lesser General Public License as
6 published by the Free Software Foundation; either version 2.1 of the
7 License, or (at your option) any later version.
8
9 This file is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU Lesser General Public License for more details.
13
14 You should have received a copy of the GNU Lesser General Public License
15 along with this program. If not, see <https://www.gnu.org/licenses/>. */
16
17 /* Written by Bruno Haible <bruno@clisp.org>. */
18
19 /* The macros in this file implement forward iteration through a
20 multi-byte string.
21
22 With these macros, an iteration loop that looks like
23
24 char *iter;
25 for (iter = buf; iter < buf + buflen; iter++)
26 {
27 do_something (*iter);
28 }
29
30 becomes
31
32 mbi_iterator_t iter;
33 for (mbi_init (iter, buf, buflen); mbi_avail (iter); mbi_advance (iter))
34 {
35 do_something (mbi_cur_ptr (iter), mb_len (mbi_cur (iter)));
36 }
37
38 The benefit of these macros over plain use of mbrtowc is:
39 - Handling of invalid multibyte sequences is possible without
40 making the code more complicated, while still preserving the
41 invalid multibyte sequences.
42
43 mbi_iterator_t
44 is a type usable for variable declarations.
45
46 mbi_init (iter, startptr, length)
47 initializes the iterator, starting at startptr and crossing length bytes.
48
49 mbi_avail (iter)
50 returns true if there are more multibyte characters available before
51 the end of string is reached. In this case, mbi_cur (iter) is
52 initialized to the next multibyte character.
53
54 mbi_advance (iter)
55 advances the iterator by one multibyte character.
56
57 mbi_cur (iter)
58 returns the current multibyte character, of type mbchar_t. All the
59 macros defined in mbchar.h can be used on it.
60
61 mbi_cur_ptr (iter)
62 return a pointer to the beginning of the current multibyte character.
63
64 mbi_reloc (iter, ptrdiff)
65 relocates iterator when the string is moved by ptrdiff bytes.
66
67 mbi_copy (&destiter, &srciter)
68 copies srciter to destiter.
69
70 Here are the function prototypes of the macros.
71
72 extern void mbi_init (mbi_iterator_t iter,
73 const char *startptr, size_t length);
74 extern bool mbi_avail (mbi_iterator_t iter);
75 extern void mbi_advance (mbi_iterator_t iter);
76 extern mbchar_t mbi_cur (mbi_iterator_t iter);
77 extern const char * mbi_cur_ptr (mbi_iterator_t iter);
78 extern void mbi_reloc (mbi_iterator_t iter, ptrdiff_t ptrdiff);
79 extern void mbi_copy (mbi_iterator_t *new, const mbi_iterator_t *old);
80 */
81
82 #ifndef _MBITER_H
83 #define _MBITER_H 1
84
85 /* This file uses _GL_INLINE_HEADER_BEGIN, _GL_INLINE,
86 _GL_ATTRIBUTE_ALWAYS_INLINE. */
87 #if !_GL_CONFIG_H_INCLUDED
88 #error "Please include config.h first."
89 #endif
90
91 #include <assert.h>
92 #include <stddef.h>
93 #include <string.h>
94 #include <uchar.h>
95 #include <wchar.h>
96
97 #include "mbchar.h"
98
99 _GL_INLINE_HEADER_BEGIN
100 #ifndef MBITER_INLINE
101 # define MBITER_INLINE _GL_INLINE _GL_ATTRIBUTE_ALWAYS_INLINE
102 #endif
103
104 struct mbiter_multi
105 {
106 const char *limit; /* pointer to end of string */
107 #if !GNULIB_MBRTOC32_REGULAR
108 bool in_shift; /* true if next byte may not be interpreted as ASCII */
109 /* If GNULIB_MBRTOC32_REGULAR, it is always false,
110 so optimize it away. */
111 #endif
112 mbstate_t state; /* if in_shift: current shift state */
113 /* If GNULIB_MBRTOC32_REGULAR, it is in an initial state
114 before and after every mbiter_multi_next invocation.
115 */
116 bool next_done; /* true if mbi_avail has already filled the following */
117 struct mbchar cur; /* the current character:
118 const char *cur.ptr pointer to current character
119 The following are only valid after mbi_avail.
120 size_t cur.bytes number of bytes of current character
121 bool cur.wc_valid true if wc is a valid 32-bit wide character
122 char32_t cur.wc if wc_valid: the current character
123 */
124 };
125
126 MBITER_INLINE void
127 mbiter_multi_next (struct mbiter_multi *iter)
128 {
129 if (iter->next_done)
130 return;
131 #if !GNULIB_MBRTOC32_REGULAR
132 if (iter->in_shift)
133 goto with_shift;
134 #endif
135 /* Handle most ASCII characters quickly, without calling mbrtowc(). */
136 if (is_basic (*iter->cur.ptr))
137 {
138 /* These characters are part of the POSIX portable character set.
139 For most of them, namely those in the ISO C basic character set,
140 ISO C 99 guarantees that their wide character code is identical to
141 their char code. For the few other ones, this is the case as well,
142 in all locale encodings that are in use. The 32-bit wide character
143 code is the same as well. */
144 iter->cur.bytes = 1;
145 iter->cur.wc = *iter->cur.ptr;
146 iter->cur.wc_valid = true;
147 }
148 else
149 {
150 assert (mbsinit (&iter->state));
151 #if !GNULIB_MBRTOC32_REGULAR
152 iter->in_shift = true;
153 with_shift:
154 #endif
155 iter->cur.bytes = mbrtoc32 (&iter->cur.wc, iter->cur.ptr,
156 iter->limit - iter->cur.ptr, &iter->state);
157 if (iter->cur.bytes == (size_t) -1)
158 {
159 /* An invalid multibyte sequence was encountered. */
160 iter->cur.bytes = 1;
161 iter->cur.wc_valid = false;
162 /* Allow the next invocation to continue from a sane state. */
163 #if !GNULIB_MBRTOC32_REGULAR
164 iter->in_shift = false;
165 #endif
166 mbszero (&iter->state);
167 }
168 else if (iter->cur.bytes == (size_t) -2)
169 {
170 /* An incomplete multibyte character at the end. */
171 iter->cur.bytes = iter->limit - iter->cur.ptr;
172 iter->cur.wc_valid = false;
173 #if !GNULIB_MBRTOC32_REGULAR
174 /* Cause the next mbi_avail invocation to return false. */
175 iter->in_shift = false;
176 #endif
177 /* Whether to reset iter->state or not is not important; the
178 string end is reached anyway. */
179 }
180 else
181 {
182 if (iter->cur.bytes == 0)
183 {
184 /* A null wide character was encountered. */
185 iter->cur.bytes = 1;
186 assert (*iter->cur.ptr == '\0');
187 assert (iter->cur.wc == 0);
188 }
189 #if !GNULIB_MBRTOC32_REGULAR
190 else if (iter->cur.bytes == (size_t) -3)
191 /* The previous multibyte sequence produced an additional 32-bit
192 wide character. */
193 iter->cur.bytes = 0;
194 #endif
195 iter->cur.wc_valid = true;
196
197 /* When in an initial state, we can go back treating ASCII
198 characters more quickly. */
199 #if !GNULIB_MBRTOC32_REGULAR
200 if (mbsinit (&iter->state))
201 iter->in_shift = false;
202 #endif
203 }
204 }
205 iter->next_done = true;
206 }
207
208 MBITER_INLINE void
209 mbiter_multi_reloc (struct mbiter_multi *iter, ptrdiff_t ptrdiff)
210 {
211 iter->cur.ptr += ptrdiff;
212 iter->limit += ptrdiff;
213 }
214
215 MBITER_INLINE void
216 mbiter_multi_copy (struct mbiter_multi *new_iter, const struct mbiter_multi *old_iter)
217 {
218 new_iter->limit = old_iter->limit;
219 #if !GNULIB_MBRTOC32_REGULAR
220 if ((new_iter->in_shift = old_iter->in_shift))
221 memcpy (&new_iter->state, &old_iter->state, sizeof (mbstate_t));
222 else
223 #endif
224 mbszero (&new_iter->state);
225 new_iter->next_done = old_iter->next_done;
226 mb_copy (&new_iter->cur, &old_iter->cur);
227 }
228
229 /* Iteration macros. */
230 typedef struct mbiter_multi mbi_iterator_t;
231 #if !GNULIB_MBRTOC32_REGULAR
232 #define mbi_init(iter, startptr, length) \
233 ((iter).cur.ptr = (startptr), (iter).limit = (iter).cur.ptr + (length), \
234 (iter).in_shift = false, mbszero (&(iter).state), \
235 (iter).next_done = false)
236 #else
237 /* Optimized: no in_shift. */
238 #define mbi_init(iter, startptr, length) \
239 ((iter).cur.ptr = (startptr), (iter).limit = (iter).cur.ptr + (length), \
240 mbszero (&(iter).state), \
241 (iter).next_done = false)
242 #endif
243 #if !GNULIB_MBRTOC32_REGULAR
244 #define mbi_avail(iter) \
245 (((iter).cur.ptr < (iter).limit || (iter).in_shift) \
246 && (mbiter_multi_next (&(iter)), true))
247 #else
248 /* Optimized: no in_shift. */
249 #define mbi_avail(iter) \
250 ((iter).cur.ptr < (iter).limit \
251 && (mbiter_multi_next (&(iter)), true))
252 #endif
253 #define mbi_advance(iter) \
254 ((iter).cur.ptr += (iter).cur.bytes, (iter).next_done = false)
255
256 /* Access to the current character. */
257 #define mbi_cur(iter) (iter).cur
258 #define mbi_cur_ptr(iter) (iter).cur.ptr
259
260 /* Relocation. */
261 #define mbi_reloc(iter, ptrdiff) mbiter_multi_reloc (&iter, ptrdiff)
262
263 /* Copying an iterator. */
264 #define mbi_copy mbiter_multi_copy
265
266 _GL_INLINE_HEADER_END
267
268 #endif /* _MBITER_H */