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 3 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, without knowing its length a-priori.
21
22 With these macros, an iteration loop that looks like
23
24 char *iter;
25 for (iter = buf; *iter != '\0'; iter++)
26 {
27 do_something (*iter);
28 }
29
30 becomes
31
32 mbui_iterator_t iter;
33 for (mbui_init (iter, buf); mbui_avail (iter); mbui_advance (iter))
34 {
35 do_something (mbui_cur_ptr (iter), mb_len (mbui_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 Compared to mbiter.h, the macros here don't need to know the string's
44 length a-priori. The downside is that at each step, the look-ahead
45 that guards against overrunning the terminating '\0' is more expensive.
46 The mbui_* macros are therefore suitable when there is a high probability
47 that only the first few multibyte characters need to be inspected.
48 Whereas the mbi_* macros are better if usually the iteration runs
49 through the entire string.
50
51 mbui_iterator_t
52 is a type usable for variable declarations.
53
54 mbui_init (iter, startptr)
55 initializes the iterator, starting at startptr.
56
57 mbui_avail (iter)
58 returns true if there are more multibyte characters available before
59 the end of string is reached. In this case, mbui_cur (iter) is
60 initialized to the next multibyte character.
61
62 mbui_advance (iter)
63 advances the iterator by one multibyte character.
64
65 mbui_cur (iter)
66 returns the current multibyte character, of type mbchar_t. All the
67 macros defined in mbchar.h can be used on it.
68
69 mbui_cur_ptr (iter)
70 return a pointer to the beginning of the current multibyte character.
71
72 mbui_reloc (iter, ptrdiff)
73 relocates iterator when the string is moved by ptrdiff bytes.
74
75 mbui_copy (&destiter, &srciter)
76 copies srciter to destiter.
77
78 Here are the function prototypes of the macros.
79
80 extern void mbui_init (mbui_iterator_t iter, const char *startptr);
81 extern bool mbui_avail (mbui_iterator_t iter);
82 extern void mbui_advance (mbui_iterator_t iter);
83 extern mbchar_t mbui_cur (mbui_iterator_t iter);
84 extern const char * mbui_cur_ptr (mbui_iterator_t iter);
85 extern void mbui_reloc (mbui_iterator_t iter, ptrdiff_t ptrdiff);
86 extern void mbui_copy (mbui_iterator_t *new, const mbui_iterator_t *old);
87 */
88
89 #ifndef _MBUITER_H
90 #define _MBUITER_H 1
91
92 /* This file uses _GL_INLINE_HEADER_BEGIN, _GL_INLINE. */
93 #if !_GL_CONFIG_H_INCLUDED
94 #error "Please include config.h first."
95 #endif
96
97 #include <assert.h>
98 #include <stddef.h>
99 #include <stdlib.h>
100 #include <string.h>
101 #include <uchar.h>
102
103 #include "mbchar.h"
104 #include "strnlen1.h"
105
106 _GL_INLINE_HEADER_BEGIN
107 #ifndef MBUITER_INLINE
108 # define MBUITER_INLINE _GL_INLINE
109 #endif
110
111 struct mbuiter_multi
112 {
113 bool in_shift; /* true if next byte may not be interpreted as ASCII */
114 mbstate_t state; /* if in_shift: current shift state */
115 bool next_done; /* true if mbui_avail has already filled the following */
116 struct mbchar cur; /* the current character:
117 const char *cur.ptr pointer to current character
118 The following are only valid after mbui_avail.
119 size_t cur.bytes number of bytes of current character
120 bool cur.wc_valid true if wc is a valid 32-bit wide character
121 char32_t cur.wc if wc_valid: the current character
122 */
123 };
124
125 MBUITER_INLINE void
126 mbuiter_multi_next (struct mbuiter_multi *iter)
127 {
128 if (iter->next_done)
129 return;
130 if (iter->in_shift)
131 goto with_shift;
132 /* Handle most ASCII characters quickly, without calling mbrtowc(). */
133 if (is_basic (*iter->cur.ptr))
134 {
135 /* These characters are part of the basic character set. ISO C 99
136 guarantees that their wide character code is identical to their
137 char code. */
138 iter->cur.bytes = 1;
139 iter->cur.wc = *iter->cur.ptr;
140 iter->cur.wc_valid = true;
141 }
142 else
143 {
144 assert (mbsinit (&iter->state));
145 iter->in_shift = true;
146 with_shift:
147 iter->cur.bytes = mbrtoc32 (&iter->cur.wc, iter->cur.ptr,
148 strnlen1 (iter->cur.ptr, MB_CUR_MAX),
149 &iter->state);
150 if (iter->cur.bytes == (size_t) -1)
151 {
152 /* An invalid multibyte sequence was encountered. */
153 iter->cur.bytes = 1;
154 iter->cur.wc_valid = false;
155 /* Whether to set iter->in_shift = false and reset iter->state
156 or not is not very important; the string is bogus anyway. */
157 }
158 else if (iter->cur.bytes == (size_t) -2)
159 {
160 /* An incomplete multibyte character at the end. */
161 iter->cur.bytes = strlen (iter->cur.ptr);
162 iter->cur.wc_valid = false;
163 /* Whether to set iter->in_shift = false and reset iter->state
164 or not is not important; the string end is reached anyway. */
165 }
166 else
167 {
168 if (iter->cur.bytes == 0)
169 {
170 /* A null wide character was encountered. */
171 iter->cur.bytes = 1;
172 assert (*iter->cur.ptr == '\0');
173 assert (iter->cur.wc == 0);
174 }
175 iter->cur.wc_valid = true;
176
177 /* When in the initial state, we can go back treating ASCII
178 characters more quickly. */
179 if (mbsinit (&iter->state))
180 iter->in_shift = false;
181 }
182 }
183 iter->next_done = true;
184 }
185
186 MBUITER_INLINE void
187 mbuiter_multi_reloc (struct mbuiter_multi *iter, ptrdiff_t ptrdiff)
188 {
189 iter->cur.ptr += ptrdiff;
190 }
191
192 MBUITER_INLINE void
193 mbuiter_multi_copy (struct mbuiter_multi *new_iter, const struct mbuiter_multi *old_iter)
194 {
195 if ((new_iter->in_shift = old_iter->in_shift))
196 memcpy (&new_iter->state, &old_iter->state, sizeof (mbstate_t));
197 else
198 memset (&new_iter->state, 0, sizeof (mbstate_t));
199 new_iter->next_done = old_iter->next_done;
200 mb_copy (&new_iter->cur, &old_iter->cur);
201 }
202
203 /* Iteration macros. */
204 typedef struct mbuiter_multi mbui_iterator_t;
205 #define mbui_init(iter, startptr) \
206 ((iter).cur.ptr = (startptr), \
207 (iter).in_shift = false, memset (&(iter).state, '\0', sizeof (mbstate_t)), \
208 (iter).next_done = false)
209 #define mbui_avail(iter) \
210 (mbuiter_multi_next (&(iter)), !mb_isnul ((iter).cur))
211 #define mbui_advance(iter) \
212 ((iter).cur.ptr += (iter).cur.bytes, (iter).next_done = false)
213
214 /* Access to the current character. */
215 #define mbui_cur(iter) (iter).cur
216 #define mbui_cur_ptr(iter) (iter).cur.ptr
217
218 /* Relocation. */
219 #define mbui_reloc(iter, ptrdiff) mbuiter_multi_reloc (&iter, ptrdiff)
220
221 /* Copying an iterator. */
222 #define mbui_copy mbuiter_multi_copy
223
224 _GL_INLINE_HEADER_END
225
226 #endif /* _MBUITER_H */