1 /* Iterating through multibyte strings: macros for multi-byte encodings.
2 Copyright (C) 2001, 2005, 2007, 2009-2022 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 #include <assert.h>
93 #include <stdbool.h>
94 #include <stddef.h>
95 #include <stdlib.h>
96 #include <string.h>
97 #include <wchar.h>
98
99 #include "mbchar.h"
100 #include "strnlen1.h"
101
102 #ifndef _GL_INLINE_HEADER_BEGIN
103 #error "Please include config.h first."
104 #endif
105 _GL_INLINE_HEADER_BEGIN
106 #ifndef MBUITER_INLINE
107 # define MBUITER_INLINE _GL_INLINE
108 #endif
109
110 struct mbuiter_multi
111 {
112 bool in_shift; /* true if next byte may not be interpreted as ASCII */
113 mbstate_t state; /* if in_shift: current shift state */
114 bool next_done; /* true if mbui_avail has already filled the following */
115 struct mbchar cur; /* the current character:
116 const char *cur.ptr pointer to current character
117 The following are only valid after mbui_avail.
118 size_t cur.bytes number of bytes of current character
119 bool cur.wc_valid true if wc is a valid wide character
120 wchar_t cur.wc if wc_valid: the current character
121 */
122 };
123
124 MBUITER_INLINE void
125 mbuiter_multi_next (struct mbuiter_multi *iter)
126 {
127 if (iter->next_done)
128 return;
129 if (iter->in_shift)
130 goto with_shift;
131 /* Handle most ASCII characters quickly, without calling mbrtowc(). */
132 if (is_basic (*iter->cur.ptr))
133 {
134 /* These characters are part of the basic character set. ISO C 99
135 guarantees that their wide character code is identical to their
136 char code. */
137 iter->cur.bytes = 1;
138 iter->cur.wc = *iter->cur.ptr;
139 iter->cur.wc_valid = true;
140 }
141 else
142 {
143 assert (mbsinit (&iter->state));
144 iter->in_shift = true;
145 with_shift:
146 iter->cur.bytes = mbrtowc (&iter->cur.wc, iter->cur.ptr,
147 strnlen1 (iter->cur.ptr, MB_CUR_MAX),
148 &iter->state);
149 if (iter->cur.bytes == (size_t) -1)
150 {
151 /* An invalid multibyte sequence was encountered. */
152 iter->cur.bytes = 1;
153 iter->cur.wc_valid = false;
154 /* Whether to set iter->in_shift = false and reset iter->state
155 or not is not very important; the string is bogus anyway. */
156 }
157 else if (iter->cur.bytes == (size_t) -2)
158 {
159 /* An incomplete multibyte character at the end. */
160 iter->cur.bytes = strlen (iter->cur.ptr);
161 iter->cur.wc_valid = false;
162 /* Whether to set iter->in_shift = false and reset iter->state
163 or not is not important; the string end is reached anyway. */
164 }
165 else
166 {
167 if (iter->cur.bytes == 0)
168 {
169 /* A null wide character was encountered. */
170 iter->cur.bytes = 1;
171 assert (*iter->cur.ptr == '\0');
172 assert (iter->cur.wc == 0);
173 }
174 iter->cur.wc_valid = true;
175
176 /* When in the initial state, we can go back treating ASCII
177 characters more quickly. */
178 if (mbsinit (&iter->state))
179 iter->in_shift = false;
180 }
181 }
182 iter->next_done = true;
183 }
184
185 MBUITER_INLINE void
186 mbuiter_multi_reloc (struct mbuiter_multi *iter, ptrdiff_t ptrdiff)
187 {
188 iter->cur.ptr += ptrdiff;
189 }
190
191 MBUITER_INLINE void
192 mbuiter_multi_copy (struct mbuiter_multi *new_iter, const struct mbuiter_multi *old_iter)
193 {
194 if ((new_iter->in_shift = old_iter->in_shift))
195 memcpy (&new_iter->state, &old_iter->state, sizeof (mbstate_t));
196 else
197 memset (&new_iter->state, 0, sizeof (mbstate_t));
198 new_iter->next_done = old_iter->next_done;
199 mb_copy (&new_iter->cur, &old_iter->cur);
200 }
201
202 /* Iteration macros. */
203 typedef struct mbuiter_multi mbui_iterator_t;
204 #define mbui_init(iter, startptr) \
205 ((iter).cur.ptr = (startptr), \
206 (iter).in_shift = false, memset (&(iter).state, '\0', sizeof (mbstate_t)), \
207 (iter).next_done = false)
208 #define mbui_avail(iter) \
209 (mbuiter_multi_next (&(iter)), !mb_isnul ((iter).cur))
210 #define mbui_advance(iter) \
211 ((iter).cur.ptr += (iter).cur.bytes, (iter).next_done = false)
212
213 /* Access to the current character. */
214 #define mbui_cur(iter) (iter).cur
215 #define mbui_cur_ptr(iter) (iter).cur.ptr
216
217 /* Relocation. */
218 #define mbui_reloc(iter, ptrdiff) mbuiter_multi_reloc (&iter, ptrdiff)
219
220 /* Copying an iterator. */
221 #define mbui_copy mbuiter_multi_copy
222
223 _GL_INLINE_HEADER_END
224
225 #endif /* _MBUITER_H */