1 /* Iterating through multibyte strings: macros for multi-byte encodings. 2 Copyright (C) 2001, 2005, 2007, 2009-2015 Free Software Foundation, Inc. 3 4 This program is free software: you can redistribute it and/or modify 5 it under the terms of the GNU General Public License as published by 6 the Free Software Foundation; either version 3 of the License, or 7 (at your option) any later version. 8 9 This program 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 General Public License for more details. 13 14 You should have received a copy of the GNU General Public License 15 along with this program. If not, see <http://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 98 /* Tru64 with Desktop Toolkit C has a bug: <stdio.h> must be included before 99 <wchar.h>. 100 BSD/OS 4.1 has a bug: <stdio.h> and <time.h> must be included before 101 <wchar.h>. */ 102 #include <stdio.h> 103 #include <time.h> 104 #include <wchar.h> 105 106 #include "mbchar.h" 107 #include "strnlen1.h" 108 109 #ifndef _GL_INLINE_HEADER_BEGIN 110 #error "Please include config.h first." 111 #endif 112 _GL_INLINE_HEADER_BEGIN 113 #ifndef MBUITER_INLINE 114 # define MBUITER_INLINE _GL_INLINE 115 #endif 116 117 struct mbuiter_multi 118 { 119 bool in_shift; /* true if next byte may not be interpreted as ASCII */ 120 mbstate_t state; /* if in_shift: current shift state */ 121 bool next_done; /* true if mbui_avail has already filled the following */ 122 struct mbchar cur; /* the current character: 123 const char *cur.ptr pointer to current character 124 The following are only valid after mbui_avail. 125 size_t cur.bytes number of bytes of current character 126 bool cur.wc_valid true if wc is a valid wide character 127 wchar_t cur.wc if wc_valid: the current character 128 */ 129 }; 130 131 MBUITER_INLINE void 132 mbuiter_multi_next (struct mbuiter_multi *iter) 133 { 134 if (iter->next_done) 135 return; 136 if (iter->in_shift) 137 goto with_shift; 138 /* Handle most ASCII characters quickly, without calling mbrtowc(). */ 139 if (is_basic (*iter->cur.ptr)) 140 { 141 /* These characters are part of the basic character set. ISO C 99 142 guarantees that their wide character code is identical to their 143 char code. */ 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 iter->in_shift = true; 152 with_shift: 153 iter->cur.bytes = mbrtowc (&iter->cur.wc, iter->cur.ptr, 154 strnlen1 (iter->cur.ptr, MB_CUR_MAX), 155 &iter->state); 156 if (iter->cur.bytes == (size_t) -1) 157 { 158 /* An invalid multibyte sequence was encountered. */ 159 iter->cur.bytes = 1; 160 iter->cur.wc_valid = false; 161 /* Whether to set iter->in_shift = false and reset iter->state 162 or not is not very important; the string is bogus anyway. */ 163 } 164 else if (iter->cur.bytes == (size_t) -2) 165 { 166 /* An incomplete multibyte character at the end. */ 167 iter->cur.bytes = strlen (iter->cur.ptr); 168 iter->cur.wc_valid = false; 169 /* Whether to set iter->in_shift = false and reset iter->state 170 or not is not important; the string end is reached anyway. */ 171 } 172 else 173 { 174 if (iter->cur.bytes == 0) 175 { 176 /* A null wide character was encountered. */ 177 iter->cur.bytes = 1; 178 assert (*iter->cur.ptr == '\0'); 179 assert (iter->cur.wc == 0); 180 } 181 iter->cur.wc_valid = true; 182 183 /* When in the initial state, we can go back treating ASCII 184 characters more quickly. */ 185 if (mbsinit (&iter->state)) 186 iter->in_shift = false; 187 } 188 } 189 iter->next_done = true; 190 } 191 192 MBUITER_INLINE void 193 mbuiter_multi_reloc (struct mbuiter_multi *iter, ptrdiff_t ptrdiff) 194 { 195 iter->cur.ptr += ptrdiff; 196 } 197 198 MBUITER_INLINE void 199 mbuiter_multi_copy (struct mbuiter_multi *new_iter, const struct mbuiter_multi *old_iter) 200 { 201 if ((new_iter->in_shift = old_iter->in_shift)) 202 memcpy (&new_iter->state, &old_iter->state, sizeof (mbstate_t)); 203 else 204 memset (&new_iter->state, 0, sizeof (mbstate_t)); 205 new_iter->next_done = old_iter->next_done; 206 mb_copy (&new_iter->cur, &old_iter->cur); 207 } 208 209 /* Iteration macros. */ 210 typedef struct mbuiter_multi mbui_iterator_t; 211 #define mbui_init(iter, startptr) \ 212 ((iter).cur.ptr = (startptr), \ 213 (iter).in_shift = false, memset (&(iter).state, '\0', sizeof (mbstate_t)), \ 214 (iter).next_done = false) 215 #define mbui_avail(iter) \ 216 (mbuiter_multi_next (&(iter)), !mb_isnul ((iter).cur)) 217 #define mbui_advance(iter) \ 218 ((iter).cur.ptr += (iter).cur.bytes, (iter).next_done = false) 219 220 /* Access to the current character. */ 221 #define mbui_cur(iter) (iter).cur 222 #define mbui_cur_ptr(iter) (iter).cur.ptr 223 224 /* Relocation. */ 225 #define mbui_reloc(iter, ptrdiff) mbuiter_multi_reloc (&iter, ptrdiff) 226 227 /* Copying an iterator. */ 228 #define mbui_copy mbuiter_multi_copy 229 230 _GL_INLINE_HEADER_END 231 232 #endif /* _MBUITER_H */ 233