lib/mbuiter.h

   1 /* Iterating through multibyte strings: macros for multi-byte encodings.
   2    Copyright (C) 2001, 2005 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 2, or (at your option)
   7    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, write to the Free Software Foundation,
  16    Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  17
  18 /* Written by Bruno Haible <bruno@clisp.org>.  */
  19
  20 /* The macros in this file implement forward iteration through a
  21    multi-byte string, without knowing its length a-priori.
  22
  23    With these macros, an iteration loop that looks like
  24
  25       char *iter;
  26       for (iter = buf; *iter != '\0'; iter++)
  27         {
  28           do_something (*iter);
  29         }
  30
  31    becomes
  32
  33       mbui_iterator_t iter;
  34       for (mbui_init (iter, buf); mbui_avail (iter); mbui_advance (iter))
  35         {
  36           do_something (mbui_cur_ptr (iter), mb_len (mbui_cur (iter)));
  37         }
  38
  39    The benefit of these macros over plain use of mbrtowc is:
  40    - Handling of invalid multibyte sequences is possible without
  41      making the code more complicated, while still preserving the
  42      invalid multibyte sequences.
  43
  44    Compared to mbiter.h, the macros here don't need to know the string's
  45    length a-priori.  The downside is that at each step, the look-ahead
  46    that guards against overrunning the terminating '\0' is more expensive.
  47    The mbui_* macros are therefore suitable when there is a high probability
  48    that only the first few multibyte characters need to be inspected.
  49    Whereas the mbi_* macros are better if usually the iteration runs
  50    through the entire string.
  51
  52    mbui_iterator_t
  53      is a type usable for variable declarations.
  54
  55    mbui_init (iter, startptr)
  56      initializes the iterator, starting at startptr.
  57
  58    mbui_avail (iter)
  59      returns true if there are more multibyte chracters available before
  60      the end of string is reached. In this case, mbui_cur (iter) is
  61      initialized to the next multibyte chracter.
  62
  63    mbui_advance (iter)
  64      advances the iterator by one multibyte character.
  65
  66    mbui_cur (iter)
  67      returns the current multibyte character, of type mbchar_t.  All the
  68      macros defined in mbchar.h can be used on it.
  69
  70    mbui_cur_ptr (iter)
  71      return a pointer to the beginning of the current multibyte character.
  72
  73    mbui_reloc (iter, ptrdiff)
  74      relocates iterator when the string is moved by ptrdiff bytes.
  75
  76    Here are the function prototypes of the macros.
  77
  78    extern void          mbui_init (mbui_iterator_t iter, const char *startptr);
  79    extern bool          mbui_avail (mbui_iterator_t iter);
  80    extern void          mbui_advance (mbui_iterator_t iter);
  81    extern mbchar_t      mbui_cur (mbui_iterator_t iter);
  82    extern const char *  mbui_cur_ptr (mbui_iterator_t iter);
  83    extern void          mbui_reloc (mbui_iterator_t iter, ptrdiff_t ptrdiff);
  84  */
  85
  86 #ifndef _MBUITER_H
  87 #define _MBUITER_H 1
  88
  89 #include <assert.h>
  90 #include <stdbool.h>
  91 #include <stdlib.h>
  92
  93 /* Tru64 with Desktop Toolkit C has a bug: <stdio.h> must be included before
  94    <wchar.h>.
  95    BSD/OS 4.1 has a bug: <stdio.h> and <time.h> must be included before
  96    <wchar.h>.  */
  97 #include <stdio.h>
  98 #include <time.h>
  99 #include <wchar.h>
 100
 101 #include "mbchar.h"
 102 #include "strnlen1.h"
 103
 104 struct mbuiter_multi
 105 {
 106   bool in_shift;        /* true if next byte may not be interpreted as ASCII */
 107   mbstate_t state;      /* if in_shift: current shift state */
 108   bool next_done;       /* true if mbui_avail has already filled the following */
 109   struct mbchar cur;    /* the current character:
 110         const char *cur.ptr             pointer to current character
 111         The following are only valid after mbui_avail.
 112         size_t cur.bytes                number of bytes of current character
 113         bool cur.wc_valid               true if wc is a valid wide character
 114         wchar_t cur.wc                  if wc_valid: the current character
 115         */
 116 };
 117
 118 static inline void
 119 mbuiter_multi_next (struct mbuiter_multi *iter)
 120 {
 121   if (iter->next_done)
 122     return;
 123   if (iter->in_shift)
 124     goto with_shift;
 125   /* Handle most ASCII characters quickly, without calling mbrtowc().  */
 126   if (is_basic (*iter->cur.ptr))
 127     {
 128       /* These characters are part of the basic character set.  ISO C 99
 129          guarantees that their wide character code is identical to their
 130          char code.  */
 131       iter->cur.bytes = 1;
 132       iter->cur.wc = *iter->cur.ptr;
 133       iter->cur.wc_valid = true;
 134     }
 135   else
 136     {
 137       assert (mbsinit (&iter->state));
 138       iter->in_shift = true;
 139     with_shift:
 140       iter->cur.bytes = mbrtowc (&iter->cur.wc, iter->cur.ptr,
 141                                  strnlen1 (iter->cur.ptr, MB_CUR_MAX),
 142                                  &iter->state);
 143       if (iter->cur.bytes == (size_t) -1)
 144         {
 145           /* An invalid multibyte sequence was encountered.  */
 146           iter->cur.bytes = 1;
 147           iter->cur.wc_valid = false;
 148           /* Whether to set iter->in_shift = false and reset iter->state
 149              or not is not very important; the string is bogus anyway.  */
 150         }
 151       else if (iter->cur.bytes == (size_t) -2)
 152         {
 153           /* An incomplete multibyte character at the end.  */
 154           iter->cur.bytes = strlen (iter->cur.ptr);
 155           iter->cur.wc_valid = false;
 156           /* Whether to set iter->in_shift = false and reset iter->state
 157              or not is not important; the string end is reached anyway.  */
 158         }
 159       else
 160         {
 161           if (iter->cur.bytes == 0)
 162             {
 163               /* A null wide character was encountered.  */
 164               iter->cur.bytes = 1;
 165               assert (*iter->cur.ptr == '\0');
 166               assert (iter->cur.wc == 0);
 167             }
 168           iter->cur.wc_valid = true;
 169
 170           /* When in the initial state, we can go back treating ASCII
 171              characters more quickly.  */
 172           if (mbsinit (&iter->state))
 173             iter->in_shift = false;
 174         }
 175     }
 176   iter->next_done = true;
 177 }
 178
 179 static inline void
 180 mbuiter_multi_reloc (struct mbuiter_multi *iter, ptrdiff_t ptrdiff)
 181 {
 182   iter->cur.ptr += ptrdiff;
 183 }
 184
 185 /* Iteration macros.  */
 186 typedef struct mbuiter_multi mbui_iterator_t;
 187 #define mbui_init(iter, startptr) \
 188   ((iter).cur.ptr = (startptr), \
 189    (iter).in_shift = false, memset (&(iter).state, '\0', sizeof (mbstate_t)), \
 190    (iter).next_done = false)
 191 #define mbui_avail(iter) \
 192   (mbuiter_multi_next (&(iter)), !mb_isnul ((iter).cur))
 193 #define mbui_advance(iter) \
 194   ((iter).cur.ptr += (iter).cur.bytes, (iter).next_done = false)
 195
 196 /* Access to the current character.  */
 197 #define mbui_cur(iter) (iter).cur
 198 #define mbui_cur_ptr(iter) (iter).cur.ptr
 199
 200 /* Relocation.  */
 201 #define mbui_reloc(iter, ptrdiff) mbuiter_multi_reloc (&iter, ptrdiff)
 202
 203 #endif /* _MBUITER_H */