libgo/runtime/go-rune.c

   1 /* go-rune.c -- rune functions for Go.
   2
   3    Copyright 2009, 2010 The Go Authors. All rights reserved.
   4    Use of this source code is governed by a BSD-style
   5    license that can be found in the LICENSE file.  */
   6
   7 #include <stddef.h>
   8
   9 #include "go-string.h"
  10
  11 /* Get a character from the UTF-8 string STR, of length LEN.  Store
  12    the Unicode character, if any, in *RUNE.  Return the number of
  13    characters used from STR.  */
  14
  15 int
  16 __go_get_rune (const unsigned char *str, size_t len, int *rune)
  17 {
  18   int c, c1, c2, c3;
  19
  20   /* Default to the "replacement character".  */
  21   *rune = 0xfffd;
  22
  23   if (len <= 0)
  24     return 1;
  25
  26   c = *str;
  27   if (c <= 0x7f)
  28     {
  29       *rune = c;
  30       return 1;
  31     }
  32
  33   if (len <= 1)
  34     return 1;
  35
  36   c1 = str[1];
  37   if ((c & 0xe0) == 0xc0
  38       && (c1 & 0xc0) == 0x80)
  39     {
  40       *rune = (((c & 0x1f) << 6)
  41                + (c1 & 0x3f));
  42       return 2;
  43     }
  44
  45   if (len <= 2)
  46     return 1;
  47
  48   c2 = str[2];
  49   if ((c & 0xf0) == 0xe0
  50       && (c1 & 0xc0) == 0x80
  51       && (c2 & 0xc0) == 0x80)
  52     {
  53       *rune = (((c & 0xf) << 12)
  54                + ((c1 & 0x3f) << 6)
  55                + (c2 & 0x3f));
  56
  57       if (*rune >= 0xd800 && *rune < 0xe000)
  58         {
  59           /* Invalid surrogate half; return replace character.  */
  60           *rune = 0xfffd;
  61           return 1;
  62         }
  63
  64       return 3;
  65     }
  66
  67   if (len <= 3)
  68     return 1;
  69
  70   c3 = str[3];
  71   if ((c & 0xf8) == 0xf0
  72       && (c1 & 0xc0) == 0x80
  73       && (c2 & 0xc0) == 0x80
  74       && (c3 & 0xc0) == 0x80)
  75     {
  76       *rune = (((c & 0x7) << 18)
  77                + ((c1 & 0x3f) << 12)
  78                + ((c2 & 0x3f) << 6)
  79                + (c3 & 0x3f));
  80       return 4;
  81     }
  82
  83   /* Invalid encoding.  Return 1 so that we advance.  */
  84   return 1;
  85 }