Blame lib/unistr/u8-mbtoucr.c

Packit Service fdd496
/* Look at first character in UTF-8 string, returning an error code.
Packit Service fdd496
   Copyright (C) 1999-2002, 2006-2007, 2009-2017 Free Software Foundation, Inc.
Packit Service fdd496
   Written by Bruno Haible <bruno@clisp.org>, 2001.
Packit Service fdd496
Packit Service fdd496
   This program is free software: you can redistribute it and/or modify it
Packit Service fdd496
   under the terms of the GNU General Public License as published
Packit Service fdd496
   by the Free Software Foundation; either version 3 of the License, or
Packit Service fdd496
   (at your option) any later version.
Packit Service fdd496
Packit Service fdd496
   This program is distributed in the hope that it will be useful,
Packit Service fdd496
   but WITHOUT ANY WARRANTY; without even the implied warranty of
Packit Service fdd496
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
Packit Service fdd496
   General Public License for more details.
Packit Service fdd496
Packit Service fdd496
   You should have received a copy of the GNU General Public License
Packit Service fdd496
   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
Packit Service fdd496
Packit Service fdd496
#include <config.h>
Packit Service fdd496
Packit Service fdd496
/* Specification.  */
Packit Service fdd496
#include "unistr.h"
Packit Service fdd496
Packit Service fdd496
int
Packit Service fdd496
u8_mbtoucr (ucs4_t *puc, const uint8_t *s, size_t n)
Packit Service fdd496
{
Packit Service fdd496
  uint8_t c = *s;
Packit Service fdd496
Packit Service fdd496
  if (c < 0x80)
Packit Service fdd496
    {
Packit Service fdd496
      *puc = c;
Packit Service fdd496
      return 1;
Packit Service fdd496
    }
Packit Service fdd496
  else if (c >= 0xc2)
Packit Service fdd496
    {
Packit Service fdd496
      if (c < 0xe0)
Packit Service fdd496
        {
Packit Service fdd496
          if (n >= 2)
Packit Service fdd496
            {
Packit Service fdd496
              if ((s[1] ^ 0x80) < 0x40)
Packit Service fdd496
                {
Packit Service fdd496
                  *puc = ((unsigned int) (c & 0x1f) << 6)
Packit Service fdd496
                         | (unsigned int) (s[1] ^ 0x80);
Packit Service fdd496
                  return 2;
Packit Service fdd496
                }
Packit Service fdd496
              /* invalid multibyte character */
Packit Service fdd496
            }
Packit Service fdd496
          else
Packit Service fdd496
            {
Packit Service fdd496
              /* incomplete multibyte character */
Packit Service fdd496
              *puc = 0xfffd;
Packit Service fdd496
              return -2;
Packit Service fdd496
            }
Packit Service fdd496
        }
Packit Service fdd496
      else if (c < 0xf0)
Packit Service fdd496
        {
Packit Service fdd496
          if (n >= 2)
Packit Service fdd496
            {
Packit Service fdd496
              if ((s[1] ^ 0x80) < 0x40
Packit Service fdd496
                  && (c >= 0xe1 || s[1] >= 0xa0)
Packit Service fdd496
                  && (c != 0xed || s[1] < 0xa0))
Packit Service fdd496
                {
Packit Service fdd496
                  if (n >= 3)
Packit Service fdd496
                    {
Packit Service fdd496
                      if ((s[2] ^ 0x80) < 0x40)
Packit Service fdd496
                        {
Packit Service fdd496
                          *puc = ((unsigned int) (c & 0x0f) << 12)
Packit Service fdd496
                                 | ((unsigned int) (s[1] ^ 0x80) << 6)
Packit Service fdd496
                                 | (unsigned int) (s[2] ^ 0x80);
Packit Service fdd496
                          return 3;
Packit Service fdd496
                        }
Packit Service fdd496
                      /* invalid multibyte character */
Packit Service fdd496
                    }
Packit Service fdd496
                  else
Packit Service fdd496
                    {
Packit Service fdd496
                      /* incomplete multibyte character */
Packit Service fdd496
                      *puc = 0xfffd;
Packit Service fdd496
                      return -2;
Packit Service fdd496
                    }
Packit Service fdd496
                }
Packit Service fdd496
              /* invalid multibyte character */
Packit Service fdd496
            }
Packit Service fdd496
          else
Packit Service fdd496
            {
Packit Service fdd496
              /* incomplete multibyte character */
Packit Service fdd496
              *puc = 0xfffd;
Packit Service fdd496
              return -2;
Packit Service fdd496
            }
Packit Service fdd496
        }
Packit Service fdd496
      else if (c < 0xf8)
Packit Service fdd496
        {
Packit Service fdd496
          if (n >= 2)
Packit Service fdd496
            {
Packit Service fdd496
              if ((s[1] ^ 0x80) < 0x40
Packit Service fdd496
                  && (c >= 0xf1 || s[1] >= 0x90)
Packit Service fdd496
                  && (c < 0xf4 || (c == 0xf4 && s[1] < 0x90)))
Packit Service fdd496
                {
Packit Service fdd496
                  if (n >= 3)
Packit Service fdd496
                    {
Packit Service fdd496
                      if ((s[2] ^ 0x80) < 0x40)
Packit Service fdd496
                        {
Packit Service fdd496
                          if (n >= 4)
Packit Service fdd496
                            {
Packit Service fdd496
                              if ((s[3] ^ 0x80) < 0x40)
Packit Service fdd496
                                {
Packit Service fdd496
                                  *puc = ((unsigned int) (c & 0x07) << 18)
Packit Service fdd496
                                         | ((unsigned int) (s[1] ^ 0x80) << 12)
Packit Service fdd496
                                         | ((unsigned int) (s[2] ^ 0x80) << 6)
Packit Service fdd496
                                         | (unsigned int) (s[3] ^ 0x80);
Packit Service fdd496
                                  return 4;
Packit Service fdd496
                                }
Packit Service fdd496
                              /* invalid multibyte character */
Packit Service fdd496
                            }
Packit Service fdd496
                          else
Packit Service fdd496
                            {
Packit Service fdd496
                              /* incomplete multibyte character */
Packit Service fdd496
                              *puc = 0xfffd;
Packit Service fdd496
                              return -2;
Packit Service fdd496
                            }
Packit Service fdd496
                        }
Packit Service fdd496
                      /* invalid multibyte character */
Packit Service fdd496
                    }
Packit Service fdd496
                  else
Packit Service fdd496
                    {
Packit Service fdd496
                      /* incomplete multibyte character */
Packit Service fdd496
                      *puc = 0xfffd;
Packit Service fdd496
                      return -2;
Packit Service fdd496
                    }
Packit Service fdd496
                }
Packit Service fdd496
              /* invalid multibyte character */
Packit Service fdd496
            }
Packit Service fdd496
          else
Packit Service fdd496
            {
Packit Service fdd496
              /* incomplete multibyte character */
Packit Service fdd496
              *puc = 0xfffd;
Packit Service fdd496
              return -2;
Packit Service fdd496
            }
Packit Service fdd496
        }
Packit Service fdd496
    }
Packit Service fdd496
  /* invalid multibyte character */
Packit Service fdd496
  *puc = 0xfffd;
Packit Service fdd496
  return -1;
Packit Service fdd496
}