|
Packit |
aea12f |
/* Conversion UTF-8 to UCS-4.
|
|
Packit Service |
991b93 |
Copyright (C) 2001-2002, 2006-2007, 2009-2020 Free Software Foundation, Inc.
|
|
Packit |
aea12f |
Written by Bruno Haible <bruno@clisp.org>, 2001.
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
This program is free software: you can redistribute it and/or
|
|
Packit |
aea12f |
modify it under the terms of either:
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
* the GNU Lesser General Public License as published by the Free
|
|
Packit |
aea12f |
Software Foundation; either version 3 of the License, or (at your
|
|
Packit |
aea12f |
option) any later version.
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
or
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
* the GNU General Public License as published by the Free
|
|
Packit |
aea12f |
Software Foundation; either version 2 of the License, or (at your
|
|
Packit |
aea12f |
option) any later version.
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
or both in parallel, as here.
|
|
Packit |
aea12f |
This program is distributed in the hope that it will be useful,
|
|
Packit |
aea12f |
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
Packit |
aea12f |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
Packit |
aea12f |
Lesser General Public License for more details.
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
You should have received a copy of the GNU Lesser General Public License
|
|
Packit |
aea12f |
along with this program. If not, see <https://www.gnu.org/licenses/>. */
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
#include <config.h>
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
/* Specification. */
|
|
Packit |
aea12f |
#include "unistr.h"
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
#if defined IN_LIBUNISTRING || HAVE_INLINE
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
int
|
|
Packit |
aea12f |
u8_mbtouc_unsafe_aux (ucs4_t *puc, const uint8_t *s, size_t n)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
uint8_t c = *s;
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
if (c >= 0xc2)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if (c < 0xe0)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if (n >= 2)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if ((s[1] ^ 0x80) < 0x40)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
*puc = ((unsigned int) (c & 0x1f) << 6)
|
|
Packit |
aea12f |
| (unsigned int) (s[1] ^ 0x80);
|
|
Packit |
aea12f |
return 2;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
/* invalid multibyte character */
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
else
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
/* incomplete multibyte character */
|
|
Packit |
aea12f |
*puc = 0xfffd;
|
|
Packit |
aea12f |
return 1;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
else if (c < 0xf0)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if (n >= 3)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if ((s[1] ^ 0x80) < 0x40)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if ((s[2] ^ 0x80) < 0x40)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if ((c >= 0xe1 || s[1] >= 0xa0)
|
|
Packit |
aea12f |
&& (c != 0xed || s[1] < 0xa0))
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
*puc = ((unsigned int) (c & 0x0f) << 12)
|
|
Packit |
aea12f |
| ((unsigned int) (s[1] ^ 0x80) << 6)
|
|
Packit |
aea12f |
| (unsigned int) (s[2] ^ 0x80);
|
|
Packit |
aea12f |
return 3;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
/* invalid multibyte character */
|
|
Packit |
aea12f |
*puc = 0xfffd;
|
|
Packit |
aea12f |
return 3;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
/* invalid multibyte character */
|
|
Packit |
aea12f |
*puc = 0xfffd;
|
|
Packit |
aea12f |
return 2;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
/* invalid multibyte character */
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
else
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
/* incomplete multibyte character */
|
|
Packit |
aea12f |
*puc = 0xfffd;
|
|
Packit |
aea12f |
if (n == 1 || (s[1] ^ 0x80) >= 0x40)
|
|
Packit |
aea12f |
return 1;
|
|
Packit |
aea12f |
else
|
|
Packit |
aea12f |
return 2;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
else if (c < 0xf8)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if (n >= 4)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if ((s[1] ^ 0x80) < 0x40)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if ((s[2] ^ 0x80) < 0x40)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if ((s[3] ^ 0x80) < 0x40)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
if ((c >= 0xf1 || s[1] >= 0x90)
|
|
Packit |
aea12f |
&& (c < 0xf4 || (c == 0xf4 && s[1] < 0x90))
|
|
Packit |
aea12f |
)
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
*puc = ((unsigned int) (c & 0x07) << 18)
|
|
Packit |
aea12f |
| ((unsigned int) (s[1] ^ 0x80) << 12)
|
|
Packit |
aea12f |
| ((unsigned int) (s[2] ^ 0x80) << 6)
|
|
Packit |
aea12f |
| (unsigned int) (s[3] ^ 0x80);
|
|
Packit |
aea12f |
return 4;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
/* invalid multibyte character */
|
|
Packit |
aea12f |
*puc = 0xfffd;
|
|
Packit |
aea12f |
return 4;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
/* invalid multibyte character */
|
|
Packit |
aea12f |
*puc = 0xfffd;
|
|
Packit |
aea12f |
return 3;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
/* invalid multibyte character */
|
|
Packit |
aea12f |
*puc = 0xfffd;
|
|
Packit |
aea12f |
return 2;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
/* invalid multibyte character */
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
else
|
|
Packit |
aea12f |
{
|
|
Packit |
aea12f |
/* incomplete multibyte character */
|
|
Packit |
aea12f |
*puc = 0xfffd;
|
|
Packit |
aea12f |
if (n == 1 || (s[1] ^ 0x80) >= 0x40)
|
|
Packit |
aea12f |
return 1;
|
|
Packit |
aea12f |
else if (n == 2 || (s[2] ^ 0x80) >= 0x40)
|
|
Packit |
aea12f |
return 2;
|
|
Packit |
aea12f |
else
|
|
Packit |
aea12f |
return 3;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
/* invalid multibyte character */
|
|
Packit |
aea12f |
*puc = 0xfffd;
|
|
Packit |
aea12f |
return 1;
|
|
Packit |
aea12f |
}
|
|
Packit |
aea12f |
|
|
Packit |
aea12f |
#endif
|