1*16d86563SAlexander Pyhalov /*
2*16d86563SAlexander Pyhalov * CDDL HEADER START
3*16d86563SAlexander Pyhalov *
4*16d86563SAlexander Pyhalov * The contents of this file are subject to the terms of the
5*16d86563SAlexander Pyhalov * Common Development and Distribution License (the "License").
6*16d86563SAlexander Pyhalov * You may not use this file except in compliance with the License.
7*16d86563SAlexander Pyhalov *
8*16d86563SAlexander Pyhalov * You can obtain a copy of the license at src/OPENSOLARIS.LICENSE
9*16d86563SAlexander Pyhalov * or http://www.opensolaris.org/os/licensing.
10*16d86563SAlexander Pyhalov * See the License for the specific language governing permissions
11*16d86563SAlexander Pyhalov * and limitations under the License.
12*16d86563SAlexander Pyhalov *
13*16d86563SAlexander Pyhalov * When distributing Covered Code, include this CDDL HEADER in each
14*16d86563SAlexander Pyhalov * file and include the License file at src/OPENSOLARIS.LICENSE.
15*16d86563SAlexander Pyhalov * If applicable, add the following below this CDDL HEADER, with the
16*16d86563SAlexander Pyhalov * fields enclosed by brackets "[]" replaced with your own identifying
17*16d86563SAlexander Pyhalov * information: Portions Copyright [yyyy] [name of copyright owner]
18*16d86563SAlexander Pyhalov *
19*16d86563SAlexander Pyhalov * CDDL HEADER END
20*16d86563SAlexander Pyhalov */
21*16d86563SAlexander Pyhalov
22*16d86563SAlexander Pyhalov /*
23*16d86563SAlexander Pyhalov * Copyright 2002 Sun Microsystems, Inc. All rights reserved.
24*16d86563SAlexander Pyhalov * Use is subject to license terms.
25*16d86563SAlexander Pyhalov */
26*16d86563SAlexander Pyhalov
27*16d86563SAlexander Pyhalov #include <sys/types.h>
28*16d86563SAlexander Pyhalov #include "../inc/common_defs.h"
29*16d86563SAlexander Pyhalov
30*16d86563SAlexander Pyhalov /*
31*16d86563SAlexander Pyhalov * convert utf8 string to unicode
32*16d86563SAlexander Pyhalov * return value: 0 - fail
33*16d86563SAlexander Pyhalov * 1 - success
34*16d86563SAlexander Pyhalov */
35*16d86563SAlexander Pyhalov int
convert_utf8_to_ucs4(uchar_t * ib,int utf8_len,uint_t * unicode)36*16d86563SAlexander Pyhalov convert_utf8_to_ucs4(uchar_t *ib, int utf8_len, uint_t *unicode)
37*16d86563SAlexander Pyhalov {
38*16d86563SAlexander Pyhalov uchar_t first_byte = *ib;
39*16d86563SAlexander Pyhalov uint_t u4;
40*16d86563SAlexander Pyhalov
41*16d86563SAlexander Pyhalov if ( number_of_bytes_in_utf8_char[first_byte] != utf8_len) return 0;
42*16d86563SAlexander Pyhalov
43*16d86563SAlexander Pyhalov u4 = (uint_t)(*ib++ & masks_tbl[utf8_len]);
44*16d86563SAlexander Pyhalov for (; utf8_len > 1; utf8_len--)
45*16d86563SAlexander Pyhalov {
46*16d86563SAlexander Pyhalov u4 = (u4 << ICV_UTF8_BIT_SHIFT) | (((uint_t) *ib) & ICV_UTF8_BIT_MASK);
47*16d86563SAlexander Pyhalov ++ib;
48*16d86563SAlexander Pyhalov }
49*16d86563SAlexander Pyhalov
50*16d86563SAlexander Pyhalov *unicode = u4;
51*16d86563SAlexander Pyhalov
52*16d86563SAlexander Pyhalov return 1;
53*16d86563SAlexander Pyhalov }
54*16d86563SAlexander Pyhalov
55*16d86563SAlexander Pyhalov /*
56*16d86563SAlexander Pyhalov * check whether the input 'str' is valid UTF-8 byte sequence or not,
57*16d86563SAlexander Pyhalov * which lenght is specified by 'utf8_len'
58*16d86563SAlexander Pyhalov *
59*16d86563SAlexander Pyhalov * return: 0 - invalid byte sequence
60*16d86563SAlexander Pyhalov * 1 - valid byte sequence
61*16d86563SAlexander Pyhalov */
62*16d86563SAlexander Pyhalov int
is_valid_utf8_string(uchar_t * str,int utf8_len)63*16d86563SAlexander Pyhalov is_valid_utf8_string(uchar_t *str, int utf8_len)
64*16d86563SAlexander Pyhalov {
65*16d86563SAlexander Pyhalov uint_t unicode = 0;
66*16d86563SAlexander Pyhalov uchar_t *ib = str;
67*16d86563SAlexander Pyhalov uchar_t first_byte;
68*16d86563SAlexander Pyhalov int is_second_byte = 0, len=utf8_len;
69*16d86563SAlexander Pyhalov
70*16d86563SAlexander Pyhalov if (number_of_bytes_in_utf8_char[*ib] == ICV_TYPE_ILLEGAL_CHAR ||
71*16d86563SAlexander Pyhalov number_of_bytes_in_utf8_char[*ib] != utf8_len ) return 0;
72*16d86563SAlexander Pyhalov
73*16d86563SAlexander Pyhalov first_byte = *ib;
74*16d86563SAlexander Pyhalov --utf8_len;
75*16d86563SAlexander Pyhalov ++ib;
76*16d86563SAlexander Pyhalov is_second_byte = 1;
77*16d86563SAlexander Pyhalov
78*16d86563SAlexander Pyhalov while (utf8_len != 0)
79*16d86563SAlexander Pyhalov {
80*16d86563SAlexander Pyhalov if (is_second_byte)
81*16d86563SAlexander Pyhalov {
82*16d86563SAlexander Pyhalov if ( *ib < valid_min_2nd_byte[first_byte] || *ib > valid_max_2nd_byte[first_byte] )
83*16d86563SAlexander Pyhalov return 0;
84*16d86563SAlexander Pyhalov is_second_byte = 0;
85*16d86563SAlexander Pyhalov }
86*16d86563SAlexander Pyhalov else if ((*ib & 0xc0) != 0x80) /* 0x80 -- 0xbf */
87*16d86563SAlexander Pyhalov return 0;
88*16d86563SAlexander Pyhalov
89*16d86563SAlexander Pyhalov --utf8_len;
90*16d86563SAlexander Pyhalov ++ib;
91*16d86563SAlexander Pyhalov }
92*16d86563SAlexander Pyhalov
93*16d86563SAlexander Pyhalov convert_utf8_to_ucs4(str, len, &unicode);
94*16d86563SAlexander Pyhalov if (unicode == 0xFFFE || unicode == 0xFFFF) return 0;
95*16d86563SAlexander Pyhalov
96*16d86563SAlexander Pyhalov return 1;
97*16d86563SAlexander Pyhalov }
98