xref: /illumos-gate/usr/src/uts/common/sys/u8_textprep.h (revision 4eaa471005973e11a6110b69fe990530b3b95a38)
1 /*
2  * CDDL HEADER START
3  *
4  * The contents of this file are subject to the terms of the
5  * Common Development and Distribution License (the "License").
6  * You may not use this file except in compliance with the License.
7  *
8  * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9  * or http://www.opensolaris.org/os/licensing.
10  * See the License for the specific language governing permissions
11  * and limitations under the License.
12  *
13  * When distributing Covered Code, include this CDDL HEADER in each
14  * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15  * If applicable, add the following below this CDDL HEADER, with the
16  * fields enclosed by brackets "[]" replaced with your own identifying
17  * information: Portions Copyright [yyyy] [name of copyright owner]
18  *
19  * CDDL HEADER END
20  */
21 /*
22  * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
23  * Use is subject to license terms.
24  */
25 
26 #ifndef	_SYS_U8_TEXTPREP_H
27 #define	_SYS_U8_TEXTPREP_H
28 
29 #pragma ident	"%Z%%M%	%I%	%E% SMI"
30 
31 #include <sys/isa_defs.h>
32 #include <sys/types.h>
33 #include <sys/errno.h>
34 
35 #ifdef	__cplusplus
36 extern "C" {
37 #endif
38 
39 /*
40  * Unicode encoding conversion functions and their macros.
41  */
42 #define	UCONV_IN_BIG_ENDIAN		0x0001
43 #define	UCONV_OUT_BIG_ENDIAN		0x0002
44 #define	UCONV_IN_SYSTEM_ENDIAN		0x0004
45 #define	UCONV_OUT_SYSTEM_ENDIAN		0x0008
46 #define	UCONV_IN_LITTLE_ENDIAN		0x0010
47 #define	UCONV_OUT_LITTLE_ENDIAN		0x0020
48 #define	UCONV_IGNORE_NULL		0x0040
49 #define	UCONV_IN_ACCEPT_BOM		0x0080
50 #define	UCONV_OUT_EMIT_BOM		0x0100
51 
52 extern int uconv_u16tou32(const uint16_t *, size_t *, uint32_t *, size_t *,
53 	int);
54 extern int uconv_u16tou8(const uint16_t *, size_t *, uchar_t *, size_t *, int);
55 extern int uconv_u32tou16(const uint32_t *, size_t *, uint16_t *, size_t *,
56 	int);
57 extern int uconv_u32tou8(const uint32_t *, size_t *, uchar_t *, size_t *, int);
58 extern int uconv_u8tou16(const uchar_t *, size_t *, uint16_t *, size_t *, int);
59 extern int uconv_u8tou32(const uchar_t *, size_t *, uint32_t *, size_t *, int);
60 
61 /*
62  * UTF-8 text preparation functions and their macros.
63  *
64  * Among the macros defined, U8_CANON_DECOMP, U8_COMPAT_DECOMP, and
65  * U8_CANON_COMP are not public interfaces and must not be used directly
66  * at the flag input argument.
67  */
68 #define	U8_STRCMP_CS			(0x00000001)
69 #define	U8_STRCMP_CI_UPPER		(0x00000002)
70 #define	U8_STRCMP_CI_LOWER		(0x00000004)
71 
72 #define	U8_CANON_DECOMP			(0x00000010)
73 #define	U8_COMPAT_DECOMP		(0x00000020)
74 #define	U8_CANON_COMP			(0x00000040)
75 
76 #define	U8_STRCMP_NFD			(U8_CANON_DECOMP)
77 #define	U8_STRCMP_NFC			(U8_CANON_DECOMP | U8_CANON_COMP)
78 #define	U8_STRCMP_NFKD			(U8_COMPAT_DECOMP)
79 #define	U8_STRCMP_NFKC			(U8_COMPAT_DECOMP | U8_CANON_COMP)
80 
81 #define	U8_TEXTPREP_TOUPPER		(U8_STRCMP_CI_UPPER)
82 #define	U8_TEXTPREP_TOLOWER		(U8_STRCMP_CI_LOWER)
83 
84 #define	U8_TEXTPREP_NFD			(U8_STRCMP_NFD)
85 #define	U8_TEXTPREP_NFC			(U8_STRCMP_NFC)
86 #define	U8_TEXTPREP_NFKD		(U8_STRCMP_NFKD)
87 #define	U8_TEXTPREP_NFKC		(U8_STRCMP_NFKC)
88 
89 #define	U8_TEXTPREP_IGNORE_NULL		(0x00010000)
90 #define	U8_TEXTPREP_IGNORE_INVALID	(0x00020000)
91 #define	U8_TEXTPREP_NOWAIT		(0x00040000)
92 
93 #define	U8_UNICODE_320			(0)
94 #define	U8_UNICODE_500			(1)
95 #define	U8_UNICODE_LATEST		(U8_UNICODE_500)
96 
97 #define	U8_VALIDATE_ENTIRE		(0x00100000)
98 #define	U8_VALIDATE_CHECK_ADDITIONAL	(0x00200000)
99 #define	U8_VALIDATE_UCS2_RANGE		(0x00400000)
100 
101 #define	U8_ILLEGAL_CHAR			(-1)
102 #define	U8_OUT_OF_RANGE_CHAR		(-2)
103 
104 extern int u8_validate(char *, size_t, char **, int, int *);
105 extern int u8_strcmp(const char *, const char *, size_t, int, size_t, int *);
106 extern size_t u8_textprep_str(char *, size_t *, char *, size_t *, int, size_t,
107 	int *);
108 
109 #ifdef	__cplusplus
110 }
111 #endif
112 
113 #endif	/* _SYS_U8_TEXTPREP_H */
114