xref: /illumos-gate/usr/src/lib/libc/port/locale/unicode.h (revision eda3ef2de2d15b389090f6ef953edaea3daaace4)
1*eda3ef2dSRobert Mustacchi /*
2*eda3ef2dSRobert Mustacchi  * This file and its contents are supplied under the terms of the
3*eda3ef2dSRobert Mustacchi  * Common Development and Distribution License ("CDDL"), version 1.0.
4*eda3ef2dSRobert Mustacchi  * You may only use this file in accordance with the terms of version
5*eda3ef2dSRobert Mustacchi  * 1.0 of the CDDL.
6*eda3ef2dSRobert Mustacchi  *
7*eda3ef2dSRobert Mustacchi  * A full copy of the text of the CDDL should have accompanied this
8*eda3ef2dSRobert Mustacchi  * source.  A copy of the CDDL is also available via the Internet at
9*eda3ef2dSRobert Mustacchi  * http://www.illumos.org/license/CDDL.
10*eda3ef2dSRobert Mustacchi  */
11*eda3ef2dSRobert Mustacchi 
12*eda3ef2dSRobert Mustacchi /*
13*eda3ef2dSRobert Mustacchi  * Copyright 2020 Robert Mustacchi
14*eda3ef2dSRobert Mustacchi  */
15*eda3ef2dSRobert Mustacchi 
16*eda3ef2dSRobert Mustacchi #ifndef _UNICODE_H
17*eda3ef2dSRobert Mustacchi #define	_UNICODE_H
18*eda3ef2dSRobert Mustacchi 
19*eda3ef2dSRobert Mustacchi /*
20*eda3ef2dSRobert Mustacchi  * Common definitions for dealing with Unicode.
21*eda3ef2dSRobert Mustacchi  *
22*eda3ef2dSRobert Mustacchi  * UTF-16 encodes data as a series of two byte values. However, there are more
23*eda3ef2dSRobert Mustacchi  * than 16-bit of code points. Code points inside of the first 16-bits are
24*eda3ef2dSRobert Mustacchi  * referred to as existing in the 'basic multilingual plane' (BMP). Those
25*eda3ef2dSRobert Mustacchi  * outside of it are in the 'supplementary plane'. When such a code point is
26*eda3ef2dSRobert Mustacchi  * encountered, it is encoded as a series of two uint16_t values.
27*eda3ef2dSRobert Mustacchi  *
28*eda3ef2dSRobert Mustacchi  * A value which is up to 20 bits (the current limit of the unicode code point
29*eda3ef2dSRobert Mustacchi  * space) is encoded by splitting it into two 10-bit values. The upper 10 bits
30*eda3ef2dSRobert Mustacchi  * are ORed with 0xd800 and the lower 10 bits are ORed with 0xdc00.
31*eda3ef2dSRobert Mustacchi  */
32*eda3ef2dSRobert Mustacchi 
33*eda3ef2dSRobert Mustacchi #ifdef __cplusplus
34*eda3ef2dSRobert Mustacchi extern "C" {
35*eda3ef2dSRobert Mustacchi #endif
36*eda3ef2dSRobert Mustacchi 
37*eda3ef2dSRobert Mustacchi /*
38*eda3ef2dSRobert Mustacchi  * Range of Unicode code points reserved for surrogate characters.
39*eda3ef2dSRobert Mustacchi  */
40*eda3ef2dSRobert Mustacchi #define	UNICODE_SUR_MIN		0xd800
41*eda3ef2dSRobert Mustacchi #define	UNICODE_SUR_MAX		0xdfff
42*eda3ef2dSRobert Mustacchi 
43*eda3ef2dSRobert Mustacchi /*
44*eda3ef2dSRobert Mustacchi  * Range of Unicode code points in supplementary planes.
45*eda3ef2dSRobert Mustacchi  */
46*eda3ef2dSRobert Mustacchi #define	UNICODE_SUP_START	0x10000
47*eda3ef2dSRobert Mustacchi #define	UNICODE_SUP_MAX		0x10ffff
48*eda3ef2dSRobert Mustacchi 
49*eda3ef2dSRobert Mustacchi /*
50*eda3ef2dSRobert Mustacchi  * Starting constants for surrogate pairs.
51*eda3ef2dSRobert Mustacchi  */
52*eda3ef2dSRobert Mustacchi #define	UNICODE_SUR_UPPER	0xd800
53*eda3ef2dSRobert Mustacchi #define	UNICODE_SUR_LOWER	0xdc00
54*eda3ef2dSRobert Mustacchi 
55*eda3ef2dSRobert Mustacchi /*
56*eda3ef2dSRobert Mustacchi  * Macros to extract the value from a surrogate pair and to take a code point
57*eda3ef2dSRobert Mustacchi  * and transform it into the surrogate version.
58*eda3ef2dSRobert Mustacchi  */
59*eda3ef2dSRobert Mustacchi #define	UNICODE_SUR_UVALUE(x)	(((x) & 0x3ff) << 10)
60*eda3ef2dSRobert Mustacchi #define	UNICODE_SUR_LVALUE(x)	((x) & 0x3ff)
61*eda3ef2dSRobert Mustacchi #define	UNICODE_SUR_UMASK(x)	(((x) >> 10) & 0x3ff)
62*eda3ef2dSRobert Mustacchi #define	UNICODE_SUR_LMASK(x)	((x) & 0x3ff)
63*eda3ef2dSRobert Mustacchi 
64*eda3ef2dSRobert Mustacchi #ifdef __cplusplus
65*eda3ef2dSRobert Mustacchi }
66*eda3ef2dSRobert Mustacchi #endif
67*eda3ef2dSRobert Mustacchi 
68*eda3ef2dSRobert Mustacchi #endif /* _UNICODE_H */
69