1*3fc10f8cSRobert Mustacchi /*
2*3fc10f8cSRobert Mustacchi * This file and its contents are supplied under the terms of the
3*3fc10f8cSRobert Mustacchi * Common Development and Distribution License ("CDDL"), version 1.0.
4*3fc10f8cSRobert Mustacchi * You may only use this file in accordance with the terms of version
5*3fc10f8cSRobert Mustacchi * 1.0 of the CDDL.
6*3fc10f8cSRobert Mustacchi *
7*3fc10f8cSRobert Mustacchi * A full copy of the text of the CDDL should have accompanied this
8*3fc10f8cSRobert Mustacchi * source. A copy of the CDDL is also available via the Internet at
9*3fc10f8cSRobert Mustacchi * http://www.illumos.org/license/CDDL.
10*3fc10f8cSRobert Mustacchi */
11*3fc10f8cSRobert Mustacchi
12*3fc10f8cSRobert Mustacchi /*
13*3fc10f8cSRobert Mustacchi * Copyright 2020 Robert Mustacchi
14*3fc10f8cSRobert Mustacchi */
15*3fc10f8cSRobert Mustacchi
16*3fc10f8cSRobert Mustacchi /*
17*3fc10f8cSRobert Mustacchi * C11 mbrtoc16(3C) support.
18*3fc10f8cSRobert Mustacchi *
19*3fc10f8cSRobert Mustacchi * The char16_t represents a UTF-16 encoding. This means that we have to deal
20*3fc10f8cSRobert Mustacchi * with surrogate pairs.
21*3fc10f8cSRobert Mustacchi */
22*3fc10f8cSRobert Mustacchi
23*3fc10f8cSRobert Mustacchi #include <locale.h>
24*3fc10f8cSRobert Mustacchi #include <wchar.h>
25*3fc10f8cSRobert Mustacchi #include <xlocale.h>
26*3fc10f8cSRobert Mustacchi #include <uchar.h>
27*3fc10f8cSRobert Mustacchi #include "mblocal.h"
28*3fc10f8cSRobert Mustacchi #include "unicode.h"
29*3fc10f8cSRobert Mustacchi
30*3fc10f8cSRobert Mustacchi #include <sys/debug.h>
31*3fc10f8cSRobert Mustacchi
32*3fc10f8cSRobert Mustacchi /*
33*3fc10f8cSRobert Mustacchi * Ensure that we never cause our save state to ever exceed that of the
34*3fc10f8cSRobert Mustacchi * mbstate_t. See the block comment in mblocal.h.
35*3fc10f8cSRobert Mustacchi */
36*3fc10f8cSRobert Mustacchi CTASSERT(sizeof (_CHAR16State) <= sizeof (mbstate_t));
37*3fc10f8cSRobert Mustacchi
38*3fc10f8cSRobert Mustacchi static mbstate_t mbrtoc16_state;
39*3fc10f8cSRobert Mustacchi
40*3fc10f8cSRobert Mustacchi size_t
mbrtoc16(char16_t * restrict pc16,const char * restrict str,size_t len,mbstate_t * restrict ps)41*3fc10f8cSRobert Mustacchi mbrtoc16(char16_t *restrict pc16, const char *restrict str, size_t len,
42*3fc10f8cSRobert Mustacchi mbstate_t *restrict ps)
43*3fc10f8cSRobert Mustacchi {
44*3fc10f8cSRobert Mustacchi wchar_t wc;
45*3fc10f8cSRobert Mustacchi size_t ret;
46*3fc10f8cSRobert Mustacchi char16_t out;
47*3fc10f8cSRobert Mustacchi _CHAR16State *c16s;
48*3fc10f8cSRobert Mustacchi
49*3fc10f8cSRobert Mustacchi if (ps == NULL) {
50*3fc10f8cSRobert Mustacchi ps = &mbrtoc16_state;
51*3fc10f8cSRobert Mustacchi }
52*3fc10f8cSRobert Mustacchi
53*3fc10f8cSRobert Mustacchi if (str == NULL) {
54*3fc10f8cSRobert Mustacchi pc16 = NULL;
55*3fc10f8cSRobert Mustacchi str = "";
56*3fc10f8cSRobert Mustacchi len = 1;
57*3fc10f8cSRobert Mustacchi }
58*3fc10f8cSRobert Mustacchi
59*3fc10f8cSRobert Mustacchi c16s = (_CHAR16State *)ps;
60*3fc10f8cSRobert Mustacchi if (c16s->c16_surrogate != 0) {
61*3fc10f8cSRobert Mustacchi if (pc16 != NULL) {
62*3fc10f8cSRobert Mustacchi *pc16 = c16s->c16_surrogate;
63*3fc10f8cSRobert Mustacchi }
64*3fc10f8cSRobert Mustacchi c16s->c16_surrogate = 0;
65*3fc10f8cSRobert Mustacchi return ((size_t)-3);
66*3fc10f8cSRobert Mustacchi }
67*3fc10f8cSRobert Mustacchi
68*3fc10f8cSRobert Mustacchi ret = mbrtowc_l(&wc, str, len, ps, uselocale(NULL));
69*3fc10f8cSRobert Mustacchi if ((ssize_t)ret < 0) {
70*3fc10f8cSRobert Mustacchi return (ret);
71*3fc10f8cSRobert Mustacchi }
72*3fc10f8cSRobert Mustacchi
73*3fc10f8cSRobert Mustacchi /*
74*3fc10f8cSRobert Mustacchi * If this character is not in the basic multilingual plane then we need
75*3fc10f8cSRobert Mustacchi * a surrogate character to represent it in UTF-16 and we will need to
76*3fc10f8cSRobert Mustacchi * write that out on the next iteration.
77*3fc10f8cSRobert Mustacchi */
78*3fc10f8cSRobert Mustacchi if (wc >= UNICODE_SUP_START) {
79*3fc10f8cSRobert Mustacchi wc -= UNICODE_SUP_START;
80*3fc10f8cSRobert Mustacchi c16s->c16_surrogate = UNICODE_SUR_LOWER | UNICODE_SUR_LMASK(wc);
81*3fc10f8cSRobert Mustacchi out = UNICODE_SUR_UPPER | UNICODE_SUR_UMASK(wc);
82*3fc10f8cSRobert Mustacchi } else {
83*3fc10f8cSRobert Mustacchi out = (char16_t)wc;
84*3fc10f8cSRobert Mustacchi }
85*3fc10f8cSRobert Mustacchi
86*3fc10f8cSRobert Mustacchi if (pc16 != NULL) {
87*3fc10f8cSRobert Mustacchi *pc16 = out;
88*3fc10f8cSRobert Mustacchi }
89*3fc10f8cSRobert Mustacchi
90*3fc10f8cSRobert Mustacchi return (ret);
91*3fc10f8cSRobert Mustacchi }
92