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 /*
17*eda3ef2dSRobert Mustacchi * C11 mbrtoc16(3C) support.
18*eda3ef2dSRobert Mustacchi *
19*eda3ef2dSRobert Mustacchi * The char16_t represents a UTF-16 encoding. This means that we have to deal
20*eda3ef2dSRobert Mustacchi * with surrogate pairs.
21*eda3ef2dSRobert Mustacchi */
22*eda3ef2dSRobert Mustacchi
23*eda3ef2dSRobert Mustacchi #include <locale.h>
24*eda3ef2dSRobert Mustacchi #include <wchar.h>
25*eda3ef2dSRobert Mustacchi #include <xlocale.h>
26*eda3ef2dSRobert Mustacchi #include <uchar.h>
27*eda3ef2dSRobert Mustacchi #include "mblocal.h"
28*eda3ef2dSRobert Mustacchi #include "unicode.h"
29*eda3ef2dSRobert Mustacchi
30*eda3ef2dSRobert Mustacchi #include <sys/debug.h>
31*eda3ef2dSRobert Mustacchi
32*eda3ef2dSRobert Mustacchi /*
33*eda3ef2dSRobert Mustacchi * Ensure that we never cause our save state to ever exceed that of the
34*eda3ef2dSRobert Mustacchi * mbstate_t. See the block comment in mblocal.h.
35*eda3ef2dSRobert Mustacchi */
36*eda3ef2dSRobert Mustacchi CTASSERT(sizeof (_CHAR16State) <= sizeof (mbstate_t));
37*eda3ef2dSRobert Mustacchi
38*eda3ef2dSRobert Mustacchi static mbstate_t mbrtoc16_state;
39*eda3ef2dSRobert Mustacchi
40*eda3ef2dSRobert Mustacchi size_t
mbrtoc16(char16_t * restrict pc16,const char * restrict str,size_t len,mbstate_t * restrict ps)41*eda3ef2dSRobert Mustacchi mbrtoc16(char16_t *restrict pc16, const char *restrict str, size_t len,
42*eda3ef2dSRobert Mustacchi mbstate_t *restrict ps)
43*eda3ef2dSRobert Mustacchi {
44*eda3ef2dSRobert Mustacchi wchar_t wc;
45*eda3ef2dSRobert Mustacchi size_t ret;
46*eda3ef2dSRobert Mustacchi char16_t out;
47*eda3ef2dSRobert Mustacchi _CHAR16State *c16s;
48*eda3ef2dSRobert Mustacchi
49*eda3ef2dSRobert Mustacchi if (ps == NULL) {
50*eda3ef2dSRobert Mustacchi ps = &mbrtoc16_state;
51*eda3ef2dSRobert Mustacchi }
52*eda3ef2dSRobert Mustacchi
53*eda3ef2dSRobert Mustacchi if (str == NULL) {
54*eda3ef2dSRobert Mustacchi pc16 = NULL;
55*eda3ef2dSRobert Mustacchi str = "";
56*eda3ef2dSRobert Mustacchi len = 1;
57*eda3ef2dSRobert Mustacchi }
58*eda3ef2dSRobert Mustacchi
59*eda3ef2dSRobert Mustacchi c16s = (_CHAR16State *)ps;
60*eda3ef2dSRobert Mustacchi if (c16s->c16_surrogate != 0) {
61*eda3ef2dSRobert Mustacchi if (pc16 != NULL) {
62*eda3ef2dSRobert Mustacchi *pc16 = c16s->c16_surrogate;
63*eda3ef2dSRobert Mustacchi }
64*eda3ef2dSRobert Mustacchi c16s->c16_surrogate = 0;
65*eda3ef2dSRobert Mustacchi return ((size_t)-3);
66*eda3ef2dSRobert Mustacchi }
67*eda3ef2dSRobert Mustacchi
68*eda3ef2dSRobert Mustacchi ret = mbrtowc_l(&wc, str, len, ps, uselocale(NULL));
69*eda3ef2dSRobert Mustacchi if ((ssize_t)ret < 0) {
70*eda3ef2dSRobert Mustacchi return (ret);
71*eda3ef2dSRobert Mustacchi }
72*eda3ef2dSRobert Mustacchi
73*eda3ef2dSRobert Mustacchi /*
74*eda3ef2dSRobert Mustacchi * If this character is not in the basic multilingual plane then we need
75*eda3ef2dSRobert Mustacchi * a surrogate character to represent it in UTF-16 and we will need to
76*eda3ef2dSRobert Mustacchi * write that out on the next iteration.
77*eda3ef2dSRobert Mustacchi */
78*eda3ef2dSRobert Mustacchi if (wc >= UNICODE_SUP_START) {
79*eda3ef2dSRobert Mustacchi wc -= UNICODE_SUP_START;
80*eda3ef2dSRobert Mustacchi c16s->c16_surrogate = UNICODE_SUR_LOWER | UNICODE_SUR_LMASK(wc);
81*eda3ef2dSRobert Mustacchi out = UNICODE_SUR_UPPER | UNICODE_SUR_UMASK(wc);
82*eda3ef2dSRobert Mustacchi } else {
83*eda3ef2dSRobert Mustacchi out = (char16_t)wc;
84*eda3ef2dSRobert Mustacchi }
85*eda3ef2dSRobert Mustacchi
86*eda3ef2dSRobert Mustacchi if (pc16 != NULL) {
87*eda3ef2dSRobert Mustacchi *pc16 = out;
88*eda3ef2dSRobert Mustacchi }
89*eda3ef2dSRobert Mustacchi
90*eda3ef2dSRobert Mustacchi return (ret);
91*eda3ef2dSRobert Mustacchi }
92