150c77c6eSEd Schouten /*-
2*4d846d26SWarner Losh * SPDX-License-Identifier: BSD-2-Clause
3d915a14eSPedro F. Giffuni *
450c77c6eSEd Schouten * Copyright (c) 2013 Ed Schouten <ed@FreeBSD.org>
550c77c6eSEd Schouten * All rights reserved.
650c77c6eSEd Schouten *
750c77c6eSEd Schouten * Redistribution and use in source and binary forms, with or without
850c77c6eSEd Schouten * modification, are permitted provided that the following conditions
950c77c6eSEd Schouten * are met:
1050c77c6eSEd Schouten * 1. Redistributions of source code must retain the above copyright
1150c77c6eSEd Schouten * notice, this list of conditions and the following disclaimer.
1250c77c6eSEd Schouten * 2. Redistributions in binary form must reproduce the above copyright
1350c77c6eSEd Schouten * notice, this list of conditions and the following disclaimer in the
1450c77c6eSEd Schouten * documentation and/or other materials provided with the distribution.
1550c77c6eSEd Schouten *
1650c77c6eSEd Schouten * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
1750c77c6eSEd Schouten * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
1850c77c6eSEd Schouten * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
1950c77c6eSEd Schouten * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
2050c77c6eSEd Schouten * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
2150c77c6eSEd Schouten * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
2250c77c6eSEd Schouten * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
2350c77c6eSEd Schouten * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
2450c77c6eSEd Schouten * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
2550c77c6eSEd Schouten * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
2650c77c6eSEd Schouten * SUCH DAMAGE.
2750c77c6eSEd Schouten */
2850c77c6eSEd Schouten
2950c77c6eSEd Schouten #include <uchar.h>
3098f3acfaSYuri Pankov #include "mblocal.h"
3150c77c6eSEd Schouten
3250c77c6eSEd Schouten typedef struct {
3350c77c6eSEd Schouten char16_t trail_surrogate;
3450c77c6eSEd Schouten mbstate_t c32_mbstate;
3550c77c6eSEd Schouten } _Char16State;
3650c77c6eSEd Schouten
3750c77c6eSEd Schouten size_t
mbrtoc16_l(char16_t * __restrict pc16,const char * __restrict s,size_t n,mbstate_t * __restrict ps,locale_t locale)3850c77c6eSEd Schouten mbrtoc16_l(char16_t * __restrict pc16, const char * __restrict s, size_t n,
3950c77c6eSEd Schouten mbstate_t * __restrict ps, locale_t locale)
4050c77c6eSEd Schouten {
4150c77c6eSEd Schouten _Char16State *cs;
4250c77c6eSEd Schouten char32_t c32;
4350c77c6eSEd Schouten ssize_t len;
4450c77c6eSEd Schouten
4550c77c6eSEd Schouten FIX_LOCALE(locale);
4650c77c6eSEd Schouten if (ps == NULL)
4798f3acfaSYuri Pankov ps = &(XLOCALE_CTYPE(locale)->mbrtoc16);
4850c77c6eSEd Schouten cs = (_Char16State *)ps;
4950c77c6eSEd Schouten
5050c77c6eSEd Schouten /*
5150c77c6eSEd Schouten * Call straight into mbrtoc32_l() if we don't need to return a
5250c77c6eSEd Schouten * character value. According to the spec, if s is a null
5350c77c6eSEd Schouten * pointer, the value of parameter pc16 is also ignored.
5450c77c6eSEd Schouten */
5550c77c6eSEd Schouten if (pc16 == NULL || s == NULL) {
5650c77c6eSEd Schouten cs->trail_surrogate = 0;
5750c77c6eSEd Schouten return (mbrtoc32_l(NULL, s, n, &cs->c32_mbstate, locale));
5850c77c6eSEd Schouten }
5950c77c6eSEd Schouten
6050c77c6eSEd Schouten /* Return the trail surrogate from the previous invocation. */
6150c77c6eSEd Schouten if (cs->trail_surrogate >= 0xdc00 && cs->trail_surrogate <= 0xdfff) {
6250c77c6eSEd Schouten *pc16 = cs->trail_surrogate;
6350c77c6eSEd Schouten cs->trail_surrogate = 0;
6450c77c6eSEd Schouten return ((size_t)-3);
6550c77c6eSEd Schouten }
6650c77c6eSEd Schouten
6750c77c6eSEd Schouten len = mbrtoc32_l(&c32, s, n, &cs->c32_mbstate, locale);
6850c77c6eSEd Schouten if (len >= 0) {
6950c77c6eSEd Schouten if (c32 < 0x10000) {
7050c77c6eSEd Schouten /* Fits in one UTF-16 character. */
7150c77c6eSEd Schouten *pc16 = c32;
7250c77c6eSEd Schouten } else {
7350c77c6eSEd Schouten /* Split up in a surrogate pair. */
7450c77c6eSEd Schouten c32 -= 0x10000;
7550c77c6eSEd Schouten *pc16 = 0xd800 | (c32 >> 10);
7650c77c6eSEd Schouten cs->trail_surrogate = 0xdc00 | (c32 & 0x3ff);
7750c77c6eSEd Schouten }
7850c77c6eSEd Schouten }
7950c77c6eSEd Schouten return (len);
8050c77c6eSEd Schouten }
8150c77c6eSEd Schouten
8250c77c6eSEd Schouten size_t
mbrtoc16(char16_t * __restrict pc16,const char * __restrict s,size_t n,mbstate_t * __restrict ps)8350c77c6eSEd Schouten mbrtoc16(char16_t * __restrict pc16, const char * __restrict s, size_t n,
8450c77c6eSEd Schouten mbstate_t * __restrict ps)
8550c77c6eSEd Schouten {
8650c77c6eSEd Schouten
8750c77c6eSEd Schouten return (mbrtoc16_l(pc16, s, n, ps, __get_locale()));
8850c77c6eSEd Schouten }
89