wide.c revision 475b496bc008381e64c802250441cc256622ce91
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * This file and its contents are supplied under the terms of the
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Common Development and Distribution License ("CDDL"), version 1.0.
5aec55eb0591d2fcdd38d7dd5408a6ff3456e596Garrett D'Amore * You may only use this file in accordance with the terms of version
5aec55eb0591d2fcdd38d7dd5408a6ff3456e596Garrett D'Amore * 1.0 of the CDDL.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * A full copy of the text of the CDDL should have accompanied this
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * source. A copy of the CDDL is also available via the Internet at
475b496bc008381e64c802250441cc256622ce91Garrett D'Amore * Copyright 2011 Nexenta Systems, Inc. All rights reserved.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * The functions in this file convert from the standard multibyte forms
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * to the wide character forms used internally by libc. Unfortunately,
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * this approach means that we need a method for each and every encoding.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_none(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_utf8(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_big5(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_gbk(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_gb2312(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_gb18030(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_mskanji(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_euccn(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_eucjp(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_euckr(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int towide_euctw(wchar_t *, const char *, int);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int (*_towide)(wchar_t *, const char *, int) = towide_none;
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic int (*_tomb)(char *, wchar_t) = tomb_none;
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Table of supported encodings. We only bother to list the multibyte
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * encodings here, because single byte locales are handed by "NONE".
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amorestatic struct {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* the name that the underlying libc implemenation uses */
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore /* the maximum number of bits required for priorities */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore int (*towide)(wchar_t *, const char *, int);
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * UTF8 values max out at 0x1fffff (although in theory there could
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * be later extensions, but it won't happen.) This means we only need
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * 21 bits to be able to encode the entire range of priorities.
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "UTF-8", "UTF-8", 21, towide_utf8, tomb_utf8 },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "UTF8", "UTF-8", 21, towide_utf8, tomb_utf8 },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "utf8", "UTF-8", 21, towide_utf8, tomb_utf8 },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "utf-8", "UTF-8", 21, towide_utf8, tomb_utf8 },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "EUC-CN", "EUC-CN", 16, towide_euccn, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "eucCN", "EUC-CN", 16, towide_euccn, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * Becuase the 3-byte form of EUC-JP use the same leading byte,
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * only 17 bits required to provide unique priorities. (The low
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * bit of that first byte is set.) By setting this value low,
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * we can get by with only 3 bytes in the strxfrm expansion.
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "EUC-JP", "EUC-JP", 17, towide_eucjp, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "eucJP", "EUC-JP", 17, towide_eucjp, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "EUC-KR", "EUC-KR", 16, towide_euckr, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "eucKR", "EUC-KR", 16, towide_euckr, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * EUC-TW uses 2 bytes most of the time, but 4 bytes if the
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * high order byte is 0x8E. However, with 4 byte encodings,
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * the third byte will be A0-B0. So we only need to consider
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * the lower order 24 bits for collation.
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "EUC-TW", "EUC-TW", 24, towide_euctw, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "eucTW", "EUC-TW", 24, towide_euctw, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "MS_Kanji", "MSKanji", 16, towide_mskanji, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "MSKanji", "MSKanji", 16, towide_mskanji, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "PCK", "MSKanji", 16, towide_mskanji, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "SJIS", "MSKanji", 16, towide_mskanji, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "Shift_JIS", "MSKanji", 16, towide_mskanji, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "BIG5", "BIG5", 16, towide_big5, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "big5", "BIG5", 16, towide_big5, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "Big5", "BIG5", 16, towide_big5, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * GB18030 can get away with just 31 bits. This is because the
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * high order bit is always set for 4 byte values, and the
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * at least one of the other bits in that 4 byte value will
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * be non-zero.
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "GB18030", "GB18030", 31, towide_gb18030, tomb_mbs },
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * This should probably be an aliase for euc-cn, or vice versa.
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore { "GB2312", "GB2312", 16, towide_gb2312, tomb_mbs },
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* ASCII stuff we just print */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore while (*mb != 0) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore (void) snprintf(scr, sizeof (scr), "\\x%02x", *mb);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * This is used for 8-bit encodings.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_none(wchar_t *c, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("invalid or unsupported multibyte locale");
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("invalid or unsupported multibyte locale");
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * UTF-8 stores wide characters in UTF-32 form.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_utf8(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore if ((c & 0x80) == 0) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* 7-bit ASCII */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* u80-u7ff - two bytes encoded */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* u800-uffff - three bytes encoded */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* u1000-u1fffff - four bytes encoded */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* 5 and 6 byte encodings are not legal unicode */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("utf8 encoding too large (%s)", show_mb(mb));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("incomplete utf8 sequence (%s)", show_mb(mb));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore c |= (s[i] & 0x3f);
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("illegal redundant utf8 encoding (%s)", show_mb(mb));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Several encodings share a simplistic dual byte encoding. In these
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * forms, they all indicate that a two byte sequence is to be used if
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * the first byte has its high bit set. They all store this simple
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * encoding as a 16-bit value, although a great many of the possible
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * code points are not used in most character sets. This gives a possible
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * set of just over 32,000 valid code points.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * 0x00 - 0x7f - 1 byte encoding
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * 0x80 - 0x7fff - illegal
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * 0x8000 - 0xffff - 2 byte encoding
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_dbcs(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore if ((c & 0x80) == 0) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("incomplete character sequence (%s)", show_mb(mb));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* Store both bytes as a single 16-bit wide. */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Most multibyte locales just convert the wide character to the multibyte
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * form by stripping leading null bytes, and writing the 32-bit quantity
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * in big-endian order.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore int n = 0, c;
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* ensure null termination */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * big5 is a simple dual byte character set.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_big5(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * GBK encodes wides in the same way that big5 does, the high order
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * bit of the first byte indicates a double byte character.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_gbk(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * GB2312 is another DBCS. Its cleaner than others in that the second
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * byte does not encode ASCII, but it supports characters.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_gb2312(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * GB18030. This encodes as 8, 16, or 32-bits.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * 7-bit values are in 1 byte, 4 byte sequences are used when
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * the second byte encodes 0x30-39 and all other sequences are 2 bytes.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_gb18030(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore if ((c & 0x80) == 0) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("incomplete character sequence (%s)", show_mb(mb));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* pull in the second byte */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore if (((c & 0xff) >= 0x30) && ((c & 0xff) <= 0x39)) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("incomplete 4-byte character sequence (%s)",
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * MS-Kanji (aka SJIS) is almost a clean DBCS like the others, but it
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * also has a range of single byte characters above 0x80. (0xa1-0xdf).
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_mskanji(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore if ((c < 0x80) || ((c > 0xa0) && (c < 0xe0))) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("incomplete character sequence (%s)", show_mb(mb));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* Store both bytes as a single 16-bit wide. */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * EUC forms. EUC encodings are "variable". FreeBSD carries some additional
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * variable data to encode these, but we're going to treat each as independent
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * instead. Its the only way we can sensibly move forward.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Note that the way in which the different EUC forms vary is how wide
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * CS2 and CS3 are and what the first byte of them is.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_euc_impl(wchar_t *wc, const char *mb, int n,
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore uint8_t cs2, uint8_t cs2width, uint8_t cs3, uint8_t cs3width)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * All variations of EUC encode 7-bit ASCII as one byte, and use
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * additional bytes for more than that.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore if ((c & 0x80) == 0) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * All EUC variants reserve 0xa1-0xff to identify CS1, which
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * is always two bytes wide. Note that unused CS will be zero,
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * and that cannot be true because we know that the high order
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * bit must be set.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore if (c >= 0xa1) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore } else if (c == cs2) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore } else if (c == cs3) {
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore werr("incomplete character sequence (%s)", show_mb(mb));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* pull in the next byte */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * EUC-CN encodes as follows:
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 0 (ASCII): 0x21-0x7E
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 1 (CNS 11643-1992 Plane 1): 0xA1A1-0xFEFE
723fee089c0a7e1bd9527b9a4b0f0abf5970336cGarrett D'Amore * Code set 2: unused
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 3: unused
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_euccn(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore return (towide_euc_impl(wc, mb, n, 0x8e, 4, 0, 0));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * EUC-JP encodes as follows:
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 0 (ASCII or JIS X 0201-1976 Roman): 0x21-0x7E
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 1 (JIS X 0208): 0xA1A1-0xFEFE
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 2 (half-width katakana): 0x8EA1-0x8EDF
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 3 (JIS X 0212-1990): 0x8FA1A1-0x8FFEFE
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_eucjp(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore return (towide_euc_impl(wc, mb, n, 0x8e, 2, 0x8f, 3));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * EUC-KR encodes as follows:
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 0 (ASCII or KS C 5636-1993): 0x21-0x7E
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 1 (KS C 5601-1992): 0xA1A1-0xFEFE
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 2: unused
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 3: unused
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_euckr(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore return (towide_euc_impl(wc, mb, n, 0, 0, 0, 0));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * EUC-TW encodes as follows:
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 0 (ASCII): 0x21-0x7E
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 1 (CNS 11643-1992 Plane 1): 0xA1A1-0xFEFE
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 2 (CNS 11643-1992 Planes 1-16): 0x8EA1A1A1-0x8EB0FEFE
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Code set 3: unused
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amoretowide_euctw(wchar_t *wc, const char *mb, int n)
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore return (towide_euc_impl(wc, mb, n, 0x8e, 4, 0, 0));
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore * Public entry points.
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore /* this won't fail hard */
6b5e5868e7ebf1aff3a5abd7d0c4ef0e5fbf3648Garrett D'Amore mbs = malloc((wcslen(wcs) * mb_cur_max) + 1);