xref: /illumos-gate/usr/src/lib/iconv_modules/ja/common/Unicode_TO_Shift_JIS-2004.c (revision 16d8656330ae5622ec32e5007f62145ebafdc50f)
1 /*
2  * CDDL HEADER START
3  *
4  * The contents of this file are subject to the terms of the
5  * Common Development and Distribution License (the "License").
6  * You may not use this file except in compliance with the License.
7  *
8  * You can obtain a copy of the license at src/OPENSOLARIS.LICENSE
9  * or http://www.opensolaris.org/os/licensing.
10  * See the License for the specific language governing permissions
11  * and limitations under the License.
12  *
13  * When distributing Covered Code, include this CDDL HEADER in each
14  * file and include the License file at src/OPENSOLARIS.LICENSE.
15  * If applicable, add the following below this CDDL HEADER, with the
16  * fields enclosed by brackets "[]" replaced with your own identifying
17  * information: Portions Copyright [yyyy] [name of copyright owner]
18  *
19  * CDDL HEADER END
20  */
21 /*
22  * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
23  * Use is subject to license terms.
24  */
25 
26 
27 #include <stdlib.h>
28 #include <errno.h>
29 #include <euc.h>
30 
31 #include "japanese.h"
32 #include "jfp_iconv_unicode.h"
33 
34 #define	JFP_U2E_ICONV_X0213
35 #include "jfp_ucs2_to_euc16.h"
36 
37 #define	DEF_SINGLE	'?'
38 
39 #define	INVL	0xff
40 
41 static const unsigned char rowtosj1_x0213_p1[] = {
42      /*          1     2     3     4     5     6     7 */
43 	INVL, 0x81, 0x81, 0x82, 0x82, 0x83, 0x83, 0x84,
44      /*    8     9    10    11    12    13    14    15 */
45 	0x84, 0x85, 0x85, 0x86, 0x86, 0x87, 0x87, 0x88,
46      /*   16    17    18    19    20    21    22    23 */
47 	0x88, 0x89, 0x89, 0x8a, 0x8a, 0x8b, 0x8b, 0x8c,
48      /*   24    25    26    27    28    29    30    31 */
49 	0x8c, 0x8d, 0x8d, 0x8e, 0x8e, 0x8f, 0x8f, 0x90,
50      /*   32    33    34    35    36    37    38    39 */
51 	0x90, 0x91, 0x91, 0x92, 0x92, 0x93, 0x93, 0x94,
52      /*   40    41    42    43    44    45    46    47 */
53 	0x94, 0x95, 0x95, 0x96, 0x96, 0x97, 0x97, 0x98,
54      /*   48    49    50    51    52    53    54    55 */
55 	0x98, 0x99, 0x99, 0x9a, 0x9a, 0x9b, 0x9b, 0x9c,
56      /*   56    57    58    59    60    61    62    63 */
57 	0x9c, 0x9d, 0x9d, 0x9e, 0x9e, 0x9f, 0x9f, 0xe0,
58      /*   64    65    66    67    68    69    70    71 */
59 	0xe0, 0xe1, 0xe1, 0xe2, 0xe2, 0xe3, 0xe3, 0xe4,
60      /*   72    73    74    75    76    77    78    79 */
61 	0xe4, 0xe5, 0xe5, 0xe6, 0xe6, 0xe7, 0xe7, 0xe8,
62      /*   80    81    82    83    84    85    86    87 */
63 	0xe8, 0xe9, 0xe9, 0xea, 0xea, 0xeb, 0xeb, 0xec,
64      /*   88    89    90    91    92    93    94       */
65 	0xec, 0xed, 0xed, 0xee, 0xee, 0xef, 0xef
66 };
67 
68 static const unsigned char rowtosj1_x0213_p2[] = {
69      /*          1     2     3     4     5     6     7 */
70 	INVL, 0xf0, INVL, 0xf1, 0xf1, 0xf2, INVL, INVL,
71      /*    8     9    10    11    12    13    14    15 */
72 	0xf0, INVL, INVL, INVL, 0xf2, 0xf3, 0xf3, 0xf4,
73      /*   16    17    18    19    20    21    22    23 */
74 	INVL, INVL, INVL, INVL, INVL, INVL, INVL, INVL,
75      /*   24    25    26    27    28    29    30    31 */
76 	INVL, INVL, INVL, INVL, INVL, INVL, INVL, INVL,
77      /*   32    33    34    35    36    37    38    39 */
78 	INVL, INVL, INVL, INVL, INVL, INVL, INVL, INVL,
79      /*   40    41    42    43    44    45    46    47 */
80 	INVL, INVL, INVL, INVL, INVL, INVL, INVL, INVL,
81      /*   48    49    50    51    52    53    54    55 */
82 	INVL, INVL, INVL, INVL, INVL, INVL, INVL, INVL,
83      /*   56    57    58    59    60    61    62    63 */
84 	INVL, INVL, INVL, INVL, INVL, INVL, INVL, INVL,
85      /*   64    65    66    67    68    69    70    71 */
86 	INVL, INVL, INVL, INVL, INVL, INVL, INVL, INVL,
87      /*   72    73    74    75    76    77    78    79 */
88 	INVL, INVL, INVL, INVL, INVL, INVL, 0xf4, 0xf5,
89      /*   80    81    82    83    84    85    86    87 */
90 	0xf5, 0xf6, 0xf6, 0xf7, 0xf7, 0xf8, 0xf8, 0xf9,
91      /*   88    89    90    91    92    93    94       */
92 	0xf9, 0xfa, 0xfa, 0xfb, 0xfb, 0xfc, 0xfc
93 };
94 
e16tosj_x0213(unsigned short e16,unsigned char * pc2)95 static unsigned char e16tosj_x0213(
96 	unsigned short	e16,
97 	unsigned char	*pc2)
98 {
99 	unsigned char	c1;
100 	unsigned char	r, c;
101 
102 	/* range check (if valid for plane 1 or 2) for e16
103 	   has been done by the caller side */
104 
105 	r = (e16 >> 8) - 0xa0;
106 
107 	if ((e16 & 0x8080) == 0x8080) { /* Plane 1 */
108 		c1 = rowtosj1_x0213_p1[r];
109 	} else { /* Plane 2 */
110 		c1 = rowtosj1_x0213_p2[r];
111 	}
112 
113 	c = (e16 & 0x7f) - 0x20;
114 
115 	if ((r % 2) == 1) { /* odd row */
116 		*pc2 = (c - 1) + 0x40;
117 		if (*pc2 >= 0x7f) {
118 			(*pc2)++;
119 		}
120 	} else { /* even row */
121 		*pc2 = (c - 1) + 0x9f;
122 	}
123 
124 	return (c1);
125 }
126 
127 void *
_icv_open(void)128 _icv_open(void)
129 {
130 	return (_icv_open_unicode((size_t)0));
131 }
132 
133 void
_icv_close(void * cd)134 _icv_close(void *cd)
135 {
136 	_icv_close_unicode(cd);
137 	return;
138 }
139 
140 size_t
_icv_iconv(void * cd,char ** inbuf,size_t * inbytesleft,char ** outbuf,size_t * outbytesleft)141 _icv_iconv(void *cd, char **inbuf, size_t *inbytesleft,
142 				char **outbuf, size_t *outbytesleft)
143 {
144 	unsigned int	u32;		/* UTF-32 */
145 	unsigned short	e16;		/* 16-bit EUC */
146 	unsigned char	ic1, ic2;
147 	size_t		rv = (size_t)0;
148 
149 	unsigned char	*ip;
150         size_t		ileft;
151 	char		*op;
152         size_t		oleft;
153 
154 	/*
155 	 * If inbuf and/or *inbuf are NULL, reset conversion descriptor
156 	 * and put escape sequence if needed.
157 	 */
158 	if ((inbuf == NULL) || (*inbuf == NULL)) {
159 		_icv_reset_unicode(cd);
160 		return ((size_t)0);
161 	}
162 
163 	ip = (unsigned char *)*inbuf;
164 	ileft = *inbytesleft;
165 	op = *outbuf;
166 	oleft = *outbytesleft;
167 
168 	while (ileft != 0) {
169 		GETU(&u32)
170 
171 		e16 = _jfp_u32_to_euc16(u32);
172 
173 		switch (e16 & 0x8080) {
174 		case 0x0000:	/* ASCII */
175 			ic1 = (unsigned char)e16;
176 			NPUT(ic1, "ASCII");
177 			break;
178 		case 0x8080:	/* PLANE1 */
179 			ic1 = e16tosj_x0213(e16, &ic2);
180 			NPUT(ic1, "PLANE1-1");
181 			NPUT(ic2, "PLANE1-2");
182 			break;
183 		case 0x0080:	/* KANA */
184 			ic1 = (unsigned char)e16;
185 			NPUT(ic1, "KANA");
186 			break;
187 		case 0x8000:	/* CS3 */
188 			ic1 = e16tosj_x0213(e16, &ic2);
189 			NPUT(ic1, "PLANE2-1");
190 			NPUT(ic2, "PLANE2-2");
191 			break;
192 		}
193 
194 next:
195 		/*
196 		 * One character successfully converted so update
197 		 * values outside of this function's stack.
198 		 */
199 		*inbuf = (char *)ip;
200 		*inbytesleft = ileft;
201 		*outbuf = op;
202 		*outbytesleft = oleft;
203 	}
204 
205 ret:
206 
207 	DEBUGPRINTERROR
208 
209 	/*
210 	 * Return value for successful return is not defined by XPG
211 	 * so return same as *inbytesleft as existing codes do.
212 	 */
213 	return ((rv == (size_t)-1) ? rv : *inbytesleft);
214 }
215