xref: /titanic_51/usr/src/lib/iconv_modules/vi/common/tcvn%viscii.c (revision 91e1e26ac6a73ce959289cf7d3d96c4baedbe0b8)
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 (c) 2008, by Sun Microsystems, Inc.
23  * All rights reserved.
24  */
25 
26 #include <stdio.h>
27 #include <errno.h>
28 #include <stdlib.h>
29 #include <sys/types.h>
30 #define __NEED_TCVN_2_VISCII__
31 #include <viscii_tcvn.h>    /* VISCII <-> TCVN mapping table */
32 #include <vi_combine.h>
33 #include "common_defs.h"
34 
35 
36 #define NON_ID_CHAR '?'     /* non-identified character */
37 
38 typedef struct _icv_state {
39     int	_errno;		/* internal errno */
40     unsigned short last;
41 } _iconv_st;
42 
43 
44 static int binsearch(unsigned short x, Combine_map_tcvn v[], int n);
45 
46 /*
47  * Open; called from iconv_open()
48  */
49 void *
50 _icv_open()
51 {
52     _iconv_st *st;
53 
54     if ((st = (_iconv_st *)malloc(sizeof(_iconv_st))) == NULL) {
55         errno = ENOMEM;
56 	return ((void *) -1);
57     }
58 
59     st->_errno = 0;
60     st->last = 0;
61     return ((void *) st);
62 }
63 
64 
65 /*
66  * Close; called from iconv_close()
67  */
68 void
69 _icv_close(_iconv_st *st)
70 {
71     if (!st)
72         errno = EBADF;
73     else
74         free(st);
75 }
76 
77 
78 /*
79  * Actual conversion; called from iconv()
80  */
81 size_t
82 _icv_iconv(_iconv_st *st, char **inbuf, size_t *inbytesleft,
83 				char **outbuf, size_t *outbytesleft)
84 {
85     int     unconv = 0;
86     int             idx = -1;
87     unsigned char   chout = 0;
88 #ifdef DEBUG
89     fprintf(stderr, "==========     iconv(): TCVN5712 -->UCS-2   ==========\n");
90 #endif
91     if (st == NULL) {
92         errno = EBADF;
93         return ((size_t) -1);
94     }
95 
96     if (inbuf == NULL || *inbuf == NULL) { /* Reset request. */
97         st->_errno = 0;
98         return ((size_t) 0);
99     }
100 
101     st->_errno = 0;     /* Reset internal errno */
102     errno = 0;          /* Reset external errno */
103 
104     /* Convert tcvn encoding to UCS-2 */
105     while (*inbytesleft > 0 && *outbytesleft > 0) {
106         unsigned char ch = 0;
107 
108         if (st->last != 0) {
109             if (ISCOMB_TCVN((unsigned char)**inbuf)) {
110                 /*
111                  * Composed characters with combine character
112                  */
113                 idx =  binsearch(st->last, tcvn_comb_data, VOWEL_NUM);
114                 if (idx >= 0) {
115                     ch = tcvn_comb_data[idx].composed[(unsigned char)**inbuf - 0xb0];
116                 } else {
117                     errno = EBADF;
118                     return ((size_t)-1);
119                 }
120                 st->last = 0;
121             } else {
122                 tcvn_2_viscii(st->last, &chout);
123                 if (st->last != 0x0 && chout == 0x0) {
124                     unconv++;
125                     chout = NON_ID_CHAR;
126                 }
127 
128                 *(*outbuf)++ = chout;
129                 (*outbytesleft) -= 1;
130                 ch = (unsigned char)**inbuf;
131             }
132             st->last = 0;
133         } else {
134             ch = (unsigned char)**inbuf;
135             if (ch >= 0x41 && ch <= 0xad
136                 && ((tcvn_comp_bases_mask0[(ch-0x40) >> 5] >> (ch & 0x1f)) & 1)) {
137                 /*
138                  * uni is vowel, it's a possible match with combine character.
139                  * Buffer it.
140                  * */
141                 st->last = ch;
142                 (*inbuf)++;
143                 (*inbytesleft)--;
144                 continue;
145             }
146         }
147 
148 
149         tcvn_2_viscii(ch, &chout);
150         if (ch != 0x0 && chout == 0x0) {
151             unconv++;
152             chout = NON_ID_CHAR;
153         }
154 
155         *(*outbuf)++ = chout;
156         (*outbytesleft) -= 1;
157 	(*inbuf)++;
158         (*inbytesleft)--;
159 
160     }
161 
162     if ( *inbytesleft > 0 && *outbytesleft <= 0 ) {
163         errno = E2BIG;
164         st->last = 0;
165         return ((size_t)-1);
166     }
167 
168     if (st->last != 0) {
169         tcvn_2_viscii(st->last, &chout);
170         if (**inbuf != 0x0 && chout == 0x0) {
171             unconv++;
172             chout = NON_ID_CHAR;
173         }
174         st->last = 0;
175         *(*outbuf)++ = chout;
176         (*outbytesleft) -= 1;
177     }
178     return ((size_t)unconv);
179 
180 }
181 
182 /* binsearch: find x in v[0] <= v[1] <= ... <= v[n-1] */
183 static int binsearch(unsigned short x, Combine_map_tcvn v[], int n)
184 {
185     int low = 0;
186     int mid = 0;
187     int high = n - 1;
188 
189     low = 0;
190     while (low <= high) {
191         mid = (low + high) / 2;
192         if (x < (unsigned short)v[mid].base)
193             high = mid - 1;
194         else if (x > (unsigned short)v[mid].base)
195             low = mid + 1;
196         else
197             /* found match */
198             return mid;
199     }
200 
201     /* no match */
202     return (-1);
203 }
204