1 /* This program is free software; you can redistribute it and/or modify *
2 * it under the terms of the GNU General Public License as published by *
3 * the Free Software Foundation; either version 2 of the License, or *
4 * (at your option) any later version. *
6 (C) 2001 by The Genes Development Team
7 Original author: Peter Verthez (Peter.Verthez@advalvas.be)
20 #define INTERNAL_ENCODING "UTF8"
21 #define ENCODING_CONF_FILE "gedcom.enc"
24 static iconv_t cd_to_internal = (iconv_t) -1;
25 static void *encoding_mapping = NULL;
26 static ENCODING the_enc = ONE_BYTE;
33 char* charwidth_string[] = { "1", "2_HILO", "2_LOHI" };
35 int node_compare(const void *node1, const void *node2)
37 return strcmp(((const struct node *) node1)->gedcom_name,
38 ((const struct node *) node2)->gedcom_name);
41 void add_encoding(char *gedcom_n, char* charwidth, char *iconv_n)
44 struct node *nodeptr = (struct node *) malloc(sizeof *nodeptr);
45 nodeptr->gedcom_name = (char *) malloc(strlen(gedcom_n)
46 + strlen(charwidth) + 3);
47 nodeptr->iconv_name = (char *) malloc(strlen(iconv_n) + 1);
48 sprintf(nodeptr->gedcom_name, "%s(%s)", gedcom_n, charwidth);
49 strcpy(nodeptr->iconv_name, iconv_n);
50 datum = tsearch(nodeptr, &encoding_mapping, node_compare);
51 if ((datum == NULL) || (*datum != nodeptr)) {
52 gedcom_warning("Duplicate entry found for encoding '%s', ignoring",
57 char* get_encoding(char* gedcom_n, ENCODING enc)
60 struct node search_node;
61 char buffer[MAXBUF + 1];
62 sprintf(buffer, "%s(%s)", gedcom_n, charwidth_string[enc]);
63 search_node.gedcom_name = buffer;
64 datum = tfind(&search_node, &encoding_mapping, node_compare);
66 gedcom_error("No encoding found for '%s'", gedcom_n);
70 return ((const struct node *) *datum)->iconv_name;
76 if (encoding_mapping == NULL) {
78 char buffer[MAXBUF + 1];
79 char gedcom_n[MAXBUF + 1];
80 char charwidth[MAXBUF + 1];
81 char iconv_n[MAXBUF + 1];
82 in = fopen(ENCODING_CONF_FILE, "r");
84 while (fgets(buffer, sizeof(buffer), in) != NULL) {
85 if (buffer[strlen(buffer) - 1] != '\n') {
86 gedcom_error("Line too long in encoding configuration file '%s'",
90 else if ((buffer[0] != '#') && (strcmp(buffer, "\n") != 0)) {
91 if (sscanf(buffer, "%s %s %s", gedcom_n, charwidth, iconv_n) == 3) {
92 add_encoding(gedcom_n, charwidth, iconv_n);
95 gedcom_error("Missing data in encoding configuration file '%s'",
104 gedcom_warning("Could not open encoding configuration file '%s'",
110 void set_encoding_width(ENCODING enc)
115 static char conv_buf[MAXGEDCLINELEN * 2];
116 static size_t conv_buf_size;
118 int open_conv_to_internal(char* fromcode)
120 char *encoding = get_encoding(fromcode, the_enc);
121 if (cd_to_internal != (iconv_t) -1)
122 iconv_close(cd_to_internal);
123 if (encoding == NULL) {
124 cd_to_internal = (iconv_t) -1;
127 memset(conv_buf, 0, sizeof(conv_buf));
129 cd_to_internal = iconv_open(INTERNAL_ENCODING, encoding);
130 if (cd_to_internal == (iconv_t) -1) {
131 gedcom_error("Error opening conversion context for encoding %s: %s",
132 encoding, strerror(errno));
135 return (cd_to_internal != (iconv_t) -1);
138 void close_conv_to_internal()
140 iconv_close(cd_to_internal);
141 cd_to_internal = (iconv_t) -1;
144 char* to_internal(char* str, size_t len,
145 char* output_buffer, size_t out_len)
147 size_t outsize = out_len;
148 char *wrptr = output_buffer;
149 char *rdptr = conv_buf;
150 /* set up input buffer (concatenate to what was left previous time) */
151 /* can't use strcpy, because possible null bytes from unicode */
152 memcpy(conv_buf + conv_buf_size, str, len);
153 conv_buf_size += len;
154 /* set up output buffer (empty it) */
155 memset(output_buffer, 0, out_len);
156 /* do the conversion */
157 iconv(cd_to_internal, &rdptr, &conv_buf_size, &wrptr, &outsize);
158 /* then shift what is left over to the head of the input buffer */
159 memmove(conv_buf, rdptr, conv_buf_size);
160 memset(conv_buf + conv_buf_size, 0, sizeof(conv_buf) - conv_buf_size);
161 return output_buffer;