1 /* Conversion between encodings.
2 Copyright (C) 2001,2002 The Genes Development Team
3 This file is part of the Gedcom parser library.
4 Contributed by Peter Verthez <Peter.Verthez@advalvas.be>, 2001.
6 The Gedcom parser library is free software; you can redistribute it
7 and/or modify it under the terms of the GNU Lesser General Public
8 License as published by the Free Software Foundation; either
9 version 2.1 of the License, or (at your option) any later version.
11 The Gedcom parser library is distributed in the hope that it will be
12 useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 Lesser General Public License for more details.
16 You should have received a copy of the GNU Lesser General Public
17 License along with the Gedcom parser library; if not, write to the
18 Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
28 #include "gedcom_internal.h"
31 #include "encoding_state.h"
33 #include "utf8tools.h"
35 #define ENCODING_CONF_FILE "gedcom.enc"
36 #define GCONV_SEARCH_PATH "GCONV_PATH"
39 static hash_t *encodings = NULL;
41 const char* charwidth_string[] = { "1", "2_HILO", "2_LOHI" };
43 hnode_t *node_alloc(void *c UNUSED)
45 return (hnode_t *)malloc(sizeof *node_alloc(NULL));
48 void node_free(hnode_t *n, void *c UNUSED)
50 free((void*)hnode_getkey(n));
55 void add_encoding(const char *gedcom_n, const char* charwidth,
60 key = (char *) malloc(strlen(gedcom_n) + strlen(charwidth) + 3);
61 val = (char *) malloc(strlen(iconv_n) + 1);
64 /* sprintf is safe here (malloc'ed before) */
65 sprintf(key, "%s(%s)", gedcom_n, charwidth);
68 if (hash_lookup(encodings, key)) {
69 gedcom_warning(_("Duplicate entry found for encoding '%s', ignoring"),
75 hash_alloc_insert(encodings, key, val);
82 char* get_encoding(const char* gedcom_n, Encoding enc)
87 if (encodings == NULL) return NULL;
89 key = (char*)malloc(strlen(gedcom_n) + strlen(charwidth_string[enc]) + 3);
93 while ((sp_pos = strchr(gedcom_n, ' ')) != NULL) {
96 /* sprintf is safe here (malloc'ed before) */
97 sprintf(key, "%s(%s)", gedcom_n, charwidth_string[enc]);
99 node = hash_lookup(encodings, key);
102 return hnode_get(node);
105 gedcom_error(_("No encoding defined for '%s'"), gedcom_n);
115 void cleanup_encodings()
117 hash_free(encodings);
120 #ifdef USE_GLIBC_ICONV
122 static char *new_gconv_path;
124 void cleanup_gconv_path()
126 /* Clean up environment */
127 putenv(GCONV_SEARCH_PATH);
129 free(new_gconv_path);
132 /* Let function be called before main() */
133 void update_gconv_search_path() __attribute__ ((constructor));
135 #endif /* USE_GLIBC_ICONV */
139 The environment variable GCONV_PATH has to be adjusted before the very
140 first call of iconv_open. For the most general case, it means that we
141 have to make our own constructor here (in case some of the other library
142 constructors would use iconv_open).
144 However, it looks like a change of an environment variable in a constructor
145 doesn't always survive until the main() function. This is the case if
146 the environment variable is a new one, for which there was no room yet
147 in the initial environment. The initial environment is located on the
148 stack, but when variables are added, it is moved to the heap (to be able
149 to grow). Now, the main function takes again the one from the stack, not
150 from the heap, so changes are lost.
152 For this, the function below will also be called in gedcom_init(), which
153 needs to be called as early as possible in the program.
156 void update_gconv_search_path()
158 #ifdef USE_GLIBC_ICONV
160 /* Add gedcom data directory to gconv search path */
161 gconv_path = getenv(GCONV_SEARCH_PATH);
162 if (gconv_path == NULL || strstr(gconv_path, PKGDATADIR) == NULL) {
163 if (gconv_path == NULL) {
164 new_gconv_path = (char *)malloc(strlen(GCONV_SEARCH_PATH)
168 sprintf(new_gconv_path, "%s=%s", GCONV_SEARCH_PATH, PKGDATADIR);
171 new_gconv_path = (char *)malloc(strlen(GCONV_SEARCH_PATH)
176 sprintf(new_gconv_path, "%s=%s:%s",
177 GCONV_SEARCH_PATH, gconv_path, PKGDATADIR);
180 /* Ignore failures of putenv (can't do anything about it anyway) */
181 putenv(new_gconv_path);
183 fprintf(stderr, "Could not allocate memory at %s, %d\n",
188 if (init_called && atexit(cleanup_gconv_path) != 0) {
189 gedcom_warning(_("Could not register path cleanup function"));
191 #endif /* USE_GLIBC_ICONV */
194 void init_encodings()
196 if (encodings == NULL) {
198 char buffer[MAXBUF + 1];
199 char gedcom_n[MAXBUF + 1];
200 char charwidth[MAXBUF + 1];
201 char iconv_n[MAXBUF + 1];
203 if (atexit(cleanup_encodings) != 0) {
204 gedcom_warning(_("Could not register encoding cleanup function"));
207 encodings = hash_create(HASHCOUNT_T_MAX, NULL, NULL);
208 hash_set_allocator(encodings, node_alloc, node_free, NULL);
210 /* Open gedcom configuration file and read */
211 in = fopen(ENCODING_CONF_FILE, "r");
214 sprintf(path, "%s/%s", PKGDATADIR, ENCODING_CONF_FILE);
215 in = fopen(path, "r");
218 gedcom_warning(_("Could not open encoding configuration file '%s': %s"),
219 ENCODING_CONF_FILE, strerror(errno));
223 while (fgets(buffer, sizeof(buffer), in) != NULL) {
224 if (buffer[strlen(buffer) - 1] != '\n') {
225 gedcom_error(_("Line too long in encoding configuration file '%s'"),
230 else if ((buffer[0] != '#') && (strcmp(buffer, "\n") != 0)) {
231 if (sscanf(buffer, "%s %s %s", gedcom_n, charwidth, iconv_n) == 3) {
232 add_encoding(gedcom_n, charwidth, iconv_n);
235 gedcom_error(_("Missing data in encoding configuration file '%s'"),
243 if (fclose(in) != 0) {
244 gedcom_warning(_("Error closing file '%s': %s"),
245 ENCODING_CONF_FILE, strerror(errno));
251 static convert_t to_int = NULL;
252 static char* error_value = "<error>";
254 int open_conv_to_internal(const char* fromcode)
256 convert_t new_to_int = NULL;
257 const char *encoding = get_encoding(fromcode, read_encoding.width);
259 if (encoding != NULL) {
260 new_to_int = initialize_utf8_conversion(encoding, 1);
261 if (new_to_int == NULL) {
262 gedcom_error(_("Error opening conversion context for encoding %s: %s"),
263 encoding, strerror(errno));
267 if (new_to_int != NULL) {
269 cleanup_utf8_conversion(to_int);
271 set_read_encoding(fromcode, encoding);
274 return (new_to_int != NULL);
277 void close_conv_to_internal()
279 if (to_int != NULL) {
280 cleanup_utf8_conversion(to_int);
286 char* to_internal(const char* str, size_t len, struct conv_buffer* output_buf)
288 if (conversion_set_output_buffer(to_int, output_buf))
289 return convert_to_utf8_incremental(to_int, str, len);