-/*
- * Copyright (C) 1996-1999, Index Data
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: charmap.c,v $
- * Revision 1.18 1999-11-30 13:48:04 adam
- * Improved installation. Updated for inclusion of YAZ header files.
- *
- * Revision 1.17 1999/09/08 12:13:21 adam
- * Fixed minor bug "replace"-mappings. Removed some logging messages.
- *
- * Revision 1.16 1999/09/07 07:19:21 adam
- * Work on character mapping. Implemented replace rules.
- *
- * Revision 1.15 1999/05/26 07:49:14 adam
- * C++ compilation.
- *
- * Revision 1.14 1998/10/13 20:09:18 adam
- * Changed call to readconf_line.
- *
- * Revision 1.13 1997/10/27 14:33:06 adam
- * Moved towards generic character mapping depending on "structure"
- * field in abstract syntax file. Fixed a few memory leaks. Fixed
- * bug with negative integers when doing searches with relational
- * operators.
- *
- * Revision 1.12 1997/09/05 15:30:11 adam
- * Changed prototype for chr_map_input - added const.
- * Added support for C++, headers uses extern "C" for public definitions.
- *
- * Revision 1.11 1997/09/05 09:52:32 adam
- * Extra argument added to function chr_read_maptab (tab path).
- *
- * Revision 1.10 1997/07/01 13:01:08 adam
- * Bug fix in routine find_entry: didn't take into account the len arg.
- *
- * Revision 1.9 1996/10/29 13:48:14 adam
- * Updated to use zebrautl.h instead of alexutil.h.
- *
- * Revision 1.8 1996/10/18 12:39:23 adam
- * Uses LOG_DEBUG instead of LOG_WARN for "Character map overlap".
- *
- * Revision 1.7 1996/06/06 12:08:56 quinn
- * Fixed bug.
- *
- * Revision 1.6 1996/06/04 13:28:00 quinn
- * More work on charmapping
- *
- * Revision 1.5 1996/06/04 08:32:15 quinn
- * Moved default keymap to keychars.c
- *
- * Revision 1.4 1996/06/03 16:32:13 quinn
- * Temporary bug-fix
- *
- * Revision 1.3 1996/06/03 15:17:46 quinn
- * Fixed bug.
- *
- * Revision 1.2 1996/06/03 10:15:09 quinn
- * Fixed bug in mapping function.
- *
- * Revision 1.1 1996/05/31 09:07:18 quinn
- * Work on character-set handling
- *
- *
- */
+/* $Id: charmap.c,v 1.35 2005-01-16 23:14:58 adam Exp $
+ Copyright (C) 1995-2005
+ Index Data ApS
+
+This file is part of the Zebra server.
+
+Zebra is free software; you can redistribute it and/or modify it under
+the terms of the GNU General Public License as published by the Free
+Software Foundation; either version 2, or (at your option) any later
+version.
+
+Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
+WARRANTY; without even the implied warranty of MERCHANTABILITY or
+FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+for more details.
+
+You should have received a copy of the GNU General Public License
+along with Zebra; see the file LICENSE.zebra. If not, write to the
+Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
+02111-1307, USA.
+*/
+
+
/*
* Support module to handle character-conversions into and out of the
#include <ctype.h>
#include <string.h>
+#include <stdlib.h>
#include <assert.h>
-#include <yaz/yaz-util.h>
+typedef unsigned ucs4_t;
+
#include <charmap.h>
+#include <yaz/yaz-util.h>
+
#define CHR_MAXSTR 1024
#define CHR_MAXEQUIV 32
+const unsigned char CHR_FIELD_BEGIN = '^';
+
const char *CHR_UNKNOWN = "\001";
const char *CHR_SPACE = "\002";
-const char *CHR_BASE = "\003";
+const char *CHR_CUT = "\003";
+const char *CHR_BASE = "\005";
struct chrmaptab_info
{
* Add an entry to the character map.
*/
static chr_t_entry *set_map_string(chr_t_entry *root, NMEM nmem,
- const char *from, int len, char *to)
+ const char *from, int len, char *to,
+ const char *from_0)
{
+ if (!from_0)
+ from_0 = from;
if (!root)
{
root = (chr_t_entry *) nmem_malloc(nmem, sizeof(*root));
{
if (!root->target || !root->target[0] || strcmp(root->target[0], to))
{
+ if (from_0 &&
+ root->target && root->target[0] && root->target[0][0] &&
+ strcmp (root->target[0], CHR_UNKNOWN))
+ {
+ yaz_log (YLOG_WARN, "duplicate entry for charmap from '%s'",
+ from_0);
+ }
root->target = (unsigned char **)
nmem_malloc(nmem, sizeof(*root->target)*2);
root->target[0] = (unsigned char *) nmem_strdup(nmem, to);
}
if (!(root->children[(unsigned char) *from] =
set_map_string(root->children[(unsigned char) *from], nmem,
- from + 1, len - 1, to)))
+ from + 1, len - 1, to, from_0)))
return 0;
}
return root;
return t->target ? t : 0;
}
-static chr_t_entry *find_entry_x(chr_t_entry *t, const char **from, int *len)
+static chr_t_entry *find_entry_x(chr_t_entry *t, const char **from, int *len, int first)
{
chr_t_entry *res;
from++;
len++;
}
- if (*len > 0 && t->children && t->children[(unsigned char) **from])
+ if (*len > 0 && t->children)
{
const char *old_from = *from;
int old_len = *len;
+
+ res = 0;
+
+ if (first && t->children[CHR_FIELD_BEGIN])
+ {
+ if ((res = find_entry_x(t->children[CHR_FIELD_BEGIN], from, len, 0)) && res != t->children[CHR_FIELD_BEGIN])
+ return res;
+ else
+ res = 0;
+ /* otherwhise there was no match on beginning of field, move on */
+ }
- (*len)--;
- (*from)++;
- if ((res = find_entry_x(t->children[(unsigned char) *old_from],
- from, len)))
- return res;
- /* no match */
- *len = old_len;
- *from = old_from;
+ if (!res && t->children[(unsigned char) **from])
+ {
+ (*len)--;
+ (*from)++;
+ if ((res = find_entry_x(t->children[(unsigned char) *old_from],
+ from, len, 0)))
+ return res;
+ /* no match */
+ *len = old_len;
+ *from = old_from;
+ }
}
/* no children match. use ourselves, if we have a target */
return t->target ? t : 0;
}
-const char **chr_map_input_x(chrmaptab maptab, const char **from, int *len)
+const char **chr_map_input_x(chrmaptab maptab, const char **from, int *len, int first)
{
chr_t_entry *t = maptab->input;
chr_t_entry *res;
- if (!(res = find_entry_x(t, from, len)))
+ if (!(res = find_entry_x(t, from, len, first)))
abort();
return (const char **) (res->target);
}
-const char **chr_map_input(chrmaptab maptab, const char **from, int len)
+const char **chr_map_input(chrmaptab maptab, const char **from, int len, int first)
{
chr_t_entry *t = maptab->input;
chr_t_entry *res;
len_tmp[0] = len;
len_tmp[1] = -1;
- if (!(res = find_entry_x(t, from, len_tmp)))
+ if (!(res = find_entry_x(t, from, len_tmp, first)))
abort();
return (const char **) (res->target);
}
unsigned char zebra_prim(char **s)
{
unsigned char c;
- unsigned int i;
-
+ unsigned int i = 0;
+
+ yaz_log (YLOG_DEBUG, "prim %.3s", *s);
+ if (**s == '\\')
+ {
+ (*s)++;
+ c = **s;
+ switch (c)
+ {
+ case '\\': c = '\\'; (*s)++; break;
+ case 'r': c = '\r'; (*s)++; break;
+ case 'n': c = '\n'; (*s)++; break;
+ case 't': c = '\t'; (*s)++; break;
+ case 's': c = ' '; (*s)++; break;
+ case 'x': sscanf(*s, "x%2x", &i); c = i; *s += 3; break;
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ case '8':
+ case '9':
+ sscanf(*s, "%3o", &i);
+ c = i;
+ *s += 3;
+ break;
+ default:
+ (*s)++;
+ }
+ }
+ else
+ {
+ c = **s;
+ ++(*s);
+ }
+ return c;
+}
+
+static int zebra_ucs4_strlen(ucs4_t *s)
+{
+ int i = 0;
+ while (*s++)
+ i++;
+ return i;
+}
+
+ucs4_t zebra_prim_w(ucs4_t **s)
+{
+ ucs4_t c;
+ ucs4_t i = 0;
+ char fmtstr[8];
+
+ yaz_log (YLOG_DEBUG, "prim_w %.3s", (char *) *s);
if (**s == '\\')
{
(*s)++;
case 'n': c = '\n'; (*s)++; break;
case 't': c = '\t'; (*s)++; break;
case 's': c = ' '; (*s)++; break;
- case 'x': sscanf(*s, "x%2x", &i); c = i; *s += 3; break;
+ case 'x':
+ if (zebra_ucs4_strlen(*s) >= 3)
+ {
+ fmtstr[0] = (*s)[1];
+ fmtstr[1] = (*s)[2];
+ fmtstr[2] = 0;
+ sscanf(fmtstr, "%x", &i);
+ c = i;
+ *s += 3;
+ }
+ break;
case '0':
case '1':
case '2':
case '7':
case '8':
case '9':
- sscanf(*s, "%3o", &i);
- c = i;
- *s += 3;
+ if (zebra_ucs4_strlen(*s) >= 3)
+ {
+ fmtstr[0] = (*s)[0];
+ fmtstr[1] = (*s)[1];
+ fmtstr[2] = (*s)[2];
+ fmtstr[3] = 0;
+ sscanf(fmtstr, "%o", &i);
+ c = i;
+ *s += 3;
+ }
break;
+ case 'L':
+ if (zebra_ucs4_strlen(*s) >= 5)
+ {
+ fmtstr[0] = (*s)[1];
+ fmtstr[1] = (*s)[2];
+ fmtstr[2] = (*s)[3];
+ fmtstr[3] = (*s)[4];
+ fmtstr[4] = 0;
+ sscanf(fmtstr, "%x", &i);
+ c = i;
+ *s += 5;
+ }
+ break;
default:
(*s)++;
}
- return c;
}
- c = **s;
- ++(*s);
+ else
+ {
+ c = **s;
+ ++(*s);
+ }
+ yaz_log (YLOG_DEBUG, "out %d", c);
return c;
}
char tmp[2];
tmp[0] = num; tmp[1] = '\0';
- tab->input = set_map_string(tab->input, tab->nmem, s, strlen(s), tmp);
+ tab->input = set_map_string(tab->input, tab->nmem, s, strlen(s), tmp, 0);
tab->output[num + tab->base_uppercase] =
(unsigned char *) nmem_strdup(tab->nmem, s);
}
{
chrmaptab tab = (chrmaptab) data;
tab->input = set_map_string(tab->input, tab->nmem, s, strlen(s),
- (char*) CHR_SPACE);
+ (char*) CHR_SPACE, 0);
+}
+
+/*
+ * Callback function.
+ * Add a space-entry to the value space.
+ */
+static void fun_addcut(const char *s, void *data, int num)
+{
+ chrmaptab tab = (chrmaptab) data;
+ tab->input = set_map_string(tab->input, tab->nmem, s, strlen(s),
+ (char*) CHR_CUT, 0);
}
/*
chrwork *arg = (chrwork *) data;
const char **res, *p = s;
- res = chr_map_input(arg->map, &s, strlen(s));
+ res = chr_map_input(arg->map, &s, strlen(s), 0);
if (*res == (char*) CHR_UNKNOWN)
- logf(LOG_WARN, "Map: '%s' has no mapping", p);
+ yaz_log(YLOG_WARN, "Map: '%s' has no mapping", p);
strncat(arg->string, *res, CHR_MAXSTR - strlen(arg->string));
arg->string[CHR_MAXSTR] = '\0';
}
chrwork *arg = (chrwork *) data;
assert(arg->map->input);
- logf (LOG_DEBUG, "set map %.*s", (int) strlen(s), s);
- set_map_string(arg->map->input, arg->map->nmem, s, strlen(s), arg->string);
+ yaz_log (YLOG_DEBUG, "set map %.*s", (int) strlen(s), s);
+ set_map_string(arg->map->input, arg->map->nmem, s, strlen(s), arg->string,
+ 0);
for (s = arg->string; *s; s++)
- logf (LOG_DEBUG, " %3d", (unsigned char) *s);
+ yaz_log (YLOG_DEBUG, " %3d", (unsigned char) *s);
}
/*
chrwork *arg = (chrwork *) data;
assert(arg->map->q_input);
- logf (LOG_DEBUG, "set qmap %.*s", (int) strlen(s), s);
+ yaz_log (YLOG_DEBUG, "set qmap %.*s", (int) strlen(s), s);
set_map_string(arg->map->q_input, arg->map->nmem, s,
- strlen(s), arg->string);
+ strlen(s), arg->string, 0);
for (s = arg->string; *s; s++)
- logf (LOG_DEBUG, " %3d", (unsigned char) *s);
+ yaz_log (YLOG_DEBUG, " %3d", (unsigned char) *s);
}
+static int scan_to_utf8 (yaz_iconv_t t, ucs4_t *from, size_t inlen,
+ char *outbuf, size_t outbytesleft)
+{
+ size_t inbytesleft = inlen * sizeof(ucs4_t);
+ char *inbuf = (char*) from;
+ size_t ret;
+
+ if (t == 0)
+ *outbuf++ = *from; /* ISO-8859-1 is OK here */
+ else
+ {
+ ret = yaz_iconv (t, &inbuf, &inbytesleft, &outbuf, &outbytesleft);
+ if (ret == (size_t) (-1))
+ {
+ yaz_log(YLOG_LOG, "from: %2X %2X %2X %2X",
+ from[0], from[1], from[2], from[3]);
+ yaz_log (YLOG_WARN|YLOG_ERRNO, "bad unicode sequence");
+ return -1;
+ }
+ }
+ *outbuf = '\0';
+ return 0;
+}
-static int scan_string(char *s,
+static int scan_string(char *s_native,
+ yaz_iconv_t t_unicode, yaz_iconv_t t_utf8,
void (*fun)(const char *c, void *data, int num),
void *data, int *num)
{
- unsigned char c, str[1024], begin, end, *p;
-
+ char str[1024];
+
+ ucs4_t arg[512];
+ ucs4_t arg_prim[512];
+ ucs4_t *s0, *s = arg;
+ ucs4_t c, begin, end;
+ size_t i;
+
+ if (t_unicode != 0)
+ {
+ char *outbuf = (char *) arg;
+ char *inbuf = s_native;
+ size_t outbytesleft = sizeof(arg)-4;
+ size_t inbytesleft = strlen(s_native);
+ size_t ret;
+ ret = yaz_iconv(t_unicode, &inbuf, &inbytesleft,
+ &outbuf, &outbytesleft);
+ if (ret == (size_t)(-1))
+ return -1;
+ i = (outbuf - (char*) arg)/sizeof(ucs4_t);
+ }
+ else
+ {
+ for (i = 0; s_native[i]; i++)
+ arg[i] = s_native[i] & 255; /* ISO-8859-1 conversion */
+ }
+ arg[i] = 0; /* terminate */
+ if (s[0] == 0xfeff || s[0] == 0xfeff) /* skip byte Order Mark */
+ s++;
while (*s)
{
switch (*s)
{
case '{':
s++;
- begin = zebra_prim(&s);
+ begin = zebra_prim_w(&s);
if (*s != '-')
{
- logf(LOG_FATAL, "Bad range in char-map");
+ yaz_log(YLOG_FATAL, "Bad range in char-map");
return -1;
}
s++;
- end = zebra_prim(&s);
+ end = zebra_prim_w(&s);
if (end <= begin)
{
- logf(LOG_FATAL, "Bad range in char-map");
+ yaz_log(YLOG_FATAL, "Bad range in char-map");
return -1;
}
s++;
for (c = begin; c <= end; c++)
{
- str[0] = c; str[1] = '\0';
- (*fun)((char *) str, data, num ? (*num)++ : 0);
+ if (scan_to_utf8 (t_utf8, &c, 1, str, sizeof(str)-1))
+ return -1;
+ (*fun)(str, data, num ? (*num)++ : 0);
}
break;
case '[': s++; abort(); break;
case '(':
- p = (unsigned char*) ++s;
- /* Find the end-marker, ignoring escapes */
- do
- {
- if (!(p = (unsigned char*) strchr((char*) p, ')')))
- {
- logf(LOG_FATAL, "Missing ')' in string");
- return -1;
- }
- }
- while (*(p - 1) == '\\');
- *p = 0;
- (*fun)(s, data, num ? (*num)++ : 0);
- s = (char*) p + 1;
+ ++s;
+ s0 = s; i = 0;
+ while (*s != ')' || s[-1] == '\\')
+ arg_prim[i++] = zebra_prim_w(&s);
+ arg_prim[i] = 0;
+ if (scan_to_utf8 (t_utf8, arg_prim, zebra_ucs4_strlen(arg_prim), str, sizeof(str)-1))
+ return -1;
+ (*fun)(str, data, num ? (*num)++ : 0);
+ s++;
break;
default:
- c = zebra_prim(&s);
- str[0] = c; str[1] = '\0';
- (*fun)((char *) str, data, num ? (*num)++ : 0);
+ c = zebra_prim_w(&s);
+ if (scan_to_utf8 (t_utf8, &c, 1, str, sizeof(str)-1))
+ return -1;
+ (*fun)(str, data, num ? (*num)++ : 0);
}
}
return 0;
}
-chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only)
+chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only,
+ const char *tabroot)
{
FILE *f;
char line[512], *argv[50];
int errors = 0;
int argc, num = (int) *CHR_BASE, i;
NMEM nmem;
+ yaz_iconv_t t_unicode = 0;
+ yaz_iconv_t t_utf8 = 0;
+ unsigned endian = 31;
+ const char *ucs4_native = "UCS-4";
+
+ if (*(char*) &endian == 31) /* little endian? */
+ ucs4_native = "UCS-4LE";
+
+ t_utf8 = yaz_iconv_open ("UTF-8", ucs4_native);
- logf (LOG_DEBUG, "maptab %s open", name);
- if (!(f = yaz_path_fopen(tabpath, name, "r")))
+ yaz_log (YLOG_DEBUG, "maptab %s open", name);
+ if (!(f = yaz_fopen(tabpath, name, "r", tabroot)))
{
- logf(LOG_WARN|LOG_ERRNO, "%s", name);
+ yaz_log(YLOG_WARN|YLOG_ERRNO, "%s", name);
return 0;
}
nmem = nmem_create ();
{
if (argc != 2)
{
- logf(LOG_FATAL, "Syntax error in charmap");
+ yaz_log(YLOG_FATAL, "Syntax error in charmap");
++errors;
}
- if (scan_string(argv[1], fun_addentry, res, &num) < 0)
+ if (scan_string(argv[1], t_unicode, t_utf8, fun_addentry,
+ res, &num) < 0)
{
- logf(LOG_FATAL, "Bad value-set specification");
+ yaz_log(YLOG_FATAL, "Bad value-set specification");
++errors;
}
res->base_uppercase = num;
{
if (!res->base_uppercase)
{
- logf(LOG_FATAL, "Uppercase directive with no lowercase set");
+ yaz_log(YLOG_FATAL, "Uppercase directive with no lowercase set");
++errors;
}
if (argc != 2)
{
- logf(LOG_FATAL, "Missing arg for uppercase directive");
+ yaz_log(YLOG_FATAL, "Missing arg for uppercase directive");
++errors;
}
- if (scan_string(argv[1], fun_addentry, res, &num) < 0)
+ if (scan_string(argv[1], t_unicode, t_utf8, fun_addentry,
+ res, &num) < 0)
{
- logf(LOG_FATAL, "Bad value-set specification");
+ yaz_log(YLOG_FATAL, "Bad value-set specification");
++errors;
}
}
{
if (argc != 2)
{
- logf(LOG_FATAL, "Syntax error in charmap");
+ yaz_log(YLOG_FATAL, "Syntax error in charmap for space");
+ ++errors;
+ }
+ if (scan_string(argv[1], t_unicode, t_utf8,
+ fun_addspace, res, 0) < 0)
+ {
+ yaz_log(YLOG_FATAL, "Bad space specification");
+ ++errors;
+ }
+ }
+ else if (!map_only && !yaz_matchstr(argv[0], "cut"))
+ {
+ if (argc != 2)
+ {
+ yaz_log(YLOG_FATAL, "Syntax error in charmap for cut");
++errors;
}
- if (scan_string(argv[1], fun_addspace, res, 0) < 0)
+ if (scan_string(argv[1], t_unicode, t_utf8,
+ fun_addcut, res, 0) < 0)
{
- logf(LOG_FATAL, "Bad space specification");
+ yaz_log(YLOG_FATAL, "Bad cut specification");
++errors;
}
}
if (argc != 3)
{
- logf(LOG_FATAL, "charmap directive map requires 2 args");
+ yaz_log(YLOG_FATAL, "charmap directive map requires 2 args");
++errors;
}
buf.map = res;
buf.string[0] = '\0';
- if (scan_string(argv[2], fun_mkstring, &buf, 0) < 0)
+ if (scan_string(argv[2], t_unicode, t_utf8,
+ fun_mkstring, &buf, 0) < 0)
{
- logf(LOG_FATAL, "Bad map target");
+ yaz_log(YLOG_FATAL, "Bad map target");
++errors;
}
- if (scan_string(argv[1], fun_add_map, &buf, 0) < 0)
+ if (scan_string(argv[1], t_unicode, t_utf8,
+ fun_add_map, &buf, 0) < 0)
{
- logf(LOG_FATAL, "Bad map source");
+ yaz_log(YLOG_FATAL, "Bad map source");
++errors;
}
}
if (argc != 3)
{
- logf(LOG_FATAL, "charmap directive qmap requires 2 args");
+ yaz_log(YLOG_FATAL, "charmap directive qmap requires 2 args");
++errors;
}
buf.map = res;
buf.string[0] = '\0';
- if (scan_string(argv[2], fun_mkstring, &buf, 0) < 0)
+ if (scan_string(argv[2], t_unicode, t_utf8,
+ fun_mkstring, &buf, 0) < 0)
{
- logf(LOG_FATAL, "Bad qmap target");
+ yaz_log(YLOG_FATAL, "Bad qmap target");
++errors;
}
- if (scan_string(argv[1], fun_add_qmap, &buf, 0) < 0)
+ if (scan_string(argv[1], t_unicode, t_utf8,
+ fun_add_qmap, &buf, 0) < 0)
{
- logf(LOG_FATAL, "Bad qmap source");
+ yaz_log(YLOG_FATAL, "Bad qmap source");
++errors;
}
}
+ else if (!yaz_matchstr(argv[0], "encoding"))
+ {
+ /*
+ * Fix me. When t_unicode==0 and use encoding directive in *.chr file the beheviour of the
+ * zebra need to comment next part of code.
+ */
+
+ /* Original code */
+#if 1
+ if (t_unicode != 0)
+ yaz_iconv_close (t_unicode);
+ t_unicode = yaz_iconv_open (ucs4_native, argv[1]);
+#endif
+ /*
+ * Fix me. It is additional staff for conversion of characters from local encoding
+ * of *.chr file to UTF-8 (internal encoding).
+ * NOTE: The derective encoding must be first directive in *.chr file.
+ */
+ /* For whatever reason Oleg enabled this.. */
+#if 0
+ if (t_utf8 != 0)
+ yaz_iconv_close(t_utf8);
+ t_utf8 = yaz_iconv_open ("UTF-8", argv[1]);
+#endif
+ }
else
{
- logf(LOG_WARN, "Syntax error at '%s' in %s", line, name);
+ yaz_log(YLOG_WARN, "Syntax error at '%s' in %s", line, name);
}
- fclose(f);
+ yaz_fclose(f);
if (errors)
{
chrmaptab_destroy(res);
res = 0;
}
- logf (LOG_DEBUG, "maptab %s close %d errors", name, errors);
+ yaz_log (YLOG_DEBUG, "maptab %s close %d errors", name, errors);
+ if (t_utf8 != 0)
+ yaz_iconv_close(t_utf8);
+ if (t_unicode != 0)
+ yaz_iconv_close(t_unicode);
return res;
}