X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=util%2Fcharmap.c;h=d38be4f6d62254142e54ccd49db52137f1774f2a;hp=cbfbf25ff5235fcee950b9158bdbe189042b0225;hb=ecb3935e78cd9bcfdebafdee0834cfb1060d7b5e;hpb=8e13fd186da8a4f5c9c37b83ea3e46fb70f1f420 diff --git a/util/charmap.c b/util/charmap.c index cbfbf25..d38be4f 100644 --- a/util/charmap.c +++ b/util/charmap.c @@ -1,11 +1,26 @@ -/* - * Copyright (C) 1996-2002, Index Data - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Id: charmap.c,v 1.21 2002-04-05 12:54:29 adam Exp $ - * - */ +/* $Id: charmap.c,v 1.40 2006-05-10 08:13:44 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 @@ -14,17 +29,24 @@ #include #include +#include #include -#include +typedef unsigned ucs4_t; + #include +#include + #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 { @@ -54,6 +76,15 @@ typedef struct chrwork } chrwork; /* + * Callback for equivalent stuff + */ +typedef struct +{ + NMEM nmem; + int no_eq; + char *eq[CHR_MAXEQUIV]; +} chr_equiv_work; +/* * Add an entry to the character map. */ static chr_t_entry *set_map_string(chr_t_entry *root, NMEM nmem, @@ -70,13 +101,14 @@ static chr_t_entry *set_map_string(chr_t_entry *root, NMEM nmem, } if (!len) { - if (!root->target || !root->target[0] || strcmp(root->target[0], to)) + if (!root->target || !root->target[0] || + strcmp((const char *) root->target[0], to)) { if (from_0 && root->target && root->target[0] && root->target[0][0] && - strcmp (root->target[0], CHR_UNKNOWN)) + strcmp((const char *) root->target[0], CHR_UNKNOWN)) { - yaz_log (LOG_WARN, "duplicate entry for charmap from '%s'", + yaz_log (YLOG_WARN, "duplicate entry for charmap from '%s'", from_0); } root->target = (unsigned char **) @@ -123,7 +155,7 @@ static chr_t_entry *find_entry(chr_t_entry *t, const char **from, int len) 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; @@ -134,35 +166,49 @@ static chr_t_entry *find_entry_x(chr_t_entry *t, const char **from, int *len) 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; @@ -170,11 +216,25 @@ const char **chr_map_input(chrmaptab maptab, const char **from, int len) 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); } +const char **chr_map_q_input(chrmaptab maptab, + const char **from, int len, int first) +{ + chr_t_entry *t = maptab->q_input; + chr_t_entry *res; + int len_tmp[2]; + + len_tmp[0] = len; + len_tmp[1] = -1; + if (!(res = find_entry_x(t, from, len_tmp, first))) + return 0; + return (const char **) (res->target); +} + const char *chr_map_output(chrmaptab maptab, const char **from, int len) { unsigned char c = ** (unsigned char **) from; @@ -182,13 +242,22 @@ const char *chr_map_output(chrmaptab maptab, const char **from, int len) return (const char*) maptab->output[c]; } -unsigned char zebra_prim(char **s) +static int zebra_ucs4_strlen(ucs4_t *s) { - unsigned char c; - unsigned int i; + 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 (LOG_DEBUG, "prim %.3s", *s); - if (**s == '\\') + yaz_log (YLOG_DEBUG, "prim_w %.3s", (char *) *s); + if (**s == '\\' && 1[*s]) { (*s)++; c = **s; @@ -199,7 +268,17 @@ unsigned char zebra_prim(char **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': @@ -210,10 +289,30 @@ unsigned char zebra_prim(char **s) 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)++; } @@ -223,7 +322,7 @@ unsigned char zebra_prim(char **s) c = **s; ++(*s); } - yaz_log (LOG_DEBUG, "out %d", c); + yaz_log (YLOG_DEBUG, "out %d", c); return c; } @@ -253,6 +352,17 @@ static void fun_addspace(const char *s, void *data, int num) (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); +} + /* * Create a string containing the mapped characters provided. */ @@ -261,14 +371,26 @@ static void fun_mkstring(const char *s, void *data, int num) 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'; } /* + * Create an unmodified string (scan_string handler). + */ +static void fun_add_equivalent_string(const char *s, void *data, int num) +{ + chr_equiv_work *arg = (chr_equiv_work *) data; + + if (arg->no_eq == CHR_MAXEQUIV) + return; + arg->eq[arg->no_eq++] = nmem_strdup(arg->nmem, s); +} + +/* * Add a map to the string contained in the argument. */ static void fun_add_map(const char *s, void *data, int num) @@ -276,89 +398,121 @@ static void fun_add_map(const char *s, void *data, int num) chrwork *arg = (chrwork *) data; assert(arg->map->input); - logf (LOG_DEBUG, "set map %.*s", (int) strlen(s), s); + 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); } -/* - * Add a query map to the string contained in the argument. - */ -static void fun_add_qmap(const char *s, void *data, int num) +static int scan_to_utf8 (yaz_iconv_t t, ucs4_t *from, size_t inlen, + char *outbuf, size_t outbytesleft) { - chrwork *arg = (chrwork *) data; - - assert(arg->map->q_input); - logf (LOG_DEBUG, "set qmap %.*s", (int) strlen(s), s); - set_map_string(arg->map->q_input, arg->map->nmem, s, - strlen(s), arg->string, 0); - for (s = arg->string; *s; s++) - logf (LOG_DEBUG, " %3d", (unsigned char) *s); + 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, - const char *tabroot) +chrmaptab chrmaptab_create(const char *tabpath, const char *name, + const char *tabroot) { FILE *f; char line[512], *argv[50]; @@ -367,13 +521,23 @@ chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only, 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"; - logf (LOG_DEBUG, "maptab %s open", name); + 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; } + + if (*(char*) &endian == 31) /* little endian? */ + ucs4_native = "UCS-4LE"; + + t_utf8 = yaz_iconv_open ("UTF-8", ucs4_native); + nmem = nmem_create (); res = (chrmaptab) nmem_malloc(nmem, sizeof(*res)); res->nmem = nmem; @@ -392,15 +556,7 @@ chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only, res->input->children[i]->target = (unsigned char **) nmem_malloc (res->nmem, 2 * sizeof(unsigned char *)); res->input->children[i]->target[1] = 0; - if (map_only) - { - res->input->children[i]->target[0] = (unsigned char *) - nmem_malloc (res->nmem, 2 * sizeof(unsigned char)); - res->input->children[i]->target[0][0] = i; - res->input->children[i]->target[0][1] = 0; - } - else - res->input->children[i]->target[0] = (unsigned char*) CHR_UNKNOWN; + res->input->children[i]->target[0] = (unsigned char*) CHR_UNKNOWN; } res->q_input = (chr_t_entry *) nmem_malloc(res->nmem, sizeof(*res->q_input)); @@ -414,16 +570,17 @@ chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only, res->base_uppercase = 0; while (!errors && (argc = readconf_line(f, &lineno, line, 512, argv, 50))) - if (!map_only && !yaz_matchstr(argv[0], "lowercase")) + if (!yaz_matchstr(argv[0], "lowercase")) { 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; @@ -431,34 +588,50 @@ chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only, res->output[(int) *CHR_UNKNOWN + num] = (unsigned char*) "@"; num = (int) *CHR_BASE; } - else if (!map_only && !yaz_matchstr(argv[0], "uppercase")) + else if (!yaz_matchstr(argv[0], "uppercase")) { 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) + { + yaz_log(YLOG_FATAL, "Missing arg for uppercase directive"); + ++errors; + } + if (scan_string(argv[1], t_unicode, t_utf8, fun_addentry, + res, &num) < 0) + { + yaz_log(YLOG_FATAL, "Bad value-set specification"); ++errors; } + } + else if (!yaz_matchstr(argv[0], "space")) + { if (argc != 2) { - logf(LOG_FATAL, "Missing arg for uppercase directive"); + yaz_log(YLOG_FATAL, "Syntax error in charmap for space"); ++errors; } - if (scan_string(argv[1], fun_addentry, res, &num) < 0) + if (scan_string(argv[1], t_unicode, t_utf8, + fun_addspace, res, 0) < 0) { - logf(LOG_FATAL, "Bad value-set specification"); + yaz_log(YLOG_FATAL, "Bad space specification"); ++errors; } } - else if (!map_only && !yaz_matchstr(argv[0], "space")) + else if (!yaz_matchstr(argv[0], "cut")) { if (argc != 2) { - logf(LOG_FATAL, "Syntax error in charmap"); + 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; } } @@ -468,47 +641,105 @@ chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only, 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; } } - else if (!yaz_matchstr(argv[0], "qmap")) + else if (!yaz_matchstr(argv[0], "equivalent")) { - chrwork buf; + chr_equiv_work w; - if (argc != 3) + if (argc != 2) { - logf(LOG_FATAL, "charmap directive qmap requires 2 args"); + yaz_log(YLOG_FATAL, "equivalent requires 1 argument"); ++errors; } - buf.map = res; - buf.string[0] = '\0'; - if (scan_string(argv[2], fun_mkstring, &buf, 0) < 0) + w.nmem = res->nmem; + w.no_eq = 0; + if (scan_string(argv[1], t_unicode, t_utf8, + fun_add_equivalent_string, &w, 0) < 0) { - logf(LOG_FATAL, "Bad qmap target"); + yaz_log(YLOG_FATAL, "equivalent: invalid string"); ++errors; } - if (scan_string(argv[1], fun_add_qmap, &buf, 0) < 0) + else if (w.no_eq == 0) { - logf(LOG_FATAL, "Bad qmap source"); + yaz_log(YLOG_FATAL, "equivalent: no strings"); ++errors; } + else + { + char *result_str; + int i, slen = 5; + + /* determine length of regular expression */ + for (i = 0; inmem, slen + 5); + + /* build the regular expression */ + *result_str = '\0'; + slen = 0; + for (i = 0; iq_input, res->nmem, + w.eq[i], strlen(w.eq[i]), + result_str, 0); + } + } } + 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); } yaz_fclose(f); @@ -517,7 +748,11 @@ chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only, 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; } @@ -528,3 +763,11 @@ void chrmaptab_destroy(chrmaptab tab) } +/* + * Local variables: + * c-basic-offset: 4 + * indent-tabs-mode: nil + * End: + * vim: shiftwidth=4 tabstop=8 expandtab + */ +