X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=util%2Fcharmap.c;h=a9ddf7258c9cecb38c191b41350fe0ca0603f649;hb=cc9f94a61cbd9dcc0df0cf7d0c7c41d2cec88189;hp=79a8210d08022dad01bfdb072371ea8cc904db77;hpb=26b8ff30091b8d7c27cc38de45af595acb0445bf;p=idzebra-moved-to-github.git diff --git a/util/charmap.c b/util/charmap.c index 79a8210..a9ddf72 100644 --- a/util/charmap.c +++ b/util/charmap.c @@ -1,21 +1,9 @@ /* - * Copyright (C) 1994, Index Data I/S + * Copyright (C) 1996-2002, Index Data * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * - * $Log: charmap.c,v $ - * 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 1.22 2002-05-03 13:46:05 adam Exp $ * */ @@ -26,47 +14,76 @@ #include #include +#include -#include -#include +#include #include -#include + +#define CHR_MAXSTR 1024 +#define CHR_MAXEQUIV 32 const char *CHR_UNKNOWN = "\001"; const char *CHR_SPACE = "\002"; const char *CHR_BASE = "\003"; -extern char *data1_tabpath; +struct chrmaptab_info +{ + chr_t_entry *input; /* mapping table for input data */ + chr_t_entry *q_input; /* mapping table for queries */ + unsigned char *output[256]; /* return mapping - for display of registers */ + int base_uppercase; /* Start of upper-case ordinals */ + NMEM nmem; +}; /* * Character map trie node. */ struct chr_t_entry { - chr_t_entry **children; /* array of children */ - unsigned char *target; /* target for this node, if any */ - unsigned char *equiv; /* equivalent to, or sumthin */ -} t_entry; + chr_t_entry **children; /* array of children */ + unsigned char **target; /* target for this node, if any */ +}; + +/* + * General argument structure for callback functions (internal use only) + */ +typedef struct chrwork +{ + chrmaptab map; + char string[CHR_MAXSTR+1]; +} chrwork; /* * Add an entry to the character map. */ -static chr_t_entry *set_map_string(chr_t_entry *root, char *from, int len, - char *to) +static chr_t_entry *set_map_string(chr_t_entry *root, NMEM nmem, + const char *from, int len, char *to, + const char *from_0) { + if (!from_0) + from_0 = from; if (!root) { - root = xmalloc(sizeof(*root)); + root = (chr_t_entry *) nmem_malloc(nmem, sizeof(*root)); root->children = 0; root->target = 0; } if (!len) { - if (!root->target || (char*) root->target == CHR_SPACE || - (char*) root->target == CHR_UNKNOWN) - root->target = (unsigned char *) xstrdup(to); - else if ((char*) to != CHR_SPACE) - logf(LOG_WARN, "Character map overlap"); + 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 (LOG_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); + root->target[1] = 0; + } } else { @@ -74,52 +91,26 @@ static chr_t_entry *set_map_string(chr_t_entry *root, char *from, int len, { int i; - root->children = xmalloc(sizeof(chr_t_entry*) * 256); + root->children = (chr_t_entry **) + nmem_malloc(nmem, sizeof(chr_t_entry*) * 256); for (i = 0; i < 256; i++) root->children[i] = 0; } if (!(root->children[(unsigned char) *from] = - set_map_string(root->children[(unsigned char) *from], from + 1, - len - 1, to))) + set_map_string(root->children[(unsigned char) *from], nmem, + from + 1, len - 1, to, from_0))) return 0; } return root; } -int chr_map_chrs(chr_t_entry *t, char **from, int len, int *read, char **to, - int max) -{ - int i = 0; - unsigned char *s; - - while (len && t->children && t->children[(unsigned char) **from]) - { - t = t->children[(unsigned char) **from]; - (*from)++; - len--; - } - /* if there were no matches, we are still at the root node, - which always has a null mapping */ - for (s = t->target; *s && max; s++) - { - **to = *s; - s++; - (*to)++; - max--; - i++; - } - return i; -} - -#if 1 - -static chr_t_entry *find_entry(chr_t_entry *t, char **from, int len) +static chr_t_entry *find_entry(chr_t_entry *t, const char **from, int len) { chr_t_entry *res; - if (t->children && t->children[(unsigned char) **from]) + if (len && t->children && t->children[(unsigned char) **from]) { - char *pos = *from; + const char *pos = *from; (*from)++; if ((res = find_entry(t->children[(unsigned char) *pos], @@ -128,97 +119,111 @@ static chr_t_entry *find_entry(chr_t_entry *t, char **from, int len) /* no match */ *from = pos; } - /* no children match. use ourselves */ - return t; + /* no children match. use ourselves, if we have a target */ + return t->target ? t : 0; } -char **chr_map_input(chr_t_entry *t, char **from, int len) +static chr_t_entry *find_entry_x(chr_t_entry *t, const char **from, int *len) { - static char *buf[2] = {0, 0}, str[2] = {0, 0}; chr_t_entry *res; - if (!t) /* no table loaded. Null mapping */ + while (*len <= 0) + { /* switch to next buffer */ + if (*len < 0) + break; + from++; + len++; + } + if (*len > 0 && t->children && t->children[(unsigned char) **from]) { - if (isalnum(**from)) - { - str[0] = **from; - buf[0] = str; - } - else - buf[0] = (char*) CHR_SPACE; + const char *old_from = *from; + int old_len = *len; + + (*len)--; (*from)++; - return buf; + if ((res = find_entry_x(t->children[(unsigned char) *old_from], + from, len))) + return res; + /* no match */ + *len = old_len; + *from = old_from; } - /* no children match. use our target string */ - if (!(res = find_entry(t, from, len))) - abort(); - buf[0] = (char *) res->target; - return buf; + /* no children match. use ourselves, if we have a target */ + return t->target ? t : 0; } -#else +const char **chr_map_input_x(chrmaptab maptab, const char **from, int *len) +{ + chr_t_entry *t = maptab->input; + chr_t_entry *res; -char **chr_map_input(chr_t_entry *t, char **from, int len) + if (!(res = find_entry_x(t, from, len))) + abort(); + return (const char **) (res->target); +} + +const char **chr_map_input(chrmaptab maptab, const char **from, int len) { - static char *buf[2] = {0, 0}, str[2] = {0, 0}; - char *start = *from; + chr_t_entry *t = maptab->input; + chr_t_entry *res; + int len_tmp[2]; - if (t) - { - while (len && t->children && t->children[(unsigned char) **from]) - { - t = t->children[(unsigned char) **from]; - (*from)++; - len--; - } - buf[0] = (char*) t->target; - } - else /* null mapping */ - { - if (isalnum(**from)) - { - str[0] = **from; - buf[0] = str; - } - else if (isspace(**from)) - buf[0] = (char*) CHR_SPACE; - else - buf[0] = (char*) CHR_UNKNOWN; - } - if (start == *from) - (*from)++; - return buf; - /* return (char*) t->target; */ + len_tmp[0] = len; + len_tmp[1] = -1; + if (!(res = find_entry_x(t, from, len_tmp))) + abort(); + return (const char **) (res->target); } -#endif +const char *chr_map_output(chrmaptab maptab, const char **from, int len) +{ + unsigned char c = ** (unsigned char **) from; + (*from)++; + return (const char*) maptab->output[c]; +} -static unsigned char prim(char **s) +unsigned char zebra_prim(char **s) { unsigned char c; - unsigned int i; + unsigned int i = 0; + yaz_log (LOG_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 '{': case '[': case '(': case '}': case ']': case ')': - (*s)++; - break; - default: sscanf(*s, "%3o", &i); c = i; *s += 3; break; + 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)++; } - return c; } - c = **s; - ++(*s); + else + { + c = **s; + ++(*s); + } + yaz_log (LOG_DEBUG, "out %d", c); return c; } @@ -226,182 +231,300 @@ static unsigned char prim(char **s) * Callback function. * Add an entry to the value space. */ -static void fun_addentry(char *s, void *data, int num) +static void fun_addentry(const char *s, void *data, int num) { - chrmaptab *tab = data; + chrmaptab tab = (chrmaptab) data; char tmp[2]; - + tmp[0] = num; tmp[1] = '\0'; - tab->input = set_map_string(tab->input, s, strlen(s), tmp); - tab->output[num + tab->base_uppercase] = (unsigned char *) xstrdup(s); + 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); } /* * Callback function. * Add a space-entry to the value space. */ -static void fun_addspace(char *s, void *data, int num) +static void fun_addspace(const char *s, void *data, int num) { - chrmaptab *tab = data; - tab->input = set_map_string(tab->input, s, strlen(s), (char*) CHR_SPACE); + chrmaptab tab = (chrmaptab) data; + tab->input = set_map_string(tab->input, tab->nmem, s, strlen(s), + (char*) CHR_SPACE, 0); } -static int scan_string(char *s, void (*fun)(char *c, void *data, int num), - void *data, int *num) +/* + * Create a string containing the mapped characters provided. + */ +static void fun_mkstring(const char *s, void *data, int num) { - unsigned char c, str[1024], begin, end, *p; + chrwork *arg = (chrwork *) data; + const char **res, *p = s; + + res = chr_map_input(arg->map, &s, strlen(s)); + if (*res == (char*) CHR_UNKNOWN) + logf(LOG_WARN, "Map: '%s' has no mapping", p); + strncat(arg->string, *res, CHR_MAXSTR - strlen(arg->string)); + arg->string[CHR_MAXSTR] = '\0'; +} + +/* + * Add a map to the string contained in the argument. + */ +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); + 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); +} + +/* + * Add a query map to the string contained in the argument. + */ +static void fun_add_qmap(const char *s, void *data, int num) +{ + 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); +} + + +static int scan_string(char *s, + void (*fun)(const char *c, void *data, int num), + void *data, int *num) +{ + unsigned char c, str[1024], begin, end, *p; + while (*s) { switch (*s) { - case '{': - s++; - begin = prim(&s); - if (*s != '-') - { - logf(LOG_FATAL, "Bad range in char-map"); - return -1; - } - s++; - end = prim(&s); - if (end <= begin) - { - logf(LOG_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); - } - break; - case '[': s++; abort(); break; - case '(': - p = (unsigned char*) ++s; + case '{': + s++; + begin = zebra_prim(&s); + if (*s != '-') + { + logf(LOG_FATAL, "Bad range in char-map"); + return -1; + } + s++; + end = zebra_prim(&s); + if (end <= begin) + { + logf(LOG_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); + } + break; + case '[': s++; abort(); break; + case '(': + p = (unsigned char*) ++s; /* Find the end-marker, ignoring escapes */ - do + do + { + if (!(p = (unsigned char*) strchr((char*) p, ')'))) { - if (!(p = (unsigned char*) strchr((char*) p, ')'))) - { - logf(LOG_FATAL, "Missing ')' in string"); - return -1; - } + logf(LOG_FATAL, "Missing ')' in string"); + return -1; } - while (*(p - 1) == '\\'); - *p = 0; - (*fun)(s, data, num ? (*num)++ : 0); - s = (char*) p + 1; - break; - default: - c = prim(&s); - str[0] = c; str[1] = '\0'; - (*fun)((char *) str, data, num ? (*num)++ : 0); + } + while (*(p - 1) == '\\'); + *p = 0; + (*fun)(s, data, num ? (*num)++ : 0); + s = (char*) p + 1; + break; + default: + c = zebra_prim(&s); + str[0] = c; str[1] = '\0'; + (*fun)((char *) str, data, num ? (*num)++ : 0); } } return 0; } -chrmaptab *chr_read_maptab(char *name) +chrmaptab chrmaptab_create(const char *tabpath, const char *name, int map_only, + const char *tabroot) { FILE *f; char line[512], *argv[50]; - chrmaptab *res = xmalloc(sizeof(*res)); + chrmaptab res; + int lineno = 0; + int errors = 0; int argc, num = (int) *CHR_BASE, i; + NMEM nmem; - if (!(f = yaz_path_fopen(data1_tabpath, name, "r"))) + logf (LOG_DEBUG, "maptab %s open", name); + if (!(f = yaz_fopen(tabpath, name, "r", tabroot))) { logf(LOG_WARN|LOG_ERRNO, "%s", name); return 0; } - res = xmalloc(sizeof(*res)); - res->input = xmalloc(sizeof(*res->input)); - res->input->target = (unsigned char*) CHR_UNKNOWN; - res->input->equiv = 0; -#if 1 - res->input->children = xmalloc(sizeof(res->input) * 256); + nmem = nmem_create (); + res = (chrmaptab) nmem_malloc(nmem, sizeof(*res)); + res->nmem = nmem; + res->input = (chr_t_entry *) nmem_malloc(res->nmem, sizeof(*res->input)); + res->input->target = (unsigned char **) + nmem_malloc(res->nmem, sizeof(*res->input->target) * 2); + res->input->target[0] = (unsigned char*) CHR_UNKNOWN; + res->input->target[1] = 0; + res->input->children = (chr_t_entry **) + nmem_malloc(res->nmem, sizeof(res->input) * 256); for (i = 0; i < 256; i++) { - res->input->children[i] = xmalloc(sizeof(*res->input)); + res->input->children[i] = (chr_t_entry *) + nmem_malloc(res->nmem, sizeof(*res->input)); res->input->children[i]->children = 0; - res->input->children[i]->target = (unsigned char*) CHR_UNKNOWN; - res->input->children[i]->equiv = 0; + 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; } -#else - res->input->children = 0; -#endif - res->query_equiv = 0; + res->q_input = (chr_t_entry *) + nmem_malloc(res->nmem, sizeof(*res->q_input)); + res->q_input->target = 0; + res->q_input->children = 0; + for (i = *CHR_BASE; i < 256; i++) res->output[i] = 0; res->output[(int) *CHR_SPACE] = (unsigned char *) " "; res->output[(int) *CHR_UNKNOWN] = (unsigned char*) "@"; res->base_uppercase = 0; - while ((argc = readconf_line(f, line, 512, argv, 50))) - if (!yaz_matchstr(argv[0], "lowercase")) + while (!errors && (argc = readconf_line(f, &lineno, line, 512, argv, 50))) + if (!map_only && !yaz_matchstr(argv[0], "lowercase")) { if (argc != 2) { logf(LOG_FATAL, "Syntax error in charmap"); - fclose(f); - return 0; + ++errors; } if (scan_string(argv[1], fun_addentry, res, &num) < 0) { logf(LOG_FATAL, "Bad value-set specification"); - fclose(f); - return 0; + ++errors; } res->base_uppercase = num; res->output[(int) *CHR_SPACE + num] = (unsigned char *) " "; res->output[(int) *CHR_UNKNOWN + num] = (unsigned char*) "@"; num = (int) *CHR_BASE; } - else if (!yaz_matchstr(argv[0], "uppercase")) + else if (!map_only && !yaz_matchstr(argv[0], "uppercase")) { if (!res->base_uppercase) { logf(LOG_FATAL, "Uppercase directive with no lowercase set"); - fclose(f); - return 0; + ++errors; } if (argc != 2) { - logf(LOG_FATAL, "Syntax error in charmap"); - fclose(f); - return 0; + logf(LOG_FATAL, "Missing arg for uppercase directive"); + ++errors; } if (scan_string(argv[1], fun_addentry, res, &num) < 0) { logf(LOG_FATAL, "Bad value-set specification"); - fclose(f); - return 0; + ++errors; } } - else if (!yaz_matchstr(argv[0], "space")) + else if (!map_only && !yaz_matchstr(argv[0], "space")) { if (argc != 2) { logf(LOG_FATAL, "Syntax error in charmap"); - fclose(f); - return 0; + ++errors; } - if (scan_string(argv[1], fun_addspace, res, 0)) + if (scan_string(argv[1], fun_addspace, res, 0) < 0) { logf(LOG_FATAL, "Bad space specification"); - fclose(f); - return 0; + ++errors; + } + } + else if (!yaz_matchstr(argv[0], "map")) + { + chrwork buf; + + if (argc != 3) + { + logf(LOG_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) + { + logf(LOG_FATAL, "Bad map target"); + ++errors; + } + if (scan_string(argv[1], fun_add_map, &buf, 0) < 0) + { + logf(LOG_FATAL, "Bad map source"); + ++errors; + } + } + else if (!yaz_matchstr(argv[0], "qmap")) + { + chrwork buf; + + if (argc != 3) + { + logf(LOG_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) + { + logf(LOG_FATAL, "Bad qmap target"); + ++errors; + } + if (scan_string(argv[1], fun_add_qmap, &buf, 0) < 0) + { + logf(LOG_FATAL, "Bad qmap source"); + ++errors; } } else { -#if 0 - logf(LOG_WARN, "Syntax error at '%s' in %s", line, file); - fclose(f); - return 0; -#endif + logf(LOG_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); return res; } + +void chrmaptab_destroy(chrmaptab tab) +{ + if (tab) + nmem_destroy (tab->nmem); +} + +