X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=dict%2Finsert.c;h=ce41de7e066b2882e3af7973b7e9dd8cba0e8c8d;hb=3d01abbe1238c82d62e4c01eb7cac760d3d53377;hp=214189d90adc5f4df818e98bebcf96bef03343a2;hpb=b879b04a092d5b00cc866cf16f755e55053d2e89;p=idzebra-moved-to-github.git diff --git a/dict/insert.c b/dict/insert.c index 214189d..ce41de7 100644 --- a/dict/insert.c +++ b/dict/insert.c @@ -1,10 +1,33 @@ /* - * Copyright (C) 1994, Index Data I/S + * Copyright (C) 1994-1996, Index Data I/S * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * * $Log: insert.c,v $ - * Revision 1.11 1995-09-04 12:33:31 adam + * Revision 1.17 1996-05-14 15:49:09 adam + * Bug fix: In function split_page. In rare cases variable best_indxp was + * referenced. + * + * Revision 1.16 1996/02/02 13:43:50 adam + * The public functions simply use char instead of Dict_char to represent + * search strings. Dict_char is used internally only. + * + * Revision 1.15 1996/02/01 20:39:59 adam + * Bug fix: insert didn't work on 8-bit characters due to unsigned char + * compares in dict_strcmp (strcmp) and signed Dict_char. Dict_char is + * unsigned now. + * + * Revision 1.14 1995/12/07 11:48:56 adam + * Insert operation obeys DICT_type = 1 (slack in page). + * Function dict_open exists if page size or magic aren't right. + * + * Revision 1.13 1995/11/28 09:06:37 adam + * Fixed potential dangling pointer. + * + * Revision 1.12 1995/09/06 10:34:44 adam + * Memcpy in clean_page edited to satisfy checkergcc. + * + * Revision 1.11 1995/09/04 12:33:31 adam * Various cleanup. YAZ util used instead. * * Revision 1.10 1994/10/05 12:16:48 adam @@ -89,7 +112,7 @@ static int split_page (Dict dict, Dict_ptr ptr, void *p) void *subp; char *info_here; Dict_ptr subptr; - int i; + int i, j; short *indxp, *best_indxp = NULL; Dict_char best_char = 0; Dict_char prev_char = 0; @@ -108,6 +131,7 @@ static int split_page (Dict dict, Dict_ptr ptr, void *p) { /* first entry met */ best_char = prev_char = dc; best_no = 1; + best_indxp = indxp; } else if (prev_char == dc) { /* same char prefix. update */ @@ -128,20 +152,23 @@ static int split_page (Dict dict, Dict_ptr ptr, void *p) if (best_no < 0) /* we didn't find any tail string entry at all! */ return -1; + j = best_indxp - (short*) p; subptr = new_page (dict, ptr, &subp); /* scan entries to see if there is a string with */ /* length 1. info_here indicates if such entry exist */ info_here = NULL; - for (indxp=best_indxp, i=0; i 0); + Dict_char dc; - info = (char*) p + *indxp; /* entry start */ - assert (*info == best_char); - slen = dict_strlen(info); + + info = (char*) p + ((short*) p)[j]; + /* entry start */ + memcpy (&dc, info, sizeof(dc)); + assert (dc == best_char); + slen = dict_strlen((Dict_char*) info); assert (slen > 0); if (slen == 1) @@ -152,7 +179,8 @@ static int split_page (Dict dict, Dict_ptr ptr, void *p) else { info1 = info+(1+slen)*sizeof(Dict_char); /* info start */ - dict_ins (dict, info+sizeof(Dict_char), subptr, *info1, info1+1); + dict_ins (dict, (Dict_char*) (info+sizeof(Dict_char)), + subptr, *info1, info1+1); dict_bf_readp (dict->dbf, ptr, &p); } } @@ -162,7 +190,7 @@ static int split_page (Dict dict, Dict_ptr ptr, void *p) } static void clean_page (Dict dict, Dict_ptr ptr, void *p, Dict_char *out, - Dict_ptr subptr, char *userinfo) + Dict_ptr subptr, char *userinfo) { char *np = xmalloc (dict->head.page_size); int i, slen, no = 0; @@ -202,7 +230,7 @@ static void clean_page (Dict dict, Dict_ptr ptr, void *p, Dict_char *out, continue; } *--indxp2 = info2 - np; - slen = (dict_strlen(info1)+1)*sizeof(Dict_char); + slen = (dict_strlen((Dict_char*) info1)+1)*sizeof(Dict_char); memcpy (info2, info1, slen); info1 += slen; info2 += slen; @@ -226,8 +254,17 @@ static void clean_page (Dict dict, Dict_ptr ptr, void *p, Dict_char *out, info2 += slen; ++no; } +#if 1 + memcpy ((char*)p+DICT_infoffset, + (char*)np+DICT_infoffset, + info2 - ((char*)np+DICT_infoffset)); + memcpy ((char*)p + ((char*)indxp2 - (char*)np), + indxp2, + ((char*) np+DICT_pagesize(dict)) - (char*)indxp2); +#else memcpy ((char*)p+DICT_infoffset, (char*)np+DICT_infoffset, DICT_pagesize(dict)-DICT_infoffset); +#endif DICT_size(p) = info2 - np; DICT_type(p) = 0; DICT_nodir(p) = no; @@ -273,7 +310,7 @@ static int dict_ins (Dict dict, const Dict_char *str, cmp = dict_strcmp((Dict_char*) info, str); if (!cmp) { - info += (dict_strlen(info)+1)*sizeof(Dict_char); + info += (dict_strlen((Dict_char*) info)+1)*sizeof(Dict_char); /* consider change of userinfo length... */ if (*info == userlen) { @@ -396,6 +433,11 @@ static int dict_ins (Dict dict, const Dict_char *str, if (DICT_size(p)+slen+userlen >= DICT_pagesize(dict) - (1+DICT_nodir(p))*sizeof(short)) /* overflow? */ { + if (DICT_type(p)) + { + clean_page (dict, ptr, p, NULL, 0, NULL); + return dict_ins (dict, str, ptr, userlen, userinfo); + } split_page (dict, ptr, p); return dict_ins (dict, str, ptr, userlen, userinfo); } @@ -436,12 +478,12 @@ static int dict_ins (Dict dict, const Dict_char *str, return 1; } -int dict_insert (Dict dict, const Dict_char *str, int userlen, void *userinfo) +int dict_insert (Dict dict, const char *str, int userlen, void *userinfo) { assert (dict->head.last > 0); if (dict->head.last == 1) - return dict_ins (dict, str, 0, userlen, userinfo); + return dict_ins (dict, (const Dict_char *) str, 0, userlen, userinfo); else - return dict_ins (dict, str, 1, userlen, userinfo); + return dict_ins (dict, (const Dict_char *) str, 1, userlen, userinfo); }