X-Git-Url: http://jsfdemo.indexdata.com/?a=blobdiff_plain;f=src%2Ftermlists.c;h=0435f2aae1082d5a73e887f7697a520687d5cad3;hb=c7b8fb38d1cd3568301aabac00c5627bd6f510ca;hp=6c8549597d9dde33301c5fa7737a20acd892edc2;hpb=77b4d51d81340f3374ae8a97aea0177a0fbcd336;p=pazpar2-moved-to-github.git diff --git a/src/termlists.c b/src/termlists.c index 6c85495..0435f2a 100644 --- a/src/termlists.c +++ b/src/termlists.c @@ -1,9 +1,10 @@ /* - * $Id: termlists.c,v 1.3 2007-01-08 18:32:35 quinn Exp $ + * $Id: termlists.c,v 1.6 2007-01-16 23:42:10 quinn Exp $ */ #include #include +#include #include #if HAVE_CONFIG_H @@ -69,7 +70,7 @@ struct termlist *termlist_create(NMEM nmem, int numterms, int highscore_size) hashsize <<= 1; res = nmem_malloc(nmem, sizeof(struct termlist)); res->hashtable = nmem_malloc(nmem, hashsize * sizeof(struct termlist_bucket*)); - bzero(res->hashtable, hashsize * sizeof(struct termlist_bucket*)); + memset(res->hashtable, 0, hashsize * sizeof(struct termlist_bucket*)); res->hashtable_size = hashsize; res->nmem = nmem; res->hashmask = hashsize - 1; // Creates a bitmask @@ -124,11 +125,19 @@ void termlist_insert(struct termlist *tl, const char *term) { unsigned int bucket; struct termlist_bucket **p; + char buf[256], *cp; - bucket = hash((unsigned char *)term) & tl->hashmask; + if (strlen(term) > 255) + return; + strcpy(buf, term); + for (cp = buf + strlen(buf) - 1; cp > buf && + (*cp == ',' || *cp == '.' || *cp == ' ' || *cp == '-'); cp--) + *cp = '\0'; + + bucket = hash((unsigned char *)buf) & tl->hashmask; for (p = &tl->hashtable[bucket]; *p; p = &(*p)->next) { - if (!strcmp(term, (*p)->term.term)) + if (!strcmp(buf, (*p)->term.term)) { (*p)->term.frequency++; update_highscore(tl, &((*p)->term)); @@ -139,7 +148,7 @@ void termlist_insert(struct termlist *tl, const char *term) { struct termlist_bucket *new = nmem_malloc(tl->nmem, sizeof(struct termlist_bucket)); - new->term.term = nmem_strdup(tl->nmem, term); + new->term.term = nmem_strdup(tl->nmem, buf); new->term.frequency = 1; new->next = 0; *p = new;