Initial checkin of the new boolean fulltext search code

include/ft_global.h

@@ -53,8 +53,8 @@ void ft_free_stopwords(void);
 FT_DOCLIST * ft_init_search(void *, uint, byte *, uint, my_bool);
 int          ft_read_next(FT_DOCLIST *, char *);
 #define      ft_close_search(handler)    my_free(((gptr)(handler)),MYF(0))
-#define      ft_get_relevance(handler)   ((handler)->doc[(handler)->curdoc].weight)
-#define      ft_get_docid(handler)       ((handler)->doc[(handler)->curdoc].dpos)
+#define      ft_get_relevance(handler)   (((FT_DOCLIST *)(handler))->doc[((FT_DOCLIST *)(handler))->curdoc].weight)
+#define      ft_get_docid(handler)       (((FT_DOCLIST *)(handler))->doc[((FT_DOCLIST *)(handler))->curdoc].dpos)
 #define      ft_reinit_search(handler)   (((FT_DOCLIST *)(handler))->curdoc=-1)
 
 #ifdef  __cplusplus

myisam/ft_search.c

@@ -38,9 +38,9 @@ FT_DOCLIST *ft_init_search(void *info, uint keynr, byte *query,
     return NULL;
 /* black magic OFF */
 
-  if (is_boolean(query, query_len))
-    dlist=ft_boolean_search(info,keynr,query,query_len);
-  else
+//  if (is_boolean(query, query_len))
+//    dlist=ft_boolean_search(info,keynr,query,query_len);
+//  else
     dlist=ft_nlq_search(info,keynr,query,query_len);
 
   if(dlist && presort)
@@ -72,3 +72,4 @@ int ft_read_next(FT_DOCLIST *handler, char *record)
   }
   return my_errno;
 }
+

sql/item_func.cc

@@ -1903,7 +1903,7 @@ err:
   return 0;
 }
 
-double Item_func_match::val()
+double Item_func_match_nl::val()
 {
   if (ft_handler==NULL)
     init_search(1);
@@ -1922,7 +1922,7 @@ double Item_func_match::val()
   /* we'll have to find ft_relevance manually in ft_handler array */
 
   int a,b,c;
-  FT_DOC  *docs=ft_handler->doc;
+  FT_DOC  *docs=((FT_DOCLIST *)ft_handler)->doc;
   my_off_t docid=table->file->row_position();
 
   if ((null_value=(docid==HA_OFFSET_ERROR)))
@@ -1930,7 +1930,7 @@ double Item_func_match::val()
 
   // Assuming docs[] is sorted by dpos...
 
-  for (a=0, b=ft_handler->ndocs, c=(a+b)/2; b-a>1; c=(a+b)/2)
+  for (a=0, b=((FT_DOCLIST *)ft_handler)->ndocs, c=(a+b)/2; b-a>1; c=(a+b)/2)
   {
     if (docs[c].dpos > docid)
       b=c;
@@ -1941,7 +1941,6 @@ double Item_func_match::val()
     return docs[a].weight;
   else
     return 0.0;
-
 }
 
 void Item_func_match::init_search(bool no_order)
@@ -1962,16 +1961,14 @@ void Item_func_match::init_search(bool no_order)
   char tmp1[FT_QUERY_MAXLEN];
   String tmp2(tmp1,sizeof(tmp1));
 
-  // MATCH ... AGAINST (NULL) is meaningless, but possible 
+  // MATCH ... AGAINST (NULL) is meaningless, but possible
   if (!(ft_tmp=key_item()->val_str(&tmp2)))
   {
     ft_tmp=&tmp2;
     tmp2.set("",0);
   }
 
-  ft_handler=(FT_DOCLIST *)
-     table->file->ft_init_ext(key, (byte*) ft_tmp->ptr(), ft_tmp->length(),
-                              join_key && !no_order);
+  ft_handler_init(ft_tmp->ptr(), ft_tmp->length(), join_key && !no_order);
 
   if (join_key)
   {
@@ -2024,7 +2021,6 @@ bool Item_func_match::fix_fields(THD *thd,struct st_table_list *tlist)
   return 0;
 }
 
-
 bool Item_func_match::fix_index()
 {
   List_iterator_fast<Item> li(fields);

sql/item_func.h

@@ -863,30 +863,40 @@ public:
   uint key;
   bool join_key;
   Item_func_match *master;
-  FT_DOCLIST *ft_handler;
+  void * ft_handler;
 
   Item_func_match(List<Item> &a, Item *b): Item_real_func(b),
-  fields(a), table(0),  join_key(0), master(0), ft_handler(0) {}
+       fields(a), table(0),  join_key(0), master(0), ft_handler(0) {}
   ~Item_func_match()
   {
-    if (!master)
+    if (!master && ft_handler)
     {
-      if (ft_handler)
-      {
-	ft_close_search(ft_handler);
-        if(join_key)
-          table->file->ft_handler=0;
-      }
+      ft_handler_close();
+      if(join_key)
+	table->file->ft_handler=0;
     }
   }
-  const char *func_name() const { return "match"; }
+  virtual int ft_handler_init(const byte *key, uint keylen, bool presort)
+                                 { return 1; }
+  virtual int ft_handler_close() { return 1; }
   enum Functype functype() const { return FT_FUNC; }
   void update_used_tables() {}
   bool fix_fields(THD *thd,struct st_table_list *tlist);
   bool eq(const Item *) const;
-  double val();
   longlong val_int() { return val()!=0.0; }
 
   bool fix_index();
   void init_search(bool no_order);
 };
+
+class Item_func_match_nl :public Item_func_match
+{
+public:
+  Item_func_match_nl(List<Item> &a, Item *b): Item_func_match(a,b) {}
+  const char *func_name() const { return "match_NL"; }
+  double val();
+  int ft_handler_init(const byte *query, uint querylen, bool presort)
+       { ft_handler=table->file->ft_init_ext(key, query, querylen, presort); }
+  int ft_handler_close() { ft_close_search(ft_handler); ft_handler=0; }
+};
+

sql/sql_select.cc

@@ -532,7 +532,7 @@ mysql_select(THD *thd,TABLE_LIST *tables,List<Item> &fields,COND *conds,
   // No cache for MATCH
   make_join_readinfo(&join,
 		     (select_options & (SELECT_DESCRIBE |
-					SELECT_NO_JOIN_CACHE)) | 
+					SELECT_NO_JOIN_CACHE)) |
 		     (ftfuncs.elements ? SELECT_NO_JOIN_CACHE : 0));
 
   /* Need to tell Innobase that to play it safe, it should fetch all
@@ -605,7 +605,7 @@ mysql_select(THD *thd,TABLE_LIST *tables,List<Item> &fields,COND *conds,
     List_iterator_fast<Item_func_match> li(ftfuncs);
     Item_func_match *ifm;
     DBUG_PRINT("info",("Performing FULLTEXT search"));
-    thd->proc_info="FULLTEXT searching";
+    thd->proc_info="FULLTEXT search init";
 
     while ((ifm=li++))
     {
@@ -1453,15 +1453,15 @@ add_ft_keys(DYNAMIC_ARRAY *keyuse_array,
                 *arg1=(Item_func *)(func->arguments()[1]);
       if ((functype == Item_func::GE_FUNC ||
            functype == Item_func::GT_FUNC)  &&
-     	   arg0->type() == Item::FUNC_ITEM          &&
+           arg0->type() == Item::FUNC_ITEM          &&
            arg0->functype() == Item_func::FT_FUNC   &&
-           arg1->const_item() && arg1->val()>=0)
+           arg1->const_item() && arg1->val()>0)
         cond_func=(Item_func_match *) arg0;
       else if ((functype == Item_func::LE_FUNC ||
                 functype == Item_func::LT_FUNC)  &&
                 arg1->type() == Item::FUNC_ITEM          &&
                 arg1->functype() == Item_func::FT_FUNC   &&
-                arg0->const_item() && arg0->val()>=0)
+                arg0->const_item() && arg0->val()>0)
         cond_func=(Item_func_match *) arg1;
     }
   }
@@ -1473,7 +1473,7 @@ add_ft_keys(DYNAMIC_ARRAY *keyuse_array,
     {
       Item *item;
       /*
-         I', (Sergei) too lazy to implement proper recursive descent here,
+         I, (Sergei) too lazy to implement proper recursive descent here,
          and anyway, nobody will use such a stupid queries
          that will require it :-)
          May be later...
@@ -3413,7 +3413,7 @@ create_tmp_table(THD *thd,TMP_TABLE_PARAM *param,List<Item> &fields,
     temp_pool_slot = bitmap_set_next(&temp_pool);
 
   if (temp_pool_slot != MY_BIT_NONE) // we got a slot
-    sprintf(path, "%s%s_%lx_%i", mysql_tmpdir, tmp_file_prefix, 
+    sprintf(path, "%s%s_%lx_%i", mysql_tmpdir, tmp_file_prefix,
 	    current_pid, temp_pool_slot);
   else // if we run out of slots or we are not using tempool
     sprintf(path,"%s%s%lx_%lx_%x",mysql_tmpdir,tmp_file_prefix,current_pid,
@@ -5063,7 +5063,7 @@ end_write_group(JOIN *join, JOIN_TAB *join_tab __attribute__((unused)),
 
 /*****************************************************************************
 ** Remove calculation with tables that aren't yet read. Remove also tests
-** against fields that are read through key where the table is not a 
+** against fields that are read through key where the table is not a
 ** outer join table.
 ** We can't remove tests that are made against columns which are stored
 ** in sorted order.
@@ -5706,7 +5706,7 @@ static int remove_dup_with_hash_index(THD *thd, TABLE *table,
       if ((error=file->delete_row(record)))
 	goto err;
       continue;
-    } 
+    }
 
     /* copy fields to key buffer */
     field_length=field_lengths;
@@ -6242,7 +6242,7 @@ count_field_types(TMP_TABLE_PARAM *param, List<Item> &fields,
   List_iterator<Item> li(fields);
   Item *field;
 
-  param->field_count=param->sum_func_count=param->func_count= 
+  param->field_count=param->sum_func_count=param->func_count=
     param->hidden_field_count=0;
   param->quick_group=1;
   while ((field=li++))

sql/sql_yacc.yy

@@ -1549,10 +1549,10 @@ simple_expr:
 	| '{' ident expr '}'	{ $$= $3; }
         | MATCH '(' ident_list ')' AGAINST '(' expr ')'
           { Select->ftfunc_list.push_back(
-                   (Item_func_match *)($$=new Item_func_match(*$3,$7))); }
+                   (Item_func_match *)($$=new Item_func_match_nl(*$3,$7))); }
         | MATCH ident_list AGAINST '(' expr ')'
           { Select->ftfunc_list.push_back(
-                   (Item_func_match *)($$=new Item_func_match(*$2,$5))); }
+                   (Item_func_match *)($$=new Item_func_match_nl(*$2,$5))); }
 	| BINARY expr %prec NEG	{ $$= new Item_func_binary($2); }
 	| CASE_SYM opt_expr WHEN_SYM when_list opt_else END
 	  { $$= new Item_func_case(* $4, $2, $5 ) }