Merge work:/home/bk/mysql-4.0

into serg.mysql.com:/usr/home/serg/Abk/mysql-4.0

libmysql/Makefile.am

@@ -35,6 +35,7 @@ link_sources:
 	  set -x; \
 	  ss=`echo $(mystringsobjects) | sed "s;\.lo;.c;g"`; \
 	  ds=`echo $(dbugobjects) | sed "s;\.lo;.c;g"`; \
+	  qs=`echo $(sqlobjects) | sed "s;\.lo;.c;g"`; \
 	  ms=`echo $(mysysobjects) | sed "s;\.lo;.c;g"`; \
 	  for f in $$ss; do \
 	    rm -f $(srcdir)/$$f; \
@@ -44,6 +45,10 @@ link_sources:
 	    rm -f $(srcdir)/$$f; \
 	    @LN_CP_F@ $(srcdir)/../strings/$$f $(srcdir)/$$f; \
 	  done; \
+	  for f in $$qs; do \
+	    rm -f $(srcdir)/$$f; \
+	    @LN_CP_F@ $(srcdir)/../sql/$$f $(srcdir)/$$f; \
+	  done; \
 	  for f in $$ds; do \
 	    rm -f $(srcdir)/$$f; \
 	    @LN_CP_F@ $(srcdir)/../dbug/$$f $(srcdir)/$$f; \

libmysql/Makefile.shared

@@ -31,9 +31,12 @@ noinst_PROGRAMS = conf_to_src
 CHARSET_OBJS=@CHARSET_OBJS@
 LTCHARSET_OBJS= ${CHARSET_OBJS:.o=.lo}
 
-target_sources = libmysql.c net.c violite.c password.c \
+target_sources = libmysql.c net.c  password.c \
 			    get_password.c errmsg.c
 
+#quick easy dirty hack to make it work after Tonu's changes
+#In my opinion, violite.c really belongs into mysys - Sasha
+sqlobjects = violite.lo
 mystringsobjects =	strmov.lo strxmov.lo strnmov.lo strmake.lo strend.lo \
 			strnlen.lo strfill.lo is_prefix.lo \
 			int2str.lo str2int.lo strinstr.lo strcont.lo \
@@ -59,7 +62,8 @@ mysysobjects1 =		my_init.lo my_static.lo my_malloc.lo my_realloc.lo \
 # Not needed in the minimum library
 mysysobjects2 =		getopt.lo getopt1.lo getvar.lo my_lib.lo
 mysysobjects =		$(mysysobjects1) $(mysysobjects2)
-target_libadd = $(mysysobjects) $(mystringsobjects) $(dbugobjects)
+target_libadd = $(mysysobjects) $(mystringsobjects) $(dbugobjects)\
+ $(sqlobjects)
 target_ldflags = -version-info @SHARED_LIB_VERSION@
 CLEANFILES =		$(target_libadd) $(SHLIBOBJS) \
 			$(target)
@@ -75,6 +79,7 @@ clean-local:
 	rm -f `echo $(mystringsobjects) | sed "s;\.lo;.c;g"` \
 	      `echo $(dbugobjects) | sed "s;\.lo;.c;g"` \
 	      `echo $(mysysobjects) | sed "s;\.lo;.c;g"` \
+	      `echo $(sqlobjects) | sed "s;\.lo;.c;g"` \
 	       $(mystringsextra) $(mysysheaders) ctype_extra_sources.c \
 	       ../linked_client_sources
 

mysql-test/r/count_distinct2.result

@@ -74,3 +74,11 @@ count(distinct n2)	n1
 1	NULL
 1	1
 3	2
+count(distinct n)
+5000
+Variable_name	Value
+Created_tmp_disk_tables	1
+count(distinct s)
+5000
+Variable_name	Value
+Created_tmp_disk_tables	1

mysql-test/t/count_distinct2-master.opt

+-O max_heap_table_size=16384

mysql-test/t/count_distinct2.test

@@ -43,3 +43,32 @@ select count(distinct n1), count(distinct n2) from t1;
 
 select count(distinct n2), n1 from t1 group by n1;
 drop table t1;
+
+# test the converstion from tree to MyISAM
+create table t1 (n int);
+let $1=5000;
+while ($1)
+{
+ eval insert into t1 values($1);
+ dec $1;
+}
+
+flush status;
+select count(distinct n) from t1;
+show status like 'Created_tmp_disk_tables';
+drop table t1;
+
+#test conversion from heap to MyISAM
+create table t1 (s text);
+let $1=5000;
+while ($1)
+{
+ eval insert into t1 values('$1');
+ dec $1;
+}
+
+flush status;
+select count(distinct s) from t1;
+show status like 'Created_tmp_disk_tables';
+drop table t1;
+

sql/item_sum.cc

@@ -830,7 +830,26 @@ int composite_key_cmp(void* arg, byte* key1, byte* key2)
   return 0;
 }
 
+// helper function for walking the tree when we dump it to MyISAM -
+// tree_walk will call it for each
+// leaf
 
+int dump_leaf(byte* key, uint32 count __attribute__((unused)),
+		     Item_sum_count_distinct* item)
+{
+  char* buf = item->table->record[0];
+  int error;
+  memset(buf, 0xff, item->rec_offset); // make up for cheating in the tree
+  memcpy(buf + item->rec_offset, key, item->tree.size_of_element);
+  if ((error = item->table->file->write_row(buf)))
+  {
+        if (error != HA_ERR_FOUND_DUPP_KEY &&
+	    error != HA_ERR_FOUND_DUPP_UNIQUE)
+	  return 1;
+  }
+
+  return 0;
+}
 
 Item_sum_count_distinct::~Item_sum_count_distinct()
 {
@@ -916,11 +935,29 @@ bool Item_sum_count_distinct::setup(THD *thd)
 		key_len, compare_key, 0, 0);
       tree.cmp_arg = cmp_arg;
       use_tree = 1;
+
+      // the only time key_len could be 0 is if someone does
+      // count(distinct) on a char(0) field - stupid thing to do,
+      // but this has to be handled - otherwise someone can crash
+      // the server with a DoS attack
+      max_elements_in_tree = (key_len) ? max_heap_table_size/key_len :
+	max_heap_table_size;
     }
   
   return 0;
 }
 
+int Item_sum_count_distinct::tree_to_myisam()
+{
+  if(create_myisam_from_heap(table, tmp_table_param,
+			     HA_ERR_RECORD_FILE_FULL, 1) ||
+     tree_walk(&tree, (tree_walk_action)&dump_leaf, (void*)this,
+	       left_root_right))
+    return 1;
+  delete_tree(&tree);
+  use_tree = 0;
+  return 0;
+}
 
 void Item_sum_count_distinct::reset()
 {
@@ -947,7 +984,14 @@ bool Item_sum_count_distinct::add()
 
   if(use_tree)
     {
-      if(!tree_insert(&tree, table->record[0] + rec_offset, 0))
+      // if the tree got too big, convert to MyISAM, otherwise
+      // insert into the tree
+      if(tree.elements_in_tree > max_elements_in_tree)
+      {
+	if(tree_to_myisam())
+	  return 1;
+      }
+      else if(!tree_insert(&tree, table->record[0] + rec_offset, 0))
 	return 1;
     }
   else if ((error=table->file->write_row(table->record[0])))

sql/item_sum.h

@@ -148,14 +148,25 @@ class Item_sum_count_distinct :public Item_sum_int
   bool fix_fields(THD *thd,TABLE_LIST *tables);
   TMP_TABLE_PARAM *tmp_table_param;
   TREE tree;
-  bool use_tree; // If there are no blobs, we can use a tree, which
+  uint max_elements_in_tree;
+  // calculated based on max_heap_table_size. If reached,
+  // walk the tree and dump it into MyISAM table
+  
+  bool use_tree;
+  // If there are no blobs, we can use a tree, which
   // is faster than heap table. In that case, we still use the table
   // to help get things set up, but we insert nothing in it
-  int rec_offset; // the first few bytes of record ( at least one)
+  
+  int rec_offset;
+  // the first few bytes of record ( at least one)
   // are just markers for deleted and NULLs. We want to skip them since
   // they will just bloat the tree without providing any valuable info
 
+  int tree_to_myisam();
+  
   friend int composite_key_cmp(void* arg, byte* key1, byte* key2);
+  friend int dump_leaf(byte* key, uint32 count __attribute__((unused)),
+		Item_sum_count_distinct* item);
 
   public:
   Item_sum_count_distinct(List<Item> &list)