Fixed some issues found by valgrind

(one testcase, one memory leak and some accesses to not initialized memory)


mysql-test/r/events_scheduling.result:
  Changed event timer to two seconds to not get problems with slow system or when running with valgrind
mysql-test/t/events_scheduling.test:
  Changed event timer to two seconds to not get problems with slow system or when running with valgrind
mysql-test/valgrind.supp:
  Avoid purify warnings from DBUG library (safe to do)
sql/ha_berkeley.cc:
  Fix problem with not freed memory
sql/sql_class.cc:
  Ensure that row_count is initalized (as we otherwise may access it uninitialized)
sql/sql_show.cc:
  c_ptr -> ptr to avoid accessing not initialized memory
sql/sql_yacc.yy:
  Fix to not access not initialized memory
sql/table.cc:
  Fix to not access not initialized memory
sql/time.cc:
  Fix to not access not initialized memory

1 files changed, 3 insertions, 1 deletions

diff --git a/mysql-test/t/events_scheduling.test b/mysql-test/t/events_scheduling.test
index a73d25cd8ee..c688864a1e6 100644
--- a/mysql-test/t/events_scheduling.test
+++ b/mysql-test/t/events_scheduling.test
@@ -6,7 +6,9 @@ CREATE TABLE table_3(a int);
 CREATE TABLE table_4(a int);
 CREATE TABLE T19170(s1 TIMESTAMP);
 SET GLOBAL event_scheduler=1;
-CREATE EVENT E19170 ON SCHEDULE EVERY 1 SECOND DO INSERT INTO T19170 VALUES(CURRENT_TIMESTAMP);
+# We need to have 2 to make it safe with valgrind. This is probably because
+# of when we calculate the timestamp value
+CREATE EVENT E19170 ON SCHEDULE EVERY 2 SECOND DO INSERT INTO T19170 VALUES(CURRENT_TIMESTAMP);
 CREATE EVENT two_sec ON SCHEDULE EVERY 2 SECOND DO INSERT INTO table_1 VALUES(1);
 CREATE EVENT start_n_end
   ON SCHEDULE EVERY 1 SECOND