ps_6bdb (this one fails circa once over 4 runs). After the merge: the same PLUS mysql_client_test&sp (due to problem in fix of BUG#14845 which has also
been noticed in 5.0 too and Konstantin is looking at it) PLUS mysqlcheck (just consequence of mysql_client_test).
And a replication bugfix (thd->spcont must be taken into account when calling tables_ok(), as in 5.0,
and so some brackets were missing which changed one test).
Pekka will look at the NDB code where I just used "ul": ndb_charset test, and storage/ndb/src/kernel/vm/SimulatedBlock.cpp
and storage/ndb/src/kernel/blocks/dbtup/DbtupRoutines.cpp (SimulatedBlock.hpp automerged);
please also check storage/ndb/src/ndbapi/NdbRecAttr.cpp. Petr please check ha_tina.cc.


mysql-test/r/ndb_charset.result:
  "ul" in merge (Pekka please fix)
mysql-test/t/ndb_charset.test:
  "ul" in merge (Pekka please fix)
sql/sql_acl.cc:
  missing ()
sql/sql_parse.cc:
  If thd->spcont is true, always replicate (as in 5.0) (bad old merge probably).