"real" table fails in JOINs".

This is a regression caused by the fix for Bug 18444. 
This fix removed the assignment of empty_c_string to table->db performed 
in add_table_to_list, as neither me nor anyone else knew what it was 
there for. Now we know it and it's covered with tests: the only case 
when a table database name can be empty is when the table is a derived 
table. The fix puts the assignment back but makes it a bit more explicit.

Additionally, finally drop sp.result.orig which was checked in by mistake. 


BitKeeper/deleted/.del-sp.result.orig:
  Delete: mysql-test/r/sp.result.orig
mysql-test/r/derived.result:
  Updated result file.
mysql-test/r/sp.result:
  Test results fixed (Bug#21002)
mysql-test/t/derived.test:
  New error return for the case when MULTI-DELETE tries to delete from
  a derived table: now derived tables belong to their own db (""), and
  MUTLI-DELETE can't find the correspondent table for it in the 
  DELETE list, as it can't resolve tables in different dbs by alias
  (See Bug#21148 for details)
mysql-test/t/sp.test:
  Add a test case for Bug#21002 "Derived table not selecting from a "real"
   table fails in JOINs"
sql/sp.cc:
  Make empty_c_string globally accessible.
sql/sql_class.cc:
  Add empty_c_string definition.
sql/sql_class.h:
  Add a comment for the constructor of Table_ident which is
  used for derived tables. Make sure this constructor also initializes
  the database name, not only the table name.
sql/sql_parse.cc:
  Don't call check_db_name for empty database. 
  Currently the only case when a table database name can be empty
  is when the table is a derived table.
  Report the right error if the database name is wrong (ER_WRONG_DB_NAME,
  not ER_WRONG_TABLE_NAME).