There was an incomplete reset of the name resolution context, that caused 
INSERT ... SELECT ... JOIN statements to resolve not by joint row type calculated
for the join.
Removed the redundant re-initialization of the context, because 
mysql_insert_select_prepare() now correctly saves/restores the context.

into moonbone.local:/work/tmp_merge-5.0-opt-mysql

into moonbone.local:/work/tmp_merge-5.0-opt-mysql

into moonbone.local:/work/tmp_merge-4.1-opt-mysql

into  clam.ndb.mysql.com:/space/pekka/ndb/version/my50

into mysql.com:/Users/kent/mysql/bk/mysql-5.0-new

  Avoid error message trying 'windoze-dsp' in obsolete directory
compile-dist:
  Avoid error message for target 'distclean' and no Makefile

into  clam.ndb.mysql.com:/space/pekka/ndb/version/my50-bug19285

into mysql.com:/Users/kent/mysql/bk/mysql-5.0-new

  Make output less verbose
  Make temporary directory name unique
  Remove temporary directory on interrupt

  Don't try copy non existing extra/{sql_state,mysqld_error}.h

into mysql.com:/Users/kent/mysql/bk/mysql-5.0-new

  Include "sql_yacc.yy" for completeness (bug#20387)

into moonbone.local:/work/tmp_merge-5.0-opt-mysql

  After merge fix

  Changes for Netware

into moonbone.local:/work/tmp_merge-5.0-opt-mysql

  Delete: netware/mysql_install.def
Makefile.am:
  Removed obsolete "mysql_install.def"

  Fix for bug#16716 for --ps-protocol mode.
item_cmpfunc.cc:
  Fix for a memory allocation/freeing problem in agg_cmp_type() after fix
for bug#16377. Few language corrections.

into  neptunus.(none):/home/msvensson/mysql/mysql-5.0

Set default valgrind options to "-v --show-reachable=yes" and make it possible to override that with --valgrind-options 

into  mysql.com:/usr/home/ram/work/mysql-5.0

into  mysql.com:/usr/home/ram/work/mysql-5.0

into  mysql.com:/home/stewart/Documents/MySQL/5.0/main

into  mysql.com:/home/stewart/Documents/MySQL/5.0/main

into  mysql.com:/usr/home/ram/work/mysql-5.0

into serg.mylan:/usr/home/serg/Abk/mysql-5.0

Don't rely on table->null_row when no index is used - it may be a multi-table search

into  rurik.mysql.com:/home/igor/mysql-5.0-opt

into  rurik.mysql.com:/home/igor/mysql-5.0-opt

into moonbone.local:/work/18175-bug-5.0-opt

resulted in a wrong error message.

The nest_level counter indicates the depth of nesting for a subselect. It is
needed to properly resolve aggregate functions in nested subselects. Obviously 
it shouldn't be incremented for UNION parts because they have the same level of
nesting. This counter was incremented by 1 in the mysql_new_select() function
for any new select and wasn't decremented for UNION parts. This resulted in
wrongly reported error messages.

Now the nest_level counter is decremented by 1 for any union part.