Original revision:
------------------------------------------------------------
revno: 3817
revision-id: guilhem@mysql.com-20100108092756-k0zzf4kvx9b7bh38
parent: guilhem@mysql.com-20100107101133-hrrgcdqg508runuf
committer: Guilhem Bichot <guilhem@mysql.com>
branch nick: mysql-6.0-codebase-bugfixing
timestamp: Fri 2010-01-08 10:27:56 +0100
message:
  fix for BUG#50120 "Valgrind errors in any test, inside mysqltest"
  Problem was that as v->name[v->name_len] may be uninitialized (which is ok per se),
  it shouldn't be used in an if(). We remove this zero_the_char/restore_it logic by
  rather zero-terminating the v->name string when we create it in var_init().
------------------------------------------------------------

Original revision:

------------------------------------------------------------
revno: 3789.1.9
revision-id: serg@mysql.com-20091229134448-phe834ukzmi0k2e3
parent: serg@mysql.com-20091227081418-bgfg952gzumn1k3h
committer: Sergei Golubchik <serg@mysql.com>
branch nick: 6.0-codebase
timestamp: Tue 2009-12-29 14:44:48 +0100
message:
  better fix for Bug#48758 mysqltest crashes on sys_vars.collation_server_basic in gcov builds
  
  use setenv instead of putenv
------------------------------------------------------------

Add mysql_release.cmake intended for use by the build team (
compile options, build settings to do MySQL releases)

(handle dependent options, allow user with change WITH_READLINE to
WITH_LIBEDIT and other way around after the first compile)

Part IV: sql instrumentation

in mysqld--help test.

When mysql is compiled with different options,columns
might have different size (this depends on parameter 
names and some parameters might only be available 
when some option is switched on)

Subtraction of two unsigned months yielded a (very large) positive value.
Conversion of this to a signed value was not necessarily well defined.

Solution: do the subtraction on signed values.

(i.e if gcc version < 4)

eliminate 3 more "directory path" variables from the test's output (it was already
the case for other similar ones likes slow-query-log-file).

"set engine_condition_pushdown" is deprecated, engine condition pushdown is controlled
by a new "set optimizer_switch=engine_condition_pushdown=on|off".

fixing wrong indentation in two Makefile.am, which prevented the Anjuta IDE from parsing the MySQL tree.

when EXTRA_DEBUG is enabled. The reason for the error message is that a query 
was issued with send, but response was not read with reap.

           subquery in the select list
      
When a dependent subquery with count(distinct <col>) was 
evaluated multiple times, the Distinct_Aggregator was reused. 
However, the Aggregator was not reset, so when the subquery was
evaluated for the next record in the outer select, old dependent
info was used.
      
The fix is to clear() the existing aggregator in 
Item_sum::set_aggregator(). This ensures that the aggregator is
reevaluated with the new dependent information.