WL#3949, Added the optimization of rpl tests if mtr started w/o a combination and --mysqld=--binlog-format=XXX

in optimization phase
- Don't check specified binlog format if hasn't been assigned one

into  shellback.(none):/home/msvensson/mysql/mysql-5.1-rpl-wl3949

into  kindahl-laptop.dnsalias.net:/home/bk/b32580-mysql-5.1-rpl

The client program 'mysqlbinlog' crashed when trying to print a User_var_log_event holding
a floating-point value since the format specifier for my_b_printf() does not support
floating-point format specifiers.

This patch prints the floating-point number to an internal buffer, and then writes
that buffer to the output instead.

- Reorganize collect a little to make it easier to apply optimizations
  and settings to collected test cases.
- Add suite/rpl/combination file
- Rename include/set_binlog_format_x.inc to .sql since thay are run by "mysql"
 

into  mysql.com:/home/ksm/commits/mysql-5.1-new-rpl

The patch I previously pushed for this bug did not compile because
a field in class THD had been renamed.
This patch renames thd->query_error to thd->is_slave_error in
log_event_old.cc by applying the same patch to log_event_old.cc as
was previously applied to log_event.cc.

into  riska.(none):/home/sven/bk/b31581-5.1.22-rbr_slave_crash

This patch has two purposes:
(1) To refactor the code so that
   {Write|Update|Delete}_rows_log_event_old does not use code from
   {Write|Update|Delete}_rows_log_event. Before refactoring there
   was the following problem: whenever we modifed the code for new
   events, it affected the old events. This is bad, as it makes
   maintainance difficult. After refactoring, we can safely edit the
   new code without affecting old events. So, if we for instance
   modify the binary format of new events, we no longer need to worry
   about how the new code reads old events.
(2) To fix BUG#31581.

These two objectives are reached by the following changes:
 - Merged Rows_log_event into Old_rows_log_event and
   {Write|Update|Delete}_rows_log_event into
   {Write|Update|Delete}_rows_log_event_old.
 - Fixed the bug by replacing {WRITE|UPDATE|DELETE}_ROWS_EVENT by
   PRE_GA_{WRITE|UPDATE|DELETE}_ROWS_EVENT.
 - Added comments to log_event_old.h

(This patch is identical to the previously committed patch which was
a collapse of three changesets, except that it adds assert(0) to
constructors for old types of row log events that should never be
called.)

into  dl145h.mysql.com:/data0/mkindahl/mysql-5.1-new-rpl

into  koti.dsl.inet.fi:/home/elkin/MySQL/TEAM/FIXES/5.1/bug31170-rpl_innodb_bug28430_fails

Adding missing drop of user created for test case.

Pushing test case for bug only. Bug already fixed as a result of the patch for BUG#19958.

Non-deterministic parameters  of  SHOW SLAVE STATUS are masked out
by means of using the standard include-macro.

The masked-out parameters are not needed by the logics of the original
tests. What is need to demonstre that replication is not stopped remains.

into  kindahl-laptop.dnsalias.net:/home/bk/b12092-mysql-5.1-rpl

into  kindahl-laptop.dnsalias.net:/home/bk/b31611-mysql-5.1-target-5.1.22

into  kindahl-laptop.dnsalias.net:/home/bk/b28086-mysql-5.1-rpl

DBUG_OFF #ifdefs to eliminate two warnings in replication tree.

into  kindahl-laptop.dnsalias.net:/home/bkroot/mysql-5.1-rpl

into  koti.dsl.inet.fi:/home/elkin/MySQL/merge-5.1

into  koti.dsl.inet.fi:/home/elkin/MySQL/TEAM/FIXES/5.0/bug27571_asyn_killed_flags

refining non-deterministic tests.
The new Bug@32148 is in the way. Adjuting the tests to be somehow useful.
 

into  lambda.hsd1.co.comcast.net.:/home/malff/TREE/mysql-5.1-runtime

tests fixing: refreshing the results file.

In BUG#30244 added FOUND_ROWS() as an unsafe function, but that
works only in mixed mode under 5.1. There is a workaround that
can be used in statement-based mode either under 5.0 or 5.1
where the result of FOUND_ROWS() is stored into a user vari-
able and used that way instead. This will replicate correctly
even under statement-based replication, since it will write
a User_var entry to the binary log. For some other cases, the
value has to be passed explicitly.

This patch adds tests to demonstrate that the workarounds docu-
mented for statement-based replication works as advertised, and
does more extensive tests for cases that does not work under sta-
tement-based replication actually work under mixed mode by switch-
ing to row-based replication.