General overview:
The logic for switching to row format when binlog_format=MIXED had
numerous flaws. The underlying problem was the lack of a consistent
architecture.
General purpose of this changeset:
This changeset introduces an architecture for switching to row format
when binlog_format=MIXED. It enforces the architecture where it has
to. It leaves some bugs to be fixed later. It adds extensive tests to
verify that unsafe statements work as expected and that appropriate
errors are produced by problems with the selection of binlog format.
It was not practical to split this into smaller pieces of work.

Problem 1:
To determine the logging mode, the code has to take several parameters
into account (namely: (1) the value of binlog_format; (2) the
capabilities of the engines; (3) the type of the current statement:
normal, unsafe, or row injection). These parameters may conflict in
several ways, namely:
 - binlog_format=STATEMENT for a row injection
 - binlog_format=STATEMENT for an unsafe statement
 - binlog_format=STATEMENT for an engine only supporting row logging
 - binlog_format=ROW for an engine only supporting statement logging
 - statement is unsafe and engine does not support row logging
 - row injection in a table that does not support statement logging
 - statement modifies one table that does not support row logging and
   one that does not support statement logging
Several of these conflicts were not detected, or were detected with
an inappropriate error message. The problem of BUG#39934 was that no
appropriate error message was written for the case when an engine
only supporting row logging executed a row injection with
binlog_format=ROW. However, all above cases must be handled.
Fix 1:
Introduce new error codes (sql/share/errmsg.txt). Ensure that all
conditions are detected and handled in decide_logging_format()

Problem 2:
The binlog format shall be determined once per statement, in
decide_logging_format(). It shall not be changed before or after that.
Before decide_logging_format() is called, all information necessary to
determine the logging format must be available. This principle ensures
that all unsafe statements are handled in a consistent way.
However, this principle is not followed:
thd->set_current_stmt_binlog_row_based_if_mixed() is called in several
places, including from code executing UPDATE..LIMIT,
INSERT..SELECT..LIMIT, DELETE..LIMIT, INSERT DELAYED, and
SET @@binlog_format. After Problem 1 was fixed, that caused
inconsistencies where these unsafe statements would not print the
appropriate warnings or errors for some of the conflicts.
Fix 2:
Remove calls to THD::set_current_stmt_binlog_row_based_if_mixed() from
code executed after decide_logging_format(). Compensate by calling the
set_current_stmt_unsafe() at parse time. This way, all unsafe statements
are detected by decide_logging_format().

Problem 3:
INSERT DELAYED is not unsafe: it is logged in statement format even if
binlog_format=MIXED, and no warning is printed even if
binlog_format=STATEMENT. This is BUG#45825.
Fix 3:
Made INSERT DELAYED set itself to unsafe at parse time. This allows
decide_logging_format() to detect that a warning should be printed or
the binlog_format changed.

Problem 4:
LIMIT clause were not marked as unsafe when executed inside stored
functions/triggers/views/prepared statements. This is
BUG#45785.
Fix 4:
Make statements containing the LIMIT clause marked as unsafe at
parse time, instead of at execution time. This allows propagating
unsafe-ness to the view.

statements missed from general log

A FLUSH LOGS is added to ensure that the log info hits
the file before attempting to process.

crashes server!

The problem affects the scenario when index merge is followed by a filesort
and the sort buffer is not big enough for all the sort keys.
In this case the filesort function will read the data to the end through the 
index merge quick access method (and thus closing the cursor etc), 
but will leave the pointer to the quick select method in place.
It will then create a temporary file to hold the results of the filesort and
will add it as a sort output file (in sort.io_cache).
Note that filesort will copy the original 'sort' structure in an automatic
variable and restore it after it's done.
As a result at exiting filesort() we have a sort.io_cache filled in and 
nothing else (as a result of close of the cursors at end of reading data 
through index merge).
Now create_sort_index() will note that there is a select and will clean it up
(as it's been used already by filesort() reading the data in). While doing that
a special case in the index merge destructor will clean up the sort.io_cache,
assuming it's an output of the index merge method and is not needed anymore.
As a result the code that tries to read the data back from the filesort output 
will get no data in both memory and disk and will crash.
      
Fixed similarly to how filesort() does it : by copying the sort.io_cache structure
to a local variable, removing the pointer to the io_cache (so that it's not freed 
by QUICK_INDEX_MERGE_SELECT::~QUICK_INDEX_MERGE_SELECT) and restoring the original 
structure (together with the valid pointer) after the cleanup is done.
This is a safe thing to do because all the structures are already cleaned up by
hitting the end of the index merge's read method (QUICK_INDEX_MERGE_SELECT::get_next()) 
and the cleanup code being written in a way that tolerates repeating cleanups.

The SQL-mode PAD_CHAR_TO_FULL_LENGTH could prevent a DROP USER
statement from privileges associated with the user being dropped.
What ocurred was that reading from the User and Host fields of
the tables tables_priv or columns_priv would yield values padded
with spaces, causing a failure to match a specified user or host 
('user' != 'user     ');

The solution is to disregard the PAD_CHAR_TO_FULL_LENGTH mode
when iterating over and matching values in the privileges tables
for a DROP USER statement.

statements missed from general log

A refinement of the test in the previous patch to avoid
using sleep as a means to ensure that timestamps are
added to the log entries.

Re-enabled tests main.init_connect and rpl.rpl_init_slave.test for non-Windows
platforms.

Please remove this code upon fixing the bug.

WHERE and GROUP BY clause

Loose index scan may use range conditions on the argument of 
the MIN/MAX aggregate functions to find the beginning/end of 
the interval that satisfies the range conditions in a single go.
These range conditions may have open or closed minimum/maximum 
values. When the comparison returns 0 (equal) the code should 
check the type of the min/max values of the current interval 
and accept or reject the row based on whether the limit is 
open or not.
There was a wrong composite condition on checking this and it was
not working in all cases.
Fixed by simplifying the conditions and reversing the logic.

Backport to MySQL 5.0/1 fix by Vladislav Vaintroub:

In Vista and later and also in when using terminal services, when
server is started from  command line, client cannot connect to it
via shared memory protocol.

This is a regression introduced when  Bug#24731 was fixed.  The
reason is that client is trying to attach to shared memory using
global kernel object  namespace (all kernel objects are prefixed
with Global\). However, server started from the command line in
Vista and later will create shared memory and events using current
session namespace. Thus, client is unable to find the server and
connection fails.

The fix for the client is to first try to find server using "local"
names  (omitting Global\  prefix) and only if server is not found,
trying global namespace.

This test uses SHOW STATUS and the like, which may be unstable in the face
of logging to table, since the CSV handler is actively executing operations
and thus incrementing the counters.

Fixed by disabling logging to table for the duration of the test and restoring
it afterwards. This causes various counters to properly start counting from zero
and never advance due to CSV operations.

Needed for substitution in some tests.

with gcc 4.3.2
      
Compiling MySQL with gcc 4.3.2 and later produces a number of 
warnings, many of which are new with the recent compiler
versions.
      
This bug will be resolved in more than one patch to limit the
size of changesets. This is the first patch, fixing a number 
of the warnings, predominantly "suggest using parentheses 
around && in ||", and empty for and while bodies.

This variable is defined by default and one should not do it
directly as the socket variable is not available on Windows.

the --big-test flag is supplied. Test is too resource intensive
under normal valgrind runs (takes more than 30min on powerful
hardware).

with gcc 4.3.2

Compiling MySQL with gcc 4.3.2 and later produces a number of 
warnings, many of which are new with the recent compiler
versions.

This bug will be resolved in more than one patch to limit the
size of changesets. This is the first patch, fixing a number 
of the warnings, predominantly "suggest using parentheses 
around && in ||", and empty for and while bodies.

Under a debug run, the trace file grows to a few gigabytes.
Under valgrind, takes more then 20 minutes due to the high
number of insert statements.

variable. The problem was that THD::connect_utime could be
used without being initialized when the main thread is used
to handle connections (--thread-handling=no-threads).

Under a debug run, the trace file grows to a few gigabytes.
Under valgrind, takes more then 20 minutes due to the high
number of insert statements.

            uninitialized variable used as subscript

Grouping select from a "constant" InnoDB table (a table
of a single row) joined with other tables caused a crash.

use same (slightly unwieldy) name in all trees; fix before this version goes "public".
bless ctype to avoid upmerge conflict, le sigh.

The problem is that when a optimization of read-only transactions
(bypass 2-phase commit) was implemented, it removed the code that
reseted the XID once a transaction wasn't active anymore:

sql/sql_parse.cc:

-  bzero(&thd->transaction.stmt, sizeof(thd->transaction.stmt));
-  if (!thd->active_transaction())
-    thd->transaction.xid_state.xid.null();
+  thd->transaction.stmt.reset();

This mostly worked fine as the transaction commit and rollback
functions (in handler.cc) reset the XID once the transaction is
ended. But those functions wouldn't reset the XID in case of
a empty transaction, leading to a assertion when a new starting
a new XA transaction.

The solution is to ensure that the XID state is reset when empty
transactions are ended (by either commit or rollback). This is
achieved by reorganizing the code so that the transaction cleanup
routine is invoked whenever a transaction is ended.