CRASHES SERVER

Flushing of MERGE table or one of its child tables, which was
locked by flushing thread using LOCK TABLES, might have caused
crashes or assertion failures if the thread failed to reopen
child or parent table.
Particularly, this might have happened when another connection
killed this FLUSH TABLE statement/connection.
Also this problem might have occurred when we failed to reopen
MERGE table or one of its children when executing DDL statement
under LOCK TABLES.

The problem was caused by the fact that reopen_tables() might
have failed to reopen child table but still tried to reopen,
reattach children for and re-lock its parent. Vice versa it
might have failed to reopen parent but kept references from
children to parent around. Since reopen_tables() closes table
it has failed to reopen and therefore frees all associated
memory such dangling references led to crashes when followed.

This patch solves this problem by ensuring that we always close
parent table and all its children if we fail to reopen this
table or one of its children. Same happens if we fail to reattach
children to parent.

Affects 5.1 only.

mysql-test/r/merge.result:
  A test case for BUG#11763712.
mysql-test/t/merge.test:
  A test case for BUG#11763712.
sql/sql_base.cc:
  When flushing tables under LOCK TABLES, all locked
  and flushed tables are released and then reopened.
  It may happen that we failed to reopen some tables,
  in this case we reopen as much tables as possible.
  
  If it was not possible to reopen MERGE child, MERGE
  parent is unusable and must be removed from thread
  open tables list.
  
  If it was not possible to reopen MERGE parent, all
  MERGE child table objects are unusable as well, at
  least because their locks are handled by MERGE parent.
  They must also be removed from thread open tables
  list.
  
  In other words if it was impossible to reopen any
  object of a MERGE table or reattach child tables,
  all objects of this MERGE table must be considered
  unusable and closed.

The problem was that wrong structure of mysql.event was not detected and
the server continued to use wrongly-structured data.

The fix is to check the structure of mysql.event after opening before
any use. That makes operations with events more strict -- some operations
that might work before throw errors now. That seems to be Ok.

Another side-effect of the patch is that if mysql.event is corrupted,
unrelated DROP DATABASE statements issue an SQL warning about inability
to open mysql.event table. 

                      .editrc on linux.

MySQL client when build with libedit support ignores
.editrc at startup.

The reason for this regression was the incluison of a
safety check, issetugid(), which is not available on
some linux platforms.

Fixed by adding an equivalent check for platforms which
have get[e][u|g]id() set of functions.


cmd-line-utils/libedit/el.c:
  Bug#11757855 - 49967: built-in libedit doesn't read
                        .editrc on linux.
  
  Added function calls to check user/group IDs on linux
  systems which does not have issetugid() function.
configure.in:
  Bug#11757855 - 49967: built-in libedit doesn't read
                        .editrc on linux.
  
  Added check for getuid, geteuid, getgid, getegid
  functions.

Add extra codes to wait_until_disconnected.inc that are present in 5.5,
but not in 5.1. The missing codes cause innodb_bug59641 to fail in 5.1 on
Windows PB2 runs. The addition of those codes in 5.5 was done in
luis.soares@sun.com-20090930233215-aup3kxy4j6ltvjfp

Partitions can have different ref_length (position data length).
Removed DBUG_ASSERT which crashed debug builds when using
MAX_ROWS on some partitions.

               FAILS WITH LIBEDIT

Fixed by checking the return value of the write()
function calls and handling the open files and fd
appropriately.


cmd-line-utils/libedit/vi.c:
  BUG#12329909 - BUILDING MYSQL WITH DEBUG SUPPORT
                 FAILS WITH LIBEDIT
  
  Added a check on the return value of the write()
  function calls.

calc_daynr() function returns negative result
if malformed date with zero year and month is used.
Attempt to calculate week day on negative value
leads to crash. The fix is return NULL for
'W', 'a', 'w' specifiers if zero year and month is used.
Additional fix for calc_daynr():
--added assertion that result can not be negative
--return 0 if zero year and month is used


mysql-test/r/func_time.result:
  test case
mysql-test/t/func_time.test:
  test case
sql-common/my_time.c:
  --added assertion that result can not be negative
  --return 0 if zero year and month is used
sql/item_timefunc.cc:
  eturn NULL for 'W', 'a', 'w' specifiers
  if zero year and month is used.

Before sorting HAVING condition is split into two parts,
first part is a table related condition and the rest of is
HAVING part. Extraction of HAVING part does not take into account
the fact that some of conditions might be non-const but
have 'used_tables' == 0 (independent subqueries)
and because of that these conditions are cut off by
make_cond_for_table() function.
The fix is to use (table_map) 0 instead of used_tables in
third argument for make_cond_for_table() function.
It allows to extract elements which belong to sorted
table and in addition elements which are independend
subqueries.


mysql-test/r/having.result:
  test case
mysql-test/t/having.test:
  test case
sql/sql_select.cc:
  The fix is to use (table_map) 0 instead of used_tables in
  third argument for make_cond_for_table() function.
  It allows to extract elements which belong to sorted
  table and in addition elements which are independend
  subqueries.

Update for previous patch according to reviewers comments.

Updated the constructors for ha_partitions to use the common
init_handler_variables functions

Added use of defines for size and offset to get better readability for the code that reads
and writes the .par file. Also refactored the get_from_handler_file function.

Replaced the hardcoded 9 with 'KILL'

             FAILS ON SOLARIS

This assertion was triggered if gethostbyaddr_r cannot do a
reverse lookup on an ip address. The reason was a missing
DBUG_RETURN macro. The problem affected only debug versions of
the server.

This patch fixes the problem by replacing return with DBUG_RETURN.
No test case added.

Bug#11764671  57533: UNINITIALISED VALUES IN COPY_AND_CONVERT (SQL_STRING.CC) WITH CERTAIN CHA
When ROUND evaluates decimal result it uses Item::decimal
value as fraction value for the result. In some cases
Item::decimal is greater than real result fraction value
and uninitialised memory of result(decimal) buffer can be
used in further calculations. Issue is introduced by
Bug33143 fix. The fix is to remove erroneous assignment.


mysql-test/r/func_math.result:
  test case
mysql-test/t/func_math.test:
  test case
sql/item_func.cc:
  remove erroneous assignment

Update test case

With this combination, outoput was directed to .trace but not all
      parts of MTR was aware of this.
Replace .err with .trace at the earliest possible place

.0

The bug was fixed by the patch for bug number BUG 11763109 - 55779: SELECT
DOES NOT WORK PROPERLY IN MYSQL SERVER VERSION "5.1.42 SUSE MYSQL (Exact same
fix as was proposed for this bug.) Since the motivation for the two bug
reports was completely different, however, it still makes sense to push the
test case.

This patch contains only the test case.

Some multibyte sequences could be considered by my_mbcharlen() functions
as multibyte character but more exact my_ismbchar() does not think so.
In such a case this multibyte sequences is pushed into 'stack' buffer which
is too small to accommodate the sequence.
The fix is to allocate stack buffer in
compliance with max character length.


mysql-test/r/loaddata.result:
  test case
mysql-test/t/loaddata.test:
  test case
sql/sql_load.cc:
  allocate stack buffer in compliance with max character length.

Instead of just filtering space, filter white space (\s)
I left the default.experimental file as is, with tabs.

Valgrind warnings were caused by comparing index values to an un-initialized field.


mysql-test/r/subselect.result:
  New test cases.
mysql-test/t/subselect.test:
  New test cases.
sql/opt_sum.cc:
  Add thd to opt_sum_query enabling it to test for errors.
  If we have a non-nullable index, we cannot use it to match null values,
  since set_null() will be ignored, and we might compare uninitialized data.
sql/sql_select.cc:
  Add thd to opt_sum_query, enabling it to test for errors.
sql/sql_select.h:
  Add thd to opt_sum_query, enabling it to test for errors.

Fix for --vs-config applied
Find.pm incorrectly tested an unitialized local variable instead
  of the global, corrected.
Find.pm is also wrong in 5.5: uses a non-existent global variable. Fix when
  merging up.

There are two problems with ANALYSE():

1. Memory leak 
   it happens because do_select() can overwrite
   JOIN::procedure field(with zero value in our case) and
   JOIN destructor don't free the memory allocated for
   JOIN::procedure. The fix is to save original JOIN::procedure
   before do_select() call and restore it after do_select
   execution.

2. Wrong result
   If ANALYSE() procedure is used for the statement with LIMIT clause
   it could retrun empty result set. It happens because of missing 
   analyse::end_of_records() call. First end_send() function call
   returns NESTED_LOOP_QUERY_LIMIT and second call of end_send() with
   end_of_records flag enabled does not happen. The fix is to return
   NESTED_LOOP_OK from end_send() if procedure is active.


mysql-test/r/analyse.result:
  test case
mysql-test/t/analyse.test:
  test case
sql/sql_select.cc:
  --save original JOIN::procedure before do_select() call and
    restore it after do_select execution.
  --return NESTED_LOOP_OK from end_send() if procedure is active

When we create temporary result table for UNION
incorrect max_length for YEAR field is used and
it leads to incorrect field value and incorrect
result string length as YEAR field value calculation
depends on field length.
The fix is to use underlying item max_length for
Item_sum_hybrid::max_length intialization.


mysql-test/r/func_group.result:
  test case
mysql-test/t/func_group.test:
  test case
sql/field.cc:
  added assert
sql/item_sum.cc:
  init Item_sum_hybrid::max_length with 
  use underlying item max_length for
  INT result type.