since Oct 1st

the Windows configuration file "include/config-win.h".

and fix a defect specific to 5.1 (wrong quoting).

support-files/mysql.spec.sh:
  Merge the general flag handling (embedding in blanks) from 5.0 to 5.1,
  add the 5.1 specific part (do not escape unquoted quote marks).
  Still part of the fix for bug#40546.

Fix yesterday's patch so that it also works for the last flag.
Still part of the fix for bug#40546.

support-files/mysql.spec.sh:
  The original "sed" command relied on a blank following the flag
  which should be removed for debug builds.
  This meant the last flag would not be matched.
  Change it so that the whole string is enclosed in blanks,
  demand leading and trailing blank for each flag,
  and finally strip the leading + trailing blank.

- Removed some copy/paste between debug and normal build in RPM spec
- Removed "mysql_upgrade_shell" from RPM build
- Removed use of "grep -q" in "configure.in", not portable
- Improved test to disable ABI check not to accidently run for icc

Other changes
- Added make file test targets 'test-bt-fast' and 'test-bt-debug-fast'
- Reenabled "jp" test suite run

Prevent this by modifying CFLAGS and CXXFLAGS.

Upmerge from 5.0, modified according to the different spec file structure.

support-files/mysql.spec.sh:
  Bug #40546  Debug server in RPM is built with compiler optimization
  
  Modify CFLAGS and CXXFLAGS for the duration of the debug server build
  (implemented by doing it in a subshell)
  so that they do not contain any optimization flag.
  Play it safe and prepare for both gcc and icc.

Prevent this by modifying CFLAGS and CXXFLAGS.

support-files/mysql.spec.sh:
  Bug #40546  Debug server in RPM is built with compiler optimization
  
  Modify CFLAGS and CXXFLAGS for the duration of the debug server build
  (implemented by doing it in a subshell)
  so that they do not contain any optimization flag.
  Play it safe and prepare for both gcc and icc.

Problem 1: BUG#36625: rpl_redirect doesn't do anything useful. It tests an
obsolete feature that was never fully implemented.
Fix 1: Remove rpl_redirect.
Problem 2: rpl_innodb_bug28430 and rpl_flushlog_loop are disabled despite the
bugs for which they were disabled have been fixed.
Fix 2: Re-enable rpl_innodb_bug28430 and rpl_flushlog_loop.


mysql-test/suite/rpl/r/rpl_redirect.result:
  Removed result file for obsolete test.\
mysql-test/suite/rpl/t/disabled.def:
  Re-enabled tests.
mysql-test/suite/rpl/t/rpl_redirect.test:
  Removed obsolete test.

order by

Problem was that the first index read was unordered,
and the next was ordered, resulting in use of
uninitialized data.

Solution was to use the correct variable to see if
the 'next' call should be ordered or not.

mysql-test/r/partition.result:
  Bug#40494: MYSQL server crashes on range access with
  partitioning and order by
  
  Added test case for the bug.
mysql-test/t/partition.test:
  Bug#40494: MYSQL server crashes on range access with
  partitioning and order by
  
  Added test case for the bug.
sql/ha_partition.cc:
  Bug#40494: MYSQL server crashes on range access with
  partitioning and order by
  
  Used the wrong variable to decide to continue with
  ordered or unordered scan.

Adding missing drop of created table and tidying display.


mysql-test/t/innodb_mysql.test:
  Adding drop of created table and cleaning variable display

      
When statement-based replication is used, and the
transaction isolation level is READ-COMMITTED or stricter,
InnoDB will print an error because statement-based
replication might lead to inconsistency between master
and slave databases. However, when the binary log is not
engaged, this is not an issue and an error should
not be printed.

This patch makes thd_binlog_format() return BINLOG_FORMAT_
UNSPEC when the binary log is not engaged for the given
thread.

mysql-test/t/innodb_mysql.test:
  Adding test that no error message is printed from inside
  InnoDB when the binary log is turned off.

Adding comments to test cases.

Problem: mysqld doesn't detect that enum data must be reinserted performing
'ALTER TABLE' in some cases.

Fix: reinsert data altering an enum field if enum values are changed.


mysql-test/r/alter_table.result:
  Fix for bug#23113: Different behavior on altering ENUM fields between 5.0 and 5.1
    - test result.
mysql-test/t/alter_table.test:
  Fix for bug#23113: Different behavior on altering ENUM fields between 5.0 and 5.1
    - test case.
sql/field.cc:
  Fix for bug#23113: Different behavior on altering ENUM fields between 5.0 and 5.1
    - Field_enum::is_equal() introduced, which is called to detect that a field
  is changing by 'ALTER TABLE'.
sql/field.h:
  Fix for bug#23113: Different behavior on altering ENUM fields between 5.0 and 5.1
    - Field_enum::is_equal() introduced, which is called to detect that a field
  is changing by 'ALTER TABLE'.

(null-merge, just the merge changesets, all base changes are already in)

Added test case to check the default value of @@binlog_format.

mysql-test/t/binlog_format_basic.test:
  Added test case to verify the default binlog_format: it should be
  STATEMENT in 5.1 and MIXED in 6.0.

mysql-test/r/xa.result:
  Update test case result.
mysql-test/t/xa.test:
  Drop table used for test.

The problem was that the server did not robustly handle a
unilateral roll back issued by the Resource Manager (RM)
due to a resource deadlock within the transaction branch.
By not acknowledging the roll back, the server (TM) would
eventually corrupt the XA transaction state and crash.

The solution is to mark the transaction as rollback-only
if the RM indicates that it rolled back its branch of the
transaction.

mysql-test/r/xa.result:
  Add test case result for Bug#28323
mysql-test/t/xa.test:
  Add test case for Bug#28323
sql/handler.cc:
  Reset XID only at the end of the global transaction.
sql/share/errmsg.txt:
  Add new error codes.
sql/sql_class.h:
  Remember the error reported by the Resource Manager.
sql/sql_parse.cc:
  Rollback the transaction if the Resource Manager reported
  a error and rolled back its branch of the transaction.

            
Debug builds of MySQL 5.1, 6.0 with Sun Studio 12 broke because of
use of gcc specific feature.
            
The fix is to replace __FUNCTION__ with the corresponding character string

In certain situations, a scan of the table will return the error
code HA_ERR_RECORD_DELETED, and this error code is not 
correctly caught in the Rows_log_event::find_row() function, which
causes an error to be returned for this case.

This patch fixes the problem by adding code to either ignore the
record and continuing with the next one, the the event of a table
scan, or change the error code to HA_ERR_KEY_NOT_FOUND, in the event
that a key lookup is attempted.

collation change made in 5.1.24-rc

Problem: 'CHECK TABLE ... FOR UPGRADE' did not check for 
incompatible collation changes made in MySQL 5.1.24-rc.

Fix: add the check.


sql/handler.cc:
    - check for incompatible collation changes made in 5.1.24-rc:
  bug #27877:
    utf8_general_ci
    ucs2_general_ci

This time the inclusion of <stdio.h> before "config.h" enabled legacy large
file support, seek64() and similar, on AIX breaking the compile of "gzio.c"