retry of autoinc read semantics. We always reread the table's autoinc counter
after attempting to initialize it i.e., we want to guarantee that a read of
autoinc valus that is returned to the caller is always covered by the
AUTOINC locking mechanism.

Bug# 30907: We don't rely on *first_value to be 0 when checking whether
get_auto_increment() has been invoked for the first time in a multi-row
INSERT. We instead use trx_t::n_autoinc_rows. Initialize trx::n_autoinc_rows
inside ha_innobase::start_stmt() too.

Bug# 30888: While adding code for the low level read of the AUTOINC value
from the index, the case for MEDIUM ints which are 3 bytes was missed
triggering an assertion.

Use UNIV_MEM_ASSERT_W() instead of UNIV_MEM_ASSERT_RW().
The memory area need not be initialized.
This mistake was made in r1815.

ChangeSet@1.2536.50.1  2007-08-02 12:45:56-07:00  igor@mysql.com

Fixed bug#28404.
This patch adds cost estimation for the queries with ORDER BY / GROUP BY
and LIMIT.
If there was a ref/range access to the table whose rows were required
to be ordered in the result set the optimizer always employed this access
though a scan by a different index that was compatible with the required
order could be cheaper to produce the first L rows of the result set.
Now for such queries the optimizer makes a choice between the cheapest
ref/range accesses not compatible with the given order and index scans
compatible with it.

innodb.result: Adjusted results for test cases affected fy the fix for
bug #28404.

when building a previous version of the row. This bug is triggered when
running queries via InnoDB's internal SQL parser; when InnoDB's optimizer
selects a secondary index for the plan.

UNIV_MEM_ASSERT_RW(): New macro, to check that the contents of a memory
area is defined.

UNIV_MEM_ASSERT_W(): New macro, to check that a memory area is writable.

UNIV_MEM_ASSERT_AND_FREE(): New macro, to check that the memory is
writable before declaring it free (unwritable).  This replaces UNIV_MEM_FREE()
in many places.

mem_init_buf(): Check that the memory is writable, and declare it undefined.

mem_erase_buf(): Check that the memory is writable, and declare it freed.

trx->n_mysql_tables_in_use only after row_lock_table_for_mysql()
returns DB_SUCCESS.  A timeout on LOCK TABLES would lead to an
inconsistent state, which would cause trx_free() to print a warning.

This was later reported as Bug #31444.

dict_table_remove_from_cache(table) and thus freeing the memory
allocated for the table, copy the table name.  This avoids reading
freed memory when name == table->name.

Approved by Sunny.

database and table names.

Approved by Heikki.

ha_prototypes.h.  Enclose the definitions in ha_prototypes.h in
#ifndef UNIV_HOTBACKUP.

row_drop_table_for_mysql(): Do not mention innodb_force_recovery
when newraw is set.

ChangeSet@1.2536.10.2  2007-07-25 10:44:45+02:00  jperkin@mysql.com

Fix for bug#29641 - $CC on Open Server is set to contain arguments
for enabling threads.  However, duplicate AC_PROG_* macros in the
innobase plug.in file were resetting $CC and causing link errors.

As AC_PROG_* macros are already used in the main configure.in file
there should be no need for them to be duplicated here too.

plug.in:

Remove AC_PROG_* macros

row_drop_table_for_mysql().

on IM with Heikki.

The variable used in the tests below was introduced in r1735.

the InnoDB storage byte order, which is big-endian.

debug assertions.

holding the AUTOINC mutex and if initialization is required then we
initialize using our normal procedure. This change is related to Bug#27950.

pending and/or granted on a table. We peek at this value to determine whether
a transaction doing a simple INSERT in innodb_autoinc_lock_mode = 1, needs to
acquire the AUTOINC lock or not. This change is related to Bug# 16979.

been adjusted in r1719, or r1264 in branches/zip.

The code was related to replication and group commit, and it was
unreachable already as of MySQL 5.0.

Approved by Heikki

mem_heap_free_heap_top(): Declare the memory freed with UNIV_MEM_FREE().
Before this change, the memory freed by mem_heap_empty() was not completely
flagged free in UNIV_DEBUG_VALGRIND builds.  After this change, Valgrind
will hopefully catch all errors caught by UNIV_MEM_DEBUG.

in #ifdef UNIV_HOTBACKUP, to match the function declaration.

mysql_master_log_pos, repl_wait_binlog_name, repl_wait_binlog_pos.

Approved by Heikki.

AUTOINC locking. There are three modes, 0 for backward compatibility, 1 for the
new style locking (default, safe for statement-based replication) and
2 for no AUTOINC locking (unsafe for statement-based replication).

adding column definitions to tables.

dict_mem_table_add_col(): Add the parameter "heap" for temporary memory
allocation.  Allow it and "name" to be NULL.  These parameters are NULL
when creating dummy indexes.

dict_add_col_name(): Remove calls to ut_malloc() and ut_free().

dict_table_get_col_name(): Allow table->col_names to be NULL.

dict_table_add_system_columns(), dict_table_add_to_cache():
Add the parameter "heap".
---
Additional changes that had to be merged from branches/zip:

dict_table_add_system_columns(): New function, factored out from
dict_table_add_to_cache().

mlog_parse_index(): Add some consistency checks, and make use of
dict_table_add_system_columns().

ha_innodb.cc.  This cleanup was made possible as of r1550:

Merge changes from MySQL AB:

ChangeSet@2007-05-10 12:59:39+03:00, monty@mysql.com 
  WL#3817: Simplify string / memory area types and make things more consistent
  (first part)
  
  The following type conversions was done:
  
  - Changed byte to uchar
...

that was inadvertently added in r832.

in lock0lock.c and lock0lock.h identical.  The typo was incorrectly fixed in
r1623.

was added by someone at MySQL.

ChangeSet@1.1810.3176.1  2007-07-20 14:17:15+03:00  gkodinov@magare.gmz

Bug #29644: alter table hangs if records locked in share mode 
by long running transaction

On Windows opened files can't be deleted. There was a special
upgraded lock mode (TL_WRITE instead of TL_WRITE_ALLOW_READ) 
in ALTER TABLE to make sure nobody has the table opened
when deleting the old table in ALTER TABLE. This special mode
was causing ALTER TABLE to hang waiting on a lock inside InnoDB.
This special lock is no longer necessary as the server is 
closing the tables it needs to delete in ALTER TABLE.
Fixed by removing the special lock.
Note that this also reverses the fix for bug 17264 that deals with
another consequence of this special lock mode being used.

ha_innodb.cc@1.202.46.1  2007-07-20 14:17:14+03:00  gkodinov@magare.gmz

Bug #29644: reverse the (now excessive) fix
for bug 17264 (but leave the test case).

ChangeSet@1.2567, 2007-07-27 14:44:31+05:00, svoj@june.mysql.com +3 -0
  Merge mysql.com:/home/svoj/devel/mysql/BUG29957/mysql-5.0-engines
  into  mysql.com:/home/svoj/devel/mysql/BUG29957/mysql-5.1-engines
  MERGE: 1.1810.2871.44

One test case in innodb.test fails because of auto-increment
changes in r1562:1653:

$diff innodb.result innodb.reject
504c504
< 3	test2		this will work
---
> 4	test2		this will work

* Use standard while-loop instead of the seq command

is one test that fails as of this commit. The updated test case should be
part of the snapshot from MySQL shortly.

Fix for bug# 27950 - Init AUTOINC from delete_row().

Fix for bug# 28781 - Use value specified by MySQL, in update_row().

Summary of structural changes:
==============================
InnoDB needs to hold a table level lock for AUTOINC allocations to overcome
the non-determinism inherent in MySQL SBR for INSERT ... SELECT. In this fix 
for simple INSERT statements (including multi-value inserts), we try and avoid
acquiring the special AUTOINC table level lock unless another transaction has
already reserved the AUTOINC table level lock, in which case we fall back
to the old behavior of acquiring the AUTOINC table level lock.

The max AUTOINC value is now read directly using the low level interface
of InnoDB.