Home | Back
------------------------------------------------------------
revno: 3501 [merge]
tags: clone-5.5.13-build
committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
branch nick: merge-5.5
timestamp: Tue 2011-05-10 16:24:34 +0300
message:
  weave null merge mysql-5.1->mysql-5.5
    ------------------------------------------------------------
    revno: 2661.753.58 [merge]
    committer: Georgi Kodinov <Georgi.Kodinov@Oracle.com>
    branch nick: merge-5.1
    timestamp: Tue 2011-05-10 16:21:44 +0300
    message:
      weave merge mysql-5.0->mysql-5.1
        ------------------------------------------------------------
        revno: 1810.4000.22 [merge]
        committer: Sunanda Menon <sunanda.menon@oracle.com>
        branch nick: mysql-5.0
        timestamp: Fri 2011-05-06 10:36:30 +0200
        message:
          Merge from mysql-5.0.93-release
    ------------------------------------------------------------
    revno: 2661.753.57 [merge]
    committer: Karen Langford <karen.langford@oracle.com>
    branch nick: mysql-5.1
    timestamp: Fri 2011-05-06 10:03:02 +0200
    message:
      Merge from mysql-5.1.57-release
        ------------------------------------------------------------
        revno: 2661.765.7
        tags: mysql-5.1.57
        author: karen.langford@oracle.com
        committer: Karen Langford <karen.langford@oracle.com>
        branch nick: mysql-5.1.57-release
        timestamp: Tue 2011-04-12 01:36:38 +0200
        message:
          Bug#11867664: Fix server crashes on update with join on partitioned table.
------------------------------------------------------------
revno: 3500
committer: Serge Kozlov <Serge.Kozlov@oracle.com>
branch nick: mysql-5.5
timestamp: Mon 2011-05-09 23:26:41 +0400
message:
  automerge 5.1->5.5
------------------------------------------------------------
revno: 3499
committer: Alexander Nozdrin <alexander.nozdrin@oracle.com>
branch nick: mysql-5.5-bug12362125
timestamp: Mon 2011-05-09 12:29:23 +0400
message:
  Patch for Bug#12362125 (SP INOUT HANDLING IS BROKEN FOR TEXT TYPE).
  
  Attempts to assign value to a table column from trigger by using
  NEW.column_name pseudo-variable might result in garbled data.
  That happened when:
    - the column had a BLOB-based type (e.g. TEXT)
      and
    - the value being assigned was retrieved from stored routine variable
      of the same type.
  
  The problem was that BLOB values were not copied correctly in this
  case. Instead of doing a copy of a real value, the value's representation
  in record buffer was copied. This representation is essentially a
  pointer to a buffer associated with the virtual table for routine
  variables where the real value is stored. Since this buffer got
  freed once trigger was left or could have changed its contents when
  new value was assigned to corresponding routine variable such a shallow
  copying resulted in garbled data in NEW.colum_name column.
  
  It worked in 5.1 due to a subtle bug in create_virtual_tmp_table():
    - in 5.1 create_virtual_tmp_table() returned a table which
      had db_low_byte_first == false.
    - in 5.5 and up create_virtual_tmp_table() returns a table which
      has db_low_byte_first == true.
  Actually, db_low_byte_first == false only for ISAM storage engine,
  which was deprecated and removed in 5.0.
  
  Having db_low_byte_first == false led to getting false in the
  complex condition for the 2nd "if" in field_conv(), which in turn
  led to copy-blob-behavior as a fall-back strategy:
    - to->table->s->db_low_byte_first was true (correct value)
    - from->table->s->db_low_byte_first was false (incorrect value)
  
  In 5.5 and up that condition is true, which means blob-values are
  not copied.
------------------------------------------------------------
revno: 3498
committer: Alexander Nozdrin <alexander.nozdrin@oracle.com>
branch nick: mysql-5.5-bug12374486
timestamp: Fri 2011-05-06 17:39:20 +0400
message:
  Patch for Bug#12374486 - SEVERE MEMORY LEAK IN PREPARED STATEMENTS
  THAT CALL STORED PROCEDURES.
  
  The bug was introduced by WL#4435. The problem was that if a stored
  procedure generated a few result sets with different set of columns,
  a new memory would be allocated after every EXECUTE for every
  result set.
  
  The fix is to introduce a new memory root in scope of MYSQL_STMT,
  and to store result-set metadata in that memory root.
------------------------------------------------------------
revno: 3497
committer: Alexander Nozdrin <alexander.nozdrin@oracle.com>
branch nick: mysql-5.5-bug60025
timestamp: Fri 2011-05-06 15:41:24 +0400
message:
  Patch for Bug#11848763 / 60025
  (SUBSTRING inside a stored function works too slow).
  
  The user-visible problem was that the server started to consume memory if a
  stored-routine of some sort is executed subsequently. The memory was freed
  only after the corresponding connection was closed.
  
  Technically, the problem was that the memory needed for temporary string
  conversions was allocated on the connection ("persistent") memory root,
  instead of statement one.
  
  The root cause of this problem was the incorrect patch for Bug 55744.
  That patch wrongly fixed a crash in prepared-statement-mode introduced by
  another patch. The patch for Bug 55744 used wrong condition to check if
  prepared statement mode is active (or whether the connection-scoped or
  statement-scoped memory root should be used). The thing is that for
  prepared statements such conversions should be done in the connection
  memory root, so that that the transformations of item-tree were correctly
  remembered in the PREPARE-phase.
  
  The fix is to use proper condition to detect prepared-statement-mode and
  use proper memory root.
------------------------------------------------------------
revno: 3496
committer: Alexander Nozdrin <alexander.nozdrin@oracle.com>
branch nick: mysql-5.5-bug60025
timestamp: Fri 2011-05-06 15:39:40 +0400
message:
  Preliminary patch for Bug#11848763 / 60025
  (SUBSTRING inside a stored function works too slow).
  
  Background:
    - THD classes derives from Query_arena, thus inherits the 'state'
      attribute and related operations (is_stmt_prepare() & co).
  
    - Although these operations are available in THD, they must not
      be used. THD has its own attribute to point to the active
      Query_arena -- stmt_arena.
  
    - So, instead of using thd->is_stmt_prepare(),
      thd->stmt_arena->is_stmt_prepare() must be used. This was the root
      cause of Bug 60025.
  
  This patch enforces the proper way of calling those operations.
  is_stmt_prepare() & co are declared as private operations
  in THD (thus, they are hidden from being called on THD instance).
  
  The patch tries to minimize changes in 5.5.
------------------------------------------------------------
revno: 3495
committer: Magnus Bl?udd <magnus.blaudd@oracle.com>
branch nick: 5.5
timestamp: Fri 2011-05-06 11:20:01 +0200
message:
  Add --with-debug=full support to BUILD/* scripts, this option has been lost
  between 5.1 and 5.5, it's purpose is to set the compiler flags in a way that
  does not optimize away the call stack(i.e don't use any -OX flags at all)
------------------------------------------------------------
revno: 3494
committer: Magnus Bl?udd <magnus.blaudd@oracle.com>
branch nick: 5.5
timestamp: Fri 2011-05-06 10:53:42 +0200
message:
  Merge in patch for bug 12380149
------------------------------------------------------------
revno: 3493 [merge]
author: bjorn.munch@oracle.com
committer: Bjorn Munch <bjorn.munch@oracle.com>
branch nick: mysql-5.5
timestamp: Fri 2011-05-06 10:27:04 +0200
message:
  Merge from mysql-5.5.12-release
    ------------------------------------------------------------
    revno: 3097.4.24 [merge]
    tags: mysql-5.5.12, clone-5.5.12-build
    committer: Alexander Nozdrin <alexander.nozdrin@oracle.com>
    branch nick: mysql-5.5-security
    timestamp: Fri 2011-04-08 18:07:59 +0400
    message:
      Auto-merge from mysql-5.5.