List:Commits« Previous MessageNext Message »
From:Dmitry Lenev Date:November 12 2010 1:58pm
Subject:bzr push into mysql-5.5-runtime branch (Dmitry.Lenev:3181 to 3182) Bug#54673
Bug#57006
View as plain text  
 3182 Dmitry Lenev	2010-11-12
      Follow-up for patch fixing bug #57006 "Deadlock between
      HANDLER and FLUSH TABLES WITH READ LOCK" and bug #54673
      "It takes too long to get readlock for 'FLUSH TABLES
      WITH READ LOCK'".
      
      Disable execution of flush_read_lock.test on embedded
      server. This test uses too many statements which are
      not supported by embedded server.

    modified:
      mysql-test/t/flush_read_lock.test
 3181 Konstantin Osipov	2010-11-12
      Implement a fix for Bug#57058 -- send SERVER_QUERY_WAS_SLOW over
      network when a query was slow.
      
      When a query is slow, sent a special flag to the client
      indicating this fact.
      
      Add a test case.
      Implement review comments.
     @ include/mysql_com.h
        Clear SERVER_QUERY_WAS_SLOW at end of each statement.
        Since this patch removes the technique when 
        thd->server_status is modified briefly only to
        execute my_eof(), reset more server status
        bit that may remain in the status from
        execution of the previous statement.
     @ sql/protocol.cc
        Always use thd->server_status to 
        in net_* functions to send the latest
        status to the client.
     @ sql/sp_head.cc
        Calculate if a query was slow before
        sending EOF packet.
     @ sql/sql_cursor.cc
        Remove juggling with thd->server_status.
        The extra status bits are reset at
        start of the next statement.
     @ sql/sql_db.cc
        Remove juggling with thd->server_status.
        The extra status bits are reset at
        start of the next statement.
     @ sql/sql_error.cc
        Remove m_server_status member,
        it's not really part of the Diagnostics_area.
     @ sql/sql_error.h
        Remove server_status member, it's
        not part of the Diagnostics_area.
        The associated hack is removed as well.
     @ sql/sql_parse.cc
        Do not calculate if a query was
        slow twice. Use a status flag in thd->server_status.
     @ tests/mysql_client_test.c
        Add a test case for Bug#57058.
        Check that the status is present
        at the client, when sent.

    modified:
      include/mysql_com.h
      sql/protocol.cc
      sql/sp_head.cc
      sql/sql_class.h
      sql/sql_cursor.cc
      sql/sql_db.cc
      sql/sql_error.cc
      sql/sql_error.h
      sql/sql_parse.cc
      tests/mysql_client_test.c
=== modified file 'mysql-test/t/flush_read_lock.test'
--- a/mysql-test/t/flush_read_lock.test	2010-11-11 17:11:05 +0000
+++ b/mysql-test/t/flush_read_lock.test	2010-11-12 13:57:08 +0000
@@ -7,6 +7,9 @@
 --source include/have_innodb.inc
 # We need the Debug Sync Facility.
 --source include/have_debug_sync.inc
+# Parts of this test use DDL on events, BINLOG statement and
+# other statements which are not supported in embedded server.
+-- source include/not_embedded.inc
 # Save the initial number of concurrent sessions.
 --source include/count_sessions.inc
 

No bundle (reason: useless for push emails).
Thread
bzr push into mysql-5.5-runtime branch (Dmitry.Lenev:3181 to 3182) Bug#54673Bug#57006Dmitry Lenev12 Nov