List:Commits« Previous MessageNext Message »
From:Andrei Elkin Date:November 3 2011 6:33pm
Subject:bzr push into mysql-trunk branch (andrei.elkin:3554 to 3555) Bug#13334470
View as plain text  
 3555 Andrei Elkin	2011-11-03
      Bug#13334470 - SLAVE_START_WORKERS() USES BOGUS VALUE INSTEAD OF --SLAVE_PARALLEL_WORKERS OPT
      
      The value of number of workers appeared to be not set when MTS starts automatically at server
      bootstrap.
      
      Fixed with caching of the global opt_mts_slave_parallel_workers into
      rli->opt_slave_parallel_workers similarly to the manual start slave branch.
     @ sql/rpl_slave.cc
        missed caching of the global opt_mts_slave_parallel_workers is added to the automatic
        slave start branch.

    modified:
      sql/rpl_slave.cc
 3554 Tor Didriksen	2011-11-03
      Bug#12711108 61708: INCORRECT HANDLING OF C-COMPILER THAT DOES NOT SUPPORT INLINE KEYWORD
      
      All currently supported C compilers have 'inline'.
      If the cmake test for inline fails, then there's something else going wrong,
      and doing
      #define inline
      is not the right solution: 
      changing the meaning of reserved words is undefined behaviour.

    modified:
      config.h.cmake
      configure.cmake
=== modified file 'sql/rpl_slave.cc'
--- a/sql/rpl_slave.cc	2011-10-20 13:06:58 +0000
+++ b/sql/rpl_slave.cc	2011-11-03 18:33:01 +0000
@@ -337,6 +337,8 @@ int init_slave()
   /* If server id is not set, start_slave_thread() will say it */
   if (active_mi->host[0] && !opt_skip_slave_start)
   {
+    /* same as in start_slave() cache the global var value into rli's member */
+    active_mi->rli->opt_slave_parallel_workers= opt_mts_slave_parallel_workers;
     if (start_slave_threads(1 /* need mutex */,
                             0 /* no wait for start*/,
                             active_mi,

No bundle (reason: useless for push emails).
Thread
bzr push into mysql-trunk branch (andrei.elkin:3554 to 3555) Bug#13334470Andrei Elkin7 Nov