Re: bzr commit into mysql-next-mr-bugfixing branch (jorgen.loland:3242) WL#4800

[Guilhem Bichot]

Me again,

Guilhem Bichot a écrit, Le 30.12.2010 12:34:
> Hello Jorgen,
> 
> Jorgen Loland a écrit, Le 21.12.2010 08:34:
>> #At 
>>
> file:///export/home/jl208045/mysql/wl4800/mysql-next-mr-opt-backporting-wl4800-revert_printquick/
> 
>> based on revid:guilhem@stripped
>>
>>  3242 Jorgen Loland    2010-12-21
>>       WL#4800:             Revert print_quick() and friends to how 
>> they are in       next-mr-opt-backporting, i.e., remove optimizer 
>> tracing from       this function since tracing of quick is performed 
>> in       TABLE_READ_PLAN::trace_basic_info() instead.
> 
>> === modified file 'sql/opt_range.cc'
>> --- a/sql/opt_range.cc    2010-12-19 14:24:03 +0000
>> +++ b/sql/opt_range.cc    2010-12-21 07:34:33 +0000
>> @@ -792,17 +792,13 @@ TABLE_READ_PLAN *get_best_disjunct_quick
>>  static
>>  TRP_GROUP_MIN_MAX *get_best_group_min_max(PARAM *param, SEL_TREE *tree,
>>                                            double read_time);
>> -#ifndef DBUG_OFF
>> +#if !defined(DBUG_OFF) || defined(OPTIMIZER_TRACE)
>>  static void print_sel_tree(PARAM *param, SEL_TREE *tree, key_map 
>> *tree_map,
>>                             const char *msg);
> 
> print_sel_tree() is defined only if DBUG_OFF is not defined, so adding 
> defined(OPTIMIZER_TRACE) to the declaration is "inconsistent".
> Same for print_ror_scans_arr().
> As for print_quick(), it's used only in DBUG_EXECUTE() so should also be 
> in just "#ifndef DBUG_OFF", for declaration and definition. Same for 
> ::dbug_dump() functions.

And when this is done, opt_range.h needn't know about OPTIMIZER_TRACE so 
needn't include opt_trace.h (which is good, as changing opt_trace.h will 
cause a recompilation of less files).