List:Commits« Previous MessageNext Message »
From:Guilhem Bichot Date:January 12 2011 9:16pm
Subject:Re: bzr commit into mysql-next-mr-bugfixing branch (tor.didriksen:3244)
WL#5257
View as plain text  
Hello,

I applied this one almost verbatim, and pushed. Indeed the diff with 
trunk shrinks by 18 kB. Thanks a lot.
Minor changes I did, below:

Tor Didriksen a écrit, Le 12.01.2011 09:01:
> #At file:///export/home/didrik/repo/next-mr-opt-backporting-wl4800/ based on
> revid:tor.didriksen@stripped
> 
>  3244 Tor Didriksen	2011-01-12
>       WL#5257 Review comments.
>       
>       remove a bunch of goto end-of-function, just return instead.
>       this also allows us to remove the extra scope introduced,
>       making the patch about 300 lines smaller compared to the original in trunk.
> 
>     modified:
>       sql/opt_range.cc
> === modified file 'sql/opt_range.cc'
> --- a/sql/opt_range.cc	2011-01-11 08:15:46 +0000
> +++ b/sql/opt_range.cc	2011-01-12 08:00:09 +0000
> @@ -10129,7 +10129,8 @@ get_best_group_min_max(PARAM *param, SEL
>        !is_agg_distinct)
>    {
>      cause= "not_group_by_or_distinct";

I deleted this line above and moved the literal into the add_alnum() 
call below.

> -    goto return_null_not_chosen_cause;
> +    trace_group.add("chosen", false).add_alnum("cause", cause);
> +    DBUG_RETURN(NULL);
>    }
>    /* Analyze the query in more detail. */
>  
> @@ -10150,7 +10151,8 @@ get_best_group_min_max(PARAM *param, SEL
>        else
>        {
>          cause= "not_applicable_aggregate_function";

I deleted this line above and moved the literal into the add_alnum() 
call below.

> -        goto return_null_not_chosen_cause;
> +        trace_group.add("chosen", false).add_alnum("cause", cause);
> +        DBUG_RETURN(NULL);
>        }
>  
>        /* The argument of MIN/MAX. */
> @@ -10185,7 +10187,8 @@ get_best_group_min_max(PARAM *param, SEL
>      if ((*tmp_group->item)->real_item()->type() != Item::FIELD_ITEM)
>      {
>        cause= "group_field_is_expression";

I deleted this line above and moved the literal into the add_alnum() 
call below.

> -      goto return_null_not_chosen_cause;
> +      trace_group.add("chosen", false).add_alnum("cause", cause);
> +      DBUG_RETURN(NULL);
>      }
>    }
>  
> @@ -10194,337 +10197,339 @@ get_best_group_min_max(PARAM *param, SEL
>      (GA1,GA2) are all TRUE. If there is more than one such index, select the
>      first one. Here we set the variables: group_prefix_len and index_info.
>    */
> -  {
> -    const uint pk= param->table->s->primary_key;
> -    KEY *cur_index_info= table->key_info;
> -    KEY *cur_index_info_end= cur_index_info + table->s->keys;
> -    SEL_ARG *cur_index_tree= NULL;
> -    ha_rows cur_quick_prefix_records= 0;
> -    uint cur_param_idx= MAX_KEY;
> -    Opt_trace_array ota(thd->opt_trace, "potential_group_range_indices");

I renamed ota to trace_indices.

> @@ -10535,7 +10540,8 @@ get_best_group_min_max(PARAM *param, SEL
>                                        Field::itMBR : Field::itRAW))
>    {
>      cause= "unsupported_predicate_on_agg_attribute";

I deleted this line above and moved the literal into the add_alnum() 
call below.

> -    goto return_null_not_usable_cause;
> +    trace_group.add("usable", false).add_alnum("cause", cause);
> +    DBUG_RETURN(NULL);
>    }
>    /* The query passes all tests, so construct a new TRP object. */
>    read_plan= new (param->mem_root)
> @@ -10568,15 +10574,11 @@ get_best_group_min_max(PARAM *param, SEL
>  
>    DBUG_RETURN(read_plan);
>  

I deleted all lines below.

> -return_null_not_chosen_cause:
> -  DBUG_ASSERT(cause);
> -  trace_group.add("chosen", false).add_alnum("cause", cause);
> -  DBUG_RETURN(NULL);
> -
> -return_null_not_usable_cause:
> -  DBUG_ASSERT(cause);
> -  trace_group.add("usable", false).add_alnum("cause", cause);
> -  DBUG_RETURN(NULL);
> +//return_null_not_chosen_cause:
> +//DBUG_ASSERT(cause);
> +
> +//return_null_not_usable_cause:
> +//DBUG_ASSERT(cause);
>  }
Thread
bzr commit into mysql-next-mr-bugfixing branch (tor.didriksen:3244) WL#5257Tor Didriksen12 Jan
  • Re: bzr commit into mysql-next-mr-bugfixing branch (tor.didriksen:3244)WL#5257Guilhem Bichot12 Jan