On 01/24/11 04:45 PM, Roy Lyseng wrote:
> Hi Øystein,
>
> thank you for the review.
>
> On 21.01.11 10.33, Øystein Grøvlen wrote:
>> Hi Roy,
>>
>> Thanks for the patch. I have verified that it fixes the issue at hand.
>> My only
>> question is that when prep_on_expr is later set, will not there be a
>> need to do
>> replace_subcondition on it?
>
> Inside simplify_joins(), prep_on_expr is set based on on_expr, which by
> that time has already been transformed using replace_subcondition.
> prep_on_expr is formed by calling copy_andor_structure() on on_expr,
> which preserves "leaf" predicates of the original condition and adds new
> AND and OR nodes.
>
> So I think there is no need for another replace_subcondition() call.
OK, approved. I hope you will follow Tor's advice and make the assert
condition a bit easier to grasp.
--
Øystein
