Re: A minor adjustment to get_cheapest_path_for_pathkeys

Hi,

>> I agree. I don't think the patch submitter is obliged to try to write
>> a good commit message, but people who contribute regularly or are
>> posting large stacks of complex patches are probably well-advised to
>> try. It makes life easier for committers and even for reviewers trying
>> to make sense of their patches.
>
>
> Fair point.  So I had a go at writing a commit message for this patch as
> attached.  Thanks for all the reviews.

+1 to Robert's and Andy's arguments above. IMO the problem with the
patch was that it was declared as a performance improvement. In such
cases we typically ask the authors to prove that the actual
improvement took place and that there were no degradations.

If we consider the patch marley as a refactoring that improves the
readability I see no reason not to merge it.

v2 LGTM.

-- 
Best regards,
Aleksander Alekseev