Re: Removing unneeded downlink field from nbtree stack struct

On Mon, Aug 12, 2019 at 9:43 AM Anastasia Lubennikova
<a(dot)lubennikova(at)postgrespro(dot)ru> wrote:
> The refactoring is clear, so I set Ready for committer status.
> I have just a couple of notes about comments:
>
> 1) I think that it's worth to add explanation of the case when we use
> right sibling to this comment:
> + * stack to work back up to the parent page. We use the
> child block
> + * number (or possibly the block number of a page to its
> right)

That appears over _bt_getstackbuf().

> 2) It took me quite some time to understand why does page deletion case
> doesn't need a lock.
> I propose to add something like "For more see comments for
> _bt_lock_branch_parent()" to this line:

I ended up removing the reference to page deletion here (actually, I
removed the general discussion about the need to keep the child page
locked). This seemed like something that was really up to the callers.

Pushed a version with that change. Thanks for the review!

--
Peter Geoghegan