Re: Record a Bitmapset of non-pruned partitions - Mailing list pgsql-hackers

From David Rowley
Subject Re: Record a Bitmapset of non-pruned partitions
Date
Msg-id CAApHDvpN-fxu0JfxE=97zxqacJiFEDDF7_HwCzTJtJgAL5Vjbg@mail.gmail.com
Whole thread Raw
In response to Re: Record a Bitmapset of non-pruned partitions  (Zhihong Yu <zyu@yugabyte.com>)
List pgsql-hackers
Thanks for having a look at this.

On Mon, 2 Aug 2021 at 02:33, Zhihong Yu <zyu@yugabyte.com> wrote:
> Here we look for partitions which
> +    * might be interleaved with other partitions and set the
> +    * interleaved_parts field with the partition indexes of any partitions
> +    * which may be interleaved with another partition.
>
> The above seems a little bit repetitive. It can be shortened to remove repetition.

I agree that the word "partition" is mentioned quite a few times. The
only one I can see that could be removed is the "partition indexes"
one. Likely the details about which bit we set can be left up to the
struct field comment in partbounds.h

I've adjusted this to become:

/*
* Calculate interleaved partitions.  Here we look for partitions which
* might be interleaved with other partitions and set a bit in
* interleaved_parts for any partitions which may be interleaved with
* another partition.
*/

David



pgsql-hackers by date:

Previous
From: Tomas Vondra
Date:
Subject: Re: slab allocator performance issues
Next
From: Peter Smith
Date:
Subject: Re: Corrected documentation of data type for the logical replication message formats.