Date: Sun, 3 Sep 2023 08:11:16 -0700
From: Nathan Bossart <nathandbossart@gmail.com>
To: Alvaro Herrera <alvherre@alvh.no-ip.org>
Cc: Robert Haas <robertmhaas@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>,
	Andrew Dunstan <andrew@dunslane.net>,
	pgsql-hackers@lists.postgresql.org
Subject: Re: Inefficiency in parallel pg_restore with many tables
Message-ID: <20230903151116.GA3460412@nathanxps13>
References: <20230902185521.GA3414119@nathanxps13>
 <20230903100400.nbmajh4ppz5tkzff@alvherre.pgsql>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <20230903100400.nbmajh4ppz5tkzff@alvherre.pgsql>
Archived-At: 
 <https://www.postgresql.org/message-id/20230903151116.GA3460412%40nathanxps13>
Precedence: bulk

On Sun, Sep 03, 2023 at 12:04:00PM +0200, Alvaro Herrera wrote:
> On 2023-Sep-02, Nathan Bossart wrote:
>> I ended up hacking together a (nowhere near committable) patch to see how
>> hard it would be to allow using any type with binaryheap.  It doesn't seem
>> too bad.
> 
> Yeah, using void * seems to lead to interfaces that are pretty much the
> same as bsearch() or qsort().

Right.  This is what I had in mind.

> (Why isn't your payload type const,
> though?)

It probably should be const.  This patch was just a proof-of-concept and
still requireѕ a bit of work.

> I do wonder why did you change _remove_first and _first to have a
> 'result' output argument instead of a return value.  Does this change
> actually buy you anything?  simplehash.h doesn't do that either.
> 
>> -extern void binaryheap_add(binaryheap *heap, Datum d);
>> -extern Datum binaryheap_first(binaryheap *heap);
>> -extern Datum binaryheap_remove_first(binaryheap *heap);
>> -extern void binaryheap_replace_first(binaryheap *heap, Datum d);
>> +extern void binaryheap_add(binaryheap *heap, void *d);
>> +extern void binaryheap_first(binaryheap *heap, void *result);
>> +extern void binaryheap_remove_first(binaryheap *heap, void *result);
>> +extern void binaryheap_replace_first(binaryheap *heap, void *d);

_first could likely just return a pointer to the data in the binary heap's
array.  However, _remove_first has to copy the data somewhere, so I think
the alternative would be to return a palloc'd value.  Is there another way
that I'm not thinking of?

-- 
Nathan Bossart
Amazon Web Services: https://aws.amazon.com