Subject: Re: Dubious code in pg_rewind's process_target_file()
To: Tom Lane <tgl@sss.pgh.pa.us>, pgsql-hackers@lists.postgresql.org
References: <1221796.1599329320@sss.pgh.pa.us>
 <1222282.1599329909@sss.pgh.pa.us>
From: Heikki Linnakangas <hlinnaka@iki.fi>
Message-ID: <6ba44dc9-2b7e-1988-de4e-99832235eae5@iki.fi>
Date: Sun, 6 Sep 2020 17:54:22 +0300
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:68.0) Gecko/20100101
 Thunderbird/68.11.0
MIME-Version: 1.0
In-Reply-To: <1222282.1599329909@sss.pgh.pa.us>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 05/09/2020 21:18, Tom Lane wrote:
> I wrote:
>> It looks to me like we could replace "exists = false" with "return",
>> rather than uselessly constructing a FILE_ACTION_REMOVE entry for
>> a file we've already proven is not there.
> 
> Or actually, maybe we should just drop the lstat call altogether?
> AFAICS it's 99.99% redundant with the lstat that traverse_datadir
> has done nanoseconds before.  Yeah, maybe somebody managed to drop
> the file in between, but the FILE_ACTION_REMOVE code would have to
> deal with such cases anyway in case a drop occurs later.

Agreed, the lstat() doesn't do anything interesting.

This is refactored away by the patches discussed at 
https://www.postgresql.org/message-id/f155aab5-1323-8d0c-9e3b-32703124bf00%40iki.fi. 
But maybe we should still clean it up in the back-branches.

- Heikki