Re: checking return value from unlink in write_relcache_init_file - Mailing list pgsql-hackers

From Zhihong Yu
Subject Re: checking return value from unlink in write_relcache_init_file
Date
Msg-id CALNJ-vQWn=dmBiB+YPCkPyZiC6piw3rSrMBFyvt5q0oghHt2mA@mail.gmail.com
Whole thread Raw
In response to Re: checking return value from unlink in write_relcache_init_file  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers


On Thu, Jun 3, 2021 at 6:16 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
> On 2021-Jun-03, Tom Lane wrote:
>> If the unlink fails, there's only really a problem if the subsequent
>> open() fails to overwrite the file --- and that stanza is perfectly
>> capable of complaining for itself.  So I think the code is fine and
>> there's no need for a separate message about the unlink.  Refusing to
>> proceed, as you've done here, is strictly worse than what we have.

> It does seem to deserve a comment explaining this.

Agreed, the existing comment there is a tad terse.

                        regards, tom lane
Hi,
Here is the patch with a bit more comment on the unlink() call.

Cheers 
Attachment

pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: checking return value from unlink in write_relcache_init_file
Next
From: "tanghy.fnst@fujitsu.com"
Date:
Subject: RE: [BUG]Update Toast data failure in logical replication