Re: [PATCH] Atomic pgrename on Windows

Hi!

On Tue, Jan 7, 2020 at 7:16 PM Tomas Vondra
<tomas(dot)vondra(at)2ndquadrant(dot)com> wrote:
> I don't have access to a Windows machine and my developer experience
> with that platform is pretty much nil, but I think this patch makes
> sense. It's not an ideal solution, but it's not clear such solution
> exists, and an improvement is better than nothing.

Thank you for your attention to this patch!

> I have two minor comments about rename_temp:
>
> 1) The name might seem to be hinting it's about files opened using
> OpenTemporaryFile, but in practice it's about files that are not
> critical. But maybe it's true.

We may invent another name. What about rename_fragile()?

> 2) I think the rename_temp comment should mention it can only be used in
> cases when the renames happen in a single process (non-concurrently). If
> we could call rename_temp() concurrently from two different processes,
> it won't work as expected. Of course, we only call rename_temp from stat
> collector and syslogger, where it obviously works.

Good point, this should be reflected in comments.

> Anyway, I'm really nitpicking here ...
>
> Are there any objections to get the current patch committed as is, so
> that it does not get pushed yet again to the next commitfest.

It would be good to commit. Let's get agreement on function name first.

------
Alexander Korotkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company