Re: Atomic rename feature for Windows. - Mailing list pgsql-hackers

From Thomas Munro
Subject Re: Atomic rename feature for Windows.
Date
Msg-id CA+hUKGKca6FoCFP1Y_=JEc5_UnmVZ65tRpfoDOkfSE4dyoYXtA@mail.gmail.com
Whole thread Raw
In response to Re: Atomic rename feature for Windows.  (Victor Spirin <v.spirin@postgrespro.ru>)
Responses Re: Atomic rename feature for Windows.  (Ian Lawrence Barwick <barwick@gmail.com>)
List pgsql-hackers
On Wed, Apr 6, 2022 at 10:40 AM Victor Spirin <v.spirin@postgrespro.ru> wrote:
> Updated patch: we use the posix semantic features in Windows build 17763
> and up.
> We found an issue with this feature on Windows Server 2016 without
> updates (Windows 1607 Build 14393)

Hi Victor,

I rebased and simplified this, and added a lot of tests to be able to
understand what it does.  I think all systems that didn't have this
are now EOL and we don't need to support them in PG16, but perhaps our
_WIN32_WINNT is not quite high enough (this requires Win10 RS1, which
itself was EOL'd in 2019); the main question I have now is what
happens when you run this on non-NTFS filesystems, and whether we want
to *require* this to work because the non-POSIX support will probably
finish up untested.  I posted all that over on a new thread where I am
tidying up lots of related stuff, and I didn't want to repost the
proposed testing framework in multiple threads...

https://www.postgresql.org/message-id/flat/CA%2BhUKG%2BajSQ_8eu2AogTncOnZ5me2D-Cn66iN_-wZnRjLN%2Bicg%40mail.gmail.com



pgsql-hackers by date:

Previous
From: Corey Huinker
Date:
Subject: Re: Getting rid of SQLValueFunction
Next
From: Melih Mutlu
Date:
Subject: Re: Mingw task for Cirrus CI