Re: Regression test PANICs with master-standby setup on samemachine

At Tue, 7 May 2019 10:55:06 +0900, Michael Paquier <michael@paquier.xyz> wrote in <20190507015506.GC1499@paquier.xyz>
> On Tue, May 07, 2019 at 10:16:54AM +0900, Kyotaro HORIGUCHI wrote:
> > The fake symlinks need correction after the data directory and
> > tablespsce directory are moved. Maybe needs to call
> > CorrectSymlink() or something at startup... Or relative
> > tablespaces should be rejected on Windows?
> 
> It took enough sweat and tears to have an implementation with junction
> points done correctly on Windows and we know that it works, so I am

Indeed. It is very ill documented and complex.

> not sure that we need an actual wrapper for readlink() and such for
> the backend code to replace junction points.  The issue with Windows
> is that perl's symlink() is not directly available on Windows.

Ugg. If we want to run tablespace-related tests involving
replication on Windows, we need to make the tests using absolute
tablespace paths. Period...?

regards.

-- 
Kyotaro Horiguchi
NTT Open Source Software Center