On Mon, May 17, 2021 at 08:57:55PM -0400, Tom Lane wrote:
> Michael Paquier writes:
>> One thing that's changing in this patch is that testtablespace would
>> be created even if the input directory does not exist when using
>> --make-testtablespace-dir.
>
> Yeah, I do not see a reason for there t
Michael Paquier writes:
> On Mon, May 17, 2021 at 05:51:54PM -0400, Tom Lane wrote:
>> Say, as attached. (Windows part is untested.)
> One thing that's changing in this patch is that testtablespace would
> be created even if the input directory does not exist when using
> --make-testtablespace-d
On Mon, May 17, 2021 at 05:51:54PM -0400, Tom Lane wrote:
> I wrote:
>> I think what we want to do is have this code invoked only in
>> test directories that explicitly ask for it, say with a new
>> "--make-testtablespace" switch for pg_regress.
>
> Say, as attached. (Windows part is untested.)
I wrote:
> I think what we want to do is have this code invoked only in
> test directories that explicitly ask for it, say with a new
> "--make-testtablespace" switch for pg_regress.
Say, as attached. (Windows part is untested.)
regards, tom lane
diff --git a/src/test/re
Michael Paquier writes:
> Move tablespace path re-creation from the makefiles to pg_regress
So this didn't seem like a problem at the time, but while building
beta1 tarballs I discovered that it leaves behind "testtablespace"
subdirectories in various places where they aren't cleaned by
"make dis
Move tablespace path re-creation from the makefiles to pg_regress
Moving this logic into pg_regress fixes a potential failure with
parallel tests when pg_upgrade and the main regression test suite both
trigger the makefile rule that cleaned up testtablespace/ under
src/test/regress. Even if pg_up
