On Fri, Dec 17, 2021 at 01:41:00PM -0500, Tom Lane wrote:
> Noah Misch <noah@leadboat.com> writes:
> > On Fri, Dec 17, 2021 at 12:52:39PM -0500, Tom Lane wrote:
> >> It's sort of annoying to fire up a psql+backend
> >> for just one command, but perhaps there's other stuff that could be
> >> put there too.
>
> > Yes. The src/test/regress suite would be in a better place if one could run
> > most test files via a schedule containing only two files, the setup file and
> > the file of interest. Adding things like the "CREATE TABLE tenk1" to the
> > setup file would help that.
>
> If we're thinking of a generalized setup file, putting it after the
> tablespace test feels pretty weird. What was your motivation for
> doing this at the end of tablespace.source rather than the start?
I did it that way so a bit of the "make check" suite would exercise the
standard user experience. That's a minor concern, so putting the setup file
before the tablespace file is fine. Various contrib and TAP suites will still
test the standard user experience.
> It doesn't look like that test in itself had any interesting
> dependencies on public not being writable.
Right.