On Fri, May 24, 2019 at 4:18 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Yes, I see that too with sufficiently high -j. I believe this is
> what Noah was trying to fix in bd1592e85, but that patch evidently
> needs a bit more work :-(
It would be nice if this was fixed, but I don't see a problem when I
use the optimum number of jobs, so I don't consider it to be urgent.
I'm happy with the new approach, since it avoids the problem of
regression.diffs files that get deleted before I have a chance to take
a look. I should thank Noah for his work on this.
--
Peter Geoghegan