On Fri, Jun 19, 2020 at 7:48 PM Michael Paquier <mich...@paquier.xyz> wrote: > We cared about such plan stability that in the past FWIW, see for > example c588df9 as work_mem is a setting that people like to change. > Why should this be different? work_mem is a popular configuration > setting.
The RMT met today. We determined that it wasn't worth adjusting this test to pass with non-standard work_mem values. "make installcheck" also fails with lower random_page_cost settings. There doesn't seem to be any reason to permit a non-standard setting to cause installcheck to fail elsewhere, while not tolerating the same issue here, with work_mem. Thanks -- Peter Geoghegan
