On Thu, Oct 21, 2021 at 3:29 PM Robert Haas <robertmh...@gmail.com> wrote: > I think the correct fix for this particular problem is just to delete > the initialization, as in the attached patch. I spent a long time > studying this today and eventually convinced myself that there's just > no way these initializations can ever do anything (details in proposed > commit message). While it is important that we do not access the > global variable when it's uninitialized, here there is no reason to > access it in the first place.
I have committed this. -- Robert Haas EDB: http://www.enterprisedb.com