On Thu, Mar 16, 2023 at 2:00 AM Alexander Lakhin <exclusion@gmail.com> wrote:
> 15.03.2023 11:43, Thomas Munro wrote:
> > Do you know how it fails in non-assert builds, without the fix?
> So at least with my test script that doesn't lead to a crash or something.
Thanks. We were wondering if the retry mechanism might somehow be
hiding this in non-assert builds, but, looking more closely, that is
tied specifically to the memory reservation operation.
I noticed that d41a178b missed a comment explaining why we used
malloc() instead of palloc(), but that isn't true anymore, so here's a
small patch to clean that up.