On Sat, Mar 21, 2020 at 03:04:51PM -0400, Tom Lane wrote:
> Noah Misch <noah@leadboat.com> writes:
> > Skip WAL for new relfilenodes, under wal_level=minimal.
>
> Various buildfarm members seem not happy with this,
> though it's not universal.
Yeah, see the main thread:
https://postgr.es/m/20200321190127.GA1763544%40rfd.leadboat.com
> Also, given the invasiveness of the patch, I'm quite astonished
> that you chose to back-patch it. Is this really worth ABI
> breakage risks in the back branches?
A month ago (https://postgr.es/m/20200219074452.GA4006615@rfd.leadboat.com), I
invited folks to talk me out of back-patching it. Nobody commented either
way, so I back-patched. Please reply to that message if you see a reason to
revert the back-patches. I found only one pgxn extension ("citus") that I
expect to notice the ABI break (change in structure size), and
089e4d405d0f3b94c74a2c6a54357a84a681754b changed the size of the same
structure. Hence, I'm not worrying about the ABI aspect. I'm more worried
that some defect that evades testing will corrupt more data than the original
bug. I don't have any specific kind of defect in mind, though.