> On 19 Aug 2026, at 10:32, Fujii Masao <[email protected]> wrote: > > On Wed, Aug 19, 2026 at 5:04 PM Daniel Gustafsson <[email protected]> wrote: >> I don't disagree with this, but I also don't think between beta3 and rc1 is a >> good time to remove fields (which have been there for a very long time) from >> system info functions. Revisiting this during v20 and cleaning it up seems >> like a very good idea though. > > Yeah, I understand your point. On the other hand, I'm still not convinced > that adding data_checksum_version_init is a good idea either. Another > option would be to drop data_checksum_version_init and revert the change > to pg_control_init().data_page_checksum_version, so that it continues > to report the current data_checksum_version, as it does in v18. > > This behavior does conflict with the docs for pg_control_init(), as you > pointed out: "pg_control_init is documented to return the state at cluster > initialization". But, this inconsistency already existed in v18. So that > seems acceptable for most users. It might also be less surprising > because the meaning of the value would remain unchanged between v18 > and v19.
That's another option. I'm not a big fan of adapting new code to bugs of the past, but at this point in the cycle there are other aspects in play as well. I'll see what others think for a little bit before reverting. -- Daniel Gustafsson
