On 27/02/2026 16:21, Maxim Orlov wrote:
> I just wanted to share my work-in-progress on switching the clog to 64
> bits. So far, I've only dealt with the clog itself. Right now, support
> for the x86 platform is lacking. But the first two patches are
> refactorings and can be committed promptly.
>
> What are your thoughts on this?
This is not useful, until we have an epoch on heap pages, and can deal
with 64-bit XIDs in many other places too. But it *would* be useful to
have an epoch on heap pages without this patch. It eliminates the need
to dirty pages when freezing, when there are no dead tuples. So I'd
suggest starting with epochs on heap pages.
- Heikki
