On Tue, 2 Dec 2025 at 03:41, Andres Freund wrote:
>
> Hi,
Hi!
> I.e. we modify the buffer without even holding a share lock on the page. That
seems ... not ok.
I was recently confused by these lines too doing some related hacking
on my work.
>
> Now, this normally won't happen, was the FSM isn
Hi,
In the context of [1] I started looking at the freespace code, in particular
fsm_vacuum_page(), which has this comment:
/*
* Reset the next slot pointer. This encourages the use of low-numbered
* pages, increasing the chances that a later vacuum can truncate the
