Re: Out-of-date comments about RecentGlobalXmin?

On Tue, 06 Sep 2022 at 16:17, Daniel Gustafsson <daniel(at)yesql(dot)se> wrote:
>> On 6 Sep 2022, at 10:10, Zhang Mingli <zmlpostgres(at)gmail(dot)com> wrote:
>> On Sep 6, 2022, 16:03 +0800, Japin Li <japinli(at)hotmail(dot)com>, wrote:
>
>> It's out-of-date, doesn't it? I'm not sure s/RecentGlobalXmin/RecentXmin/g
>> is right. Any thoughts?
>> I’m afraid not, RecentGlobalXmin is split to several GlobalVis* variables.
>> Need to check one by one.
>
> It's a set of functions actually and not variables IIRC.
>
> It's worth looking at the entire comment and not just the grep output though,
> as these three places share the exact same comment. Note the second paragraph:
>
> /*
> * Start a new transaction here before first access to db, and get a
> * snapshot. We don't have a use for the snapshot itself, but we're
> * interested in the secondary effect that it sets RecentGlobalXmin. (This
> * is critical for anything that reads heap pages, because HOT may decide
> * to prune them even if the process doesn't attempt to modify any
> * tuples.)
> *
> * FIXME: This comment is inaccurate / the code buggy. A snapshot that is
> * not pushed/active does not reliably prevent HOT pruning (->xmin could
> * e.g. be cleared when cache invalidations are processed).
> */
>
> This was added in dc7420c2c92 which removed RecentGlobalXmin, addressing that
> FIXME would of course be very welcome.

My bad! Thanks for pointing out this.

--
Regrads,
Japin Li.
ChengDu WenWu Information Technology Co.,Ltd.