On Wed, Jan 18, 2023 at 03:35:41PM -0500, Tom Lane wrote:
> Andres Freund <andres@anarazel.de> writes:
> > On 2023-01-18 15:08:38 -0500, Tom Lane wrote:
> >> Huh, I wonder why the cfbot never complained about that when
> >> it was testing this patch ... maybe it doesn't use win1252?
>
> > This is the same environment that cfbot is also using, so this is odd. And
> > cfbot has been failing on windows since the commit went into master:
>
> > https://cirrus-ci.com/github/postgresql-cfbot/postgresql/
>
> > So I guess something must have changed in the patch since then? I guess it
> > could be that something changed on the VM, but I don't think so...
>
> Hmmm ... the collate.windows.win1252 test did not exist at all before
> 3-Jan (bf03cfd16), while 1b4d280ea went in on 11-Jan. So perhaps the
> most plausible theory is that the cfbot didn't test this patch between
> those dates. It should certainly have noticed if it did run it ...
Looks like it failed the last 7 days
https://cirrus-ci.com/github/postgresql-cfbot/postgresql/commitfest/41/4048
I also saw that it failed in its previous incarnation, before the
revert, but didn't dig into it at that point.