Обсуждение: Rare link canary failure in dblink test

Hi,

A couple of recent cases where an error "libpq is incorrectly linked
to backend functions" broke the dblink test:

 https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=lorikeet&dt=2020-08-22%2002:55:22
  | REL9_6_STABLE | Windows
 https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=lorikeet&dt=2020-04-27%2022:46:16
  | HEAD          | Windows

Before that, lorikeet last said that before commit a33245a8 apparently
fixed something relevant, but curiously the message also appeared a
couple of times on two Unixen.  Perhaps we can write those off as lost
in the mists of time.

 https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=lorikeet&dt=2018-09-29%2010:56:25
  | HEAD          | Windows
 https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=damselfly&dt=2018-09-10%2011:09:29
 | HEAD          | Illumos
 https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=jaguarundi&dt=2018-09-09%2018:11:15
| HEAD          | FreeBSD

Thomas Munro <thomas.munro@gmail.com> writes:
> A couple of recent cases where an error "libpq is incorrectly linked
> to backend functions" broke the dblink test:

lorikeet seems just plain unstable these days :-(.  Don't know why.

> ... curiously the message also appeared a
> couple of times on two Unixen.  Perhaps we can write those off as lost
> in the mists of time.

No, those were expected, because they ran between ed0cdf0e0, which
intentionally introduced this error, and e3d77ea6b and 4fa3741d1
which fixed it.

            regards, tom lane