At Fri, 15 May 2020 19:24:28 -0400, Alvaro Herrera <alvherre@2ndquadrant.com> wrote in
> On 2020-May-15, Michael Paquier wrote:
>
> > On Thu, May 14, 2020 at 02:12:25PM +0900, Kyotaro Horiguchi wrote:
> > > Good catch! That's not only for CreateDecodingContet. That happens
> > > everywhere in the query loop in PostgresMain() until logreader is
> > > initialized. So that also happens, for example, by starting logical
> > > replication using invalidated slot. Checking xlogreader != NULL in
> > > WalSndErrorCleanup is sufficient. It doesn't make actual difference,
> > > but the attached explicitly initialize the pointer with NULL.
> >
> > Alvaro, are you planning to look at that? Should we have an open item
> > for this matter?
>
> On it now. I'm trying to add a test for this (needs a small change to
> PostgresNode->psql), but I'm probably doing something stupid in the Perl
> side, because it doesn't detect things as well as I'd like. Still
> trying, but I may be asked to evict the office soon ...
FWIW, and I'm not sure which of the mail and the commit 1d3743023e was
earlier, but I confirmed that the committed test in
006_logical_decoding.pl causes a crash, and the crash is fixed by the
change of code.
regards.
--
Kyotaro Horiguchi
NTT Open Source Software Center