Re: pgsql: pg_basebackup: Fix a couple of recently-introduced bugs.

On Thu, Jan 27, 2022 at 2:22 PM Justin Pryzby <pryzby(at)telsasoft(dot)com> wrote:
> On Thu, Jan 27, 2022 at 02:19:30PM -0500, Robert Haas wrote:
> > On Thu, Jan 27, 2022 at 12:45 PM Justin Pryzby <pryzby(at)telsasoft(dot)com> wrote:
> > > If I'm not wrong, this is causing failures on cfbot.
> > > https://cirrus-ci.com/task/6312181619752960?logs=build#L2056
> > >
> > > [17:24:48.348] c:\cirrus\src\backend\replication\basebackup_gzip.c(90): warning C4715: 'bbsink_gzip_new': not all control paths return a value [c:\cirrus\postgres.vcxproj]
> > > [17:24:48.348] c:\cirrus\src\bin\pg_basebackup\pg_basebackup.c(1922,67): error C2065: 'Z_DEFAULT_COMPRESSION': undeclared identifier [c:\cirrus\pg_basebackup.vcxproj]
> > >
> > > The warning seems to be what's fixed by David's patch here.
> > > https://www.postgresql.org/message-id/CA%2BTgmobSeRFM84pwa5HfZCGRx%3DQ-iFfPFZfq-4%3Dt1p04yxg_Og%40mail.gmail.com
> >
> > I can commit that patch (though so could David) but I don't understand
> > why you think it's connected to this commit particularly.
>
> The C2065: is connected to this commit, right ?
>
> I included the warning since it's 1 line away and evidences that the warning is
> easily visible , and not due to some obscure combination of windows settings.

Oh ... I see, I misread your email. Yeah, this commit introduced
Z_DEFAULT_COMPRESSION outside of a HAVE_LIBZ block. Whoops.

--
Robert Haas
EDB: http://www.enterprisedb.com