pgsql: Fix incorrect format placeholder.

Started by Tom Laneover 4 years ago8 messagescomitters
Jump to latest
#1Tom Lane
tgl@sss.pgh.pa.us

Fix incorrect format placeholder.

Per buildfarm warnings.

Branch
------
master

Details
-------
https://git.postgresql.org/pg/commitdiff/b0cf5444f9a8d915b2e9b44790025f17a7dc107f

Modified Files
--------------
src/bin/pg_basebackup/pg_receivewal.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

#2Michael Paquier
michael@paquier.xyz
In reply to: Tom Lane (#1)
Re: pgsql: Fix incorrect format placeholder.

On Mon, Nov 08, 2021 at 07:32:41PM +0000, Tom Lane wrote:

Fix incorrect format placeholder.

Per buildfarm warnings.

Thanks, that's embarassing. The last time I bumped into this issue,
the report came from lapwing which would directly report a failure,
but it did not happen this time. Which animal has reported that?
--
Michael

#3Tom Lane
tgl@sss.pgh.pa.us
In reply to: Michael Paquier (#2)
Re: pgsql: Fix incorrect format placeholder.

Michael Paquier <michael@paquier.xyz> writes:

Thanks, that's embarassing. The last time I bumped into this issue,
the report came from lapwing which would directly report a failure,
but it did not happen this time. Which animal has reported that?

Umm ... florican and dangomushi spotted it. Maybe we need a 32-bit
machine with -Werror?

regards, tom lane

#4Michael Paquier
michael@paquier.xyz
In reply to: Tom Lane (#3)
Re: pgsql: Fix incorrect format placeholder.

On Tue, Nov 09, 2021 at 12:19:36AM -0500, Tom Lane wrote:

Umm ... florican and dangomushi spotted it. Maybe we need a 32-bit
machine with -Werror?

Indeed, I can see it in the logs of dangomushi. I could add a -Werror
there, but there are also a bunch of warnings generated by plperl.
I'd rather keep --with-perl and clang, and there seems to be a fix
regarding that stuff in upstream:
https://www.nntp.perl.org/group/perl.perl5.changes/2021/07/msg57758.html

But it also looks like Archlinux ARM did not catch up that yet..
--
Michael

#5Tom Lane
tgl@sss.pgh.pa.us
In reply to: Michael Paquier (#4)
Re: pgsql: Fix incorrect format placeholder.

Michael Paquier <michael@paquier.xyz> writes:

On Tue, Nov 09, 2021 at 12:19:36AM -0500, Tom Lane wrote:

Umm ... florican and dangomushi spotted it. Maybe we need a 32-bit
machine with -Werror?

Indeed, I can see it in the logs of dangomushi. I could add a -Werror
there, but there are also a bunch of warnings generated by plperl.

Yeah, we had complaints before about -Wcompound-token-split-by-macro
warnings.

I'd rather keep --with-perl and clang, and there seems to be a fix
regarding that stuff in upstream:
https://www.nntp.perl.org/group/perl.perl5.changes/2021/07/msg57758.html

I figured the Perl guys would do something about it, but it will be a
very long time before that propagates everywhere. Maybe we should
teach configure to add -Wno-compound-token-split-by-macro? By my
count there are now five buildfarm members besides dangomushi that
are producing these warnings, and I think the number is going to
go up before it goes down.

(Meanwhile, I guess I could enable -Werror on florican.)

regards, tom lane

#6Michael Paquier
michael@paquier.xyz
In reply to: Tom Lane (#5)
Re: pgsql: Fix incorrect format placeholder.

On Tue, Nov 09, 2021 at 01:06:26AM -0500, Tom Lane wrote:

I figured the Perl guys would do something about it, but it will be a
very long time before that propagates everywhere. Maybe we should
teach configure to add -Wno-compound-token-split-by-macro? By my
count there are now five buildfarm members besides dangomushi that
are producing these warnings, and I think the number is going to
go up before it goes down.

(Meanwhile, I guess I could enable -Werror on florican.)

Yeah, you have a point here. -Wcompound-token-split-by-macro is new
as of clang-12 from last April, so I am a bit surprised that it is
that present in the buildfarm TBH.

The attached is sufficient to remove those warnings. Debian provides
clang-12, so it is easy to reproduce the problem.

What do you think?
--
Michael

Attachments:

plperl-clang-warnings.patchtext/x-diff; charset=us-asciiDownload+54-0
#7Tom Lane
tgl@sss.pgh.pa.us
In reply to: Michael Paquier (#6)
Re: pgsql: Fix incorrect format placeholder.

Michael Paquier <michael@paquier.xyz> writes:

Yeah, you have a point here. -Wcompound-token-split-by-macro is new
as of clang-12 from last April, so I am a bit surprised that it is
that present in the buildfarm TBH.

The attached is sufficient to remove those warnings. Debian provides
clang-12, so it is easy to reproduce the problem.

OK by me, but probably should raise the issue in -hackers not just
-committers. Also, please back-patch once the release freeze lifts.

regards, tom lane

#8Michael Paquier
michael@paquier.xyz
In reply to: Tom Lane (#7)
Re: pgsql: Fix incorrect format placeholder.

On Tue, Nov 09, 2021 at 09:40:47AM -0500, Tom Lane wrote:

OK by me, but probably should raise the issue in -hackers not just
-committers. Also, please back-patch once the release freeze lifts.

Sure.
--
Michael