Обсуждение: v13 latest snapshot build error

Hi,?

I'm getting build error while building latest snapshot. Any idea why? Please
note that I'm adding this patch to the  tarball:


https://git.postgresql.org/gitweb/?p=pgrpms.git;a=blob;f=rpm/redhat/master/postgresql-13/master/postgresql-13-var-run-socket.patch;h=a0292a80ae219b4c8dc1c2e686a3521f02b4330d;hb=HEAD

+ ./configure --enable-rpath --prefix=/usr/pgsql-13 --includedir=/usr/pgsql-
13/include --libdir=/usr/pgsql-13/lib --mandir=/usr/pgsql-13/share/man --
datadir=/usr/pgsql-13/share --with-icu --with-llvm --with-perl --with-python --
with-tcl --with-tclconfig=/usr/lib64 --with-openssl --with-pam --with-gssapi --
with-includes=/usr/include --with-libraries=/usr/lib64 --enable-nls --enable-
dtrace --with-uuid=e2fs --with-libxml --with-libxslt --with-ldap --with-selinux
--with-systemd --with-system-tzdata=/usr/share/zoneinfo --
sysconfdir=/etc/sysconfig/pgsql --docdir=/usr/pgsql-13/doc --
htmldir=/usr/pgsql-13/doc/html
+ MAKELEVEL=0
+ /usr/bin/make -j4 all
In file included from ../../../../src/include/c.h:55,
                 from ../../../../src/include/postgres.h:46,
                 from guc.c:17:
../../../../src/include/pg_config_manual.h:200:31: error: called object is not
a function or function pointer
 #define DEFAULT_PGSOCKET_DIR  "/var/run/postgresql"
                               ^~~~~~~~~~~~~~~~~~~~~
guc.c:4064:3: note: in expansion of macro 'DEFAULT_PGSOCKET_DIR'
   DEFAULT_PGSOCKET_DIR ", /tmp"
   ^~~~~~~~~~~~~~~~~~~~
make[4]: *** [<builtin>: guc.o] Error 1
make[3]: *** [../../../src/backend/common.mk:39: misc-recursive] Error 2
make[3]: *** Waiting for unfinished jobs....
make[2]: *** [common.mk:39: utils-recursive] Error 2
make[1]: *** [Makefile:42: all-backend-recurse] Error 2
make: *** [GNUmakefile:11: all-src-recurse] Error 2


Regards,
--
Devrim Gündüz
Open Source Solution Architect, Red Hat Certified Engineer
Twitter: @DevrimGunduz , @DevrimGunduzTR

Devrim =?ISO-8859-1?Q?G=FCnd=FCz?= <devrim@gunduz.org> writes:
> I'm getting build error while building latest snapshot. Any idea why? Please
> note that I'm adding this patch to the  tarball:
>
https://git.postgresql.org/gitweb/?p=pgrpms.git;a=blob;f=rpm/redhat/master/postgresql-13/master/postgresql-13-var-run-socket.patch;h=a0292a80ae219b4c8dc1c2e686a3521f02b4330d;hb=HEAD

(Hey, I recognize that patch ...)

> In file included from ../../../../src/include/c.h:55,
>                  from ../../../../src/include/postgres.h:46,
>                  from guc.c:17:
> ../../../../src/include/pg_config_manual.h:200:31: error: called object is not
> a function or function pointer
>  #define DEFAULT_PGSOCKET_DIR  "/var/run/postgresql"
>                                ^~~~~~~~~~~~~~~~~~~~~
> guc.c:4064:3: note: in expansion of macro 'DEFAULT_PGSOCKET_DIR'
>    DEFAULT_PGSOCKET_DIR ", /tmp"
>    ^~~~~~~~~~~~~~~~~~~~
> make[4]: *** [<builtin>: guc.o] Error 1

That is just weird.  Could it be a compiler bug?  I assume you're
using some bleeding-edge gcc version, and it's really hard to see
another reason why this would fail, especially with a nonsensical
error like that.

            regards, tom lane

Hi Tom,

[Adding Christoph]

On Wed, 2020-03-11 at 21:07 +0000, Devrim Gunduz wrote:
> This happens on RHEL 8. I don't think it's that bleeding edge.

Morever, my RHEL 7 build are broken as well:

+ MAKELEVEL=0
+ /usr/bin/make -j4 all
In file included from /usr/include/time.h:37:0,
                 from ../../../../src/include/portability/instr_time.h:64,
                 from ../../../../src/include/executor/instrument.h:16,
                 from ../../../../src/include/nodes/execnodes.h:18,
                 from ../../../../src/include/executor/execdesc.h:18,
                 from ../../../../src/include/executor/executor.h:17,
                 from ../../../../src/include/commands/explain.h:16,
                 from ../../../../src/include/commands/prepare.h:16,
                 from guc.c:40:
guc.c:4068:3: error: called object is not a function or function pointer
   NULL, NULL, NULL
   ^
make[4]: *** [guc.o] Error 1
make[3]: *** [misc-recursive] Error 2
make[3]: *** Waiting for unfinished jobs....
make[2]: *** [utils-recursive] Error 2
make[1]: *** [all-backend-recurse] Error 2
make: *** [all-src-recurse] Error 2
error: Bad exit status from /var/tmp/rpm-tmp.PXbYKE (%build)
    Bad exit status from /var/tmp/rpm-tmp.PXbYKE (%build)

IDK what is going on, but apparently something is broken somewhere.

Cheers,
--
Devrim Gündüz
Open Source Solution Architect, Red Hat Certified Engineer
Twitter: @DevrimGunduz , @DevrimGunduzTR

Devrim =?ISO-8859-1?Q?G=FCnd=FCz?= <devrim@gunduz.org> writes:
> On Wed, 2020-03-11 at 21:07 +0000, Devrim Gunduz wrote:
>> This happens on RHEL 8. I don't think it's that bleeding edge.

> Morever, my RHEL 7 build are broken as well:

Hm.  We have various RHEL and CentOS 7.x machines in the buildfarm,
and they aren't unhappy.  So seems like it must be something specific
to your build.  Are you carrying any other patches?

            regards, tom lane

Hello,

On 3/12/2020 4:44 AM, Tom Lane wrote:
> Devrim =?ISO-8859-1?Q?G=FCnd=FCz?= <devrim@gunduz.org> writes:
>> I'm getting build error while building latest snapshot. Any idea why? Please
>> note that I'm adding this patch to the  tarball:
>>
https://git.postgresql.org/gitweb/?p=pgrpms.git;a=blob;f=rpm/redhat/master/postgresql-13/master/postgresql-13-var-run-socket.patch;h=a0292a80ae219b4c8dc1c2e686a3521f02b4330d;hb=HEAD
> 
> (Hey, I recognize that patch ...)
> 
>> In file included from ../../../../src/include/c.h:55,
>>                   from ../../../../src/include/postgres.h:46,
>>                   from guc.c:17:
>> ../../../../src/include/pg_config_manual.h:200:31: error: called object is not
>> a function or function pointer
>>   #define DEFAULT_PGSOCKET_DIR  "/var/run/postgresql"
>>                                 ^~~~~~~~~~~~~~~~~~~~~
>> guc.c:4064:3: note: in expansion of macro 'DEFAULT_PGSOCKET_DIR'
>>     DEFAULT_PGSOCKET_DIR ", /tmp"
>>     ^~~~~~~~~~~~~~~~~~~~
>> make[4]: *** [<builtin>: guc.o] Error 1
> 
> That is just weird.  Could it be a compiler bug?  I assume you're
> using some bleeding-edge gcc version, and it's really hard to see
> another reason why this would fail, especially with a nonsensical
> error like that.

I'm not familiar with the patch itself. But I think there is just a lack 
of the comma here, after ", /tmp" :-)

> --- a/src/backend/utils/misc/guc.c
> +++ b/src/backend/utils/misc/guc.c
> @@ -4061,7 +4061,7 @@ static struct config_string ConfigureNamesString[] =
>                 },
>                 &Unix_socket_directories,
>  #ifdef HAVE_UNIX_SOCKETS
> -               DEFAULT_PGSOCKET_DIR,
> +               DEFAULT_PGSOCKET_DIR ", /tmp"
>  #else
>                 "",
>  #endif

So, it should be:
  > #ifdef HAVE_UNIX_SOCKETS
> -               DEFAULT_PGSOCKET_DIR,
> +               DEFAULT_PGSOCKET_DIR ", /tmp",
>  #else

-- 
Artur

Artur Zakirov <zaartur@gmail.com> writes:
> I'm not familiar with the patch itself. But I think there is just a lack 
> of the comma here, after ", /tmp" :-)

[ blink... ]  There definitely is a comma there in the version of the
patch that's in the Fedora repo.

            regards, tom lane

On 3/12/2020 11:52 AM, Tom Lane wrote:
> Artur Zakirov <zaartur@gmail.com> writes:
>> I'm not familiar with the patch itself. But I think there is just a lack
>> of the comma here, after ", /tmp" :-)
> 
> [ blink... ]  There definitely is a comma there in the version of the
> patch that's in the Fedora repo.

Ah, I see. I just saw the version Devrim sent (not sure that it is used 
in the Fedora repo):


https://git.postgresql.org/gitweb/?p=pgrpms.git;a=blob;f=rpm/redhat/master/postgresql-13/master/postgresql-13-var-run-socket.patch;h=a0292a80ae219b4c8dc1c2e686a3521f02b4330d;hb=HEAD

And I thought there should be a comma at the end of the line after 
concatenation to avoid concatenation with NULL:

DEFAULT_PGSOCKET_DIR ", /tmp"

>   {
>     {"unix_socket_directories", PGC_POSTMASTER, CONN_AUTH_SETTINGS,
>       gettext_noop("Sets the directories where Unix-domain sockets will be created."),
>       NULL,
>       GUC_SUPERUSER_ONLY
>     },
>     &Unix_socket_directories,
> #ifdef HAVE_UNIX_SOCKETS
>     DEFAULT_PGSOCKET_DIR ", /tmp"
> #else
>     "",
> #endif
>     NULL, NULL, NULL
>   },



-- 
Artur

On Wed, 2020-03-11 at 22:52 -0400, Tom Lane wrote:
> Artur Zakirov <zaartur@gmail.com> writes:
> > I'm not familiar with the patch itself. But I think there is just a lack
> > of the comma here, after ", /tmp" :-)
>
> [ blink... ]  There definitely is a comma there in the version of the
> patch that's in the Fedora repo.


Exactly.Looks like I broke the patch while trying to update the patch.

Sorry for the noise.

Regards,

--
Devrim Gündüz
Open Source Solution Architect, Red Hat Certified Engineer
Twitter: @DevrimGunduz , @DevrimGunduzTR