Re: ia64-hp-hpux11.23 configure warnings

checking build system type... ia64-hp-hpux11.23
checking host system type... ia64-hp-hpux11.23
...
checking sys/ipc.h presence... yes
configure: WARNING: sys/ipc.h: present but cannot be compiled
configure: WARNING: sys/ipc.h: check for missing prerequisite headers?
configure: WARNING: sys/ipc.h: see the Autoconf documentation
configure: WARNING: sys/ipc.h: section "Present But Cannot Be Compiled"
configure: WARNING: sys/ipc.h: proceeding with the preprocessor's result
configure: WARNING: sys/ipc.h: in the future, the compiler will take precedence
configure: WARNING: ## ---------------------------------------- ##
configure: WARNING: ## Report this to pgsql-bugs(at)postgresql(dot)org ##
configure: WARNING: ## ---------------------------------------- ##

There are several more of these sorts of warnings in the
attached complete configure output.

Thanks,
Ed

On Thursday September 8 2005 12:20 pm, Ed L. wrote:
> checking build system type... ia64-hp-hpux11.23
> checking host system type... ia64-hp-hpux11.23
> ...
> checking sys/ipc.h presence... yes
> configure: WARNING: sys/ipc.h: present but cannot be compiled
> configure: WARNING: sys/ipc.h: check for missing
> prerequisite headers? configure: WARNING: sys/ipc.h: see the
> Autoconf documentation configure: WARNING: sys/ipc.h:
> section "Present But Cannot Be Compiled" configure: WARNING:
> sys/ipc.h: proceeding with the preprocessor's result
> configure: WARNING: sys/ipc.h: in the future, the compiler
> will take precedence configure: WARNING: ##
> ---------------------------------------- ## configure:
> WARNING: ## Report this to pgsql-bugs(at)postgresql(dot)org ##
> configure: WARNING: ##
> ---------------------------------------- ##
>

Build failure of 8.1beta1 for ia64-hp-hpux11.23:

gcc -O2 -Wall -Wmissing-prototypes -Wpointer-arith -Wendif-labels -fno-strict-aliasing -I../../../src/include -D_XOPEN_SOURCE_EXTENDED -c -o pg_sema.o pg_sema.c
In file included from pg_sema.c:22:
/usr/include/sys/ipc.h:51: error: parse error before "cid_t"
/usr/include/sys/ipc.h:56: error: parse error before '}' token
In file included from pg_sema.c:25:
/usr/include/sys/sem.h:91: error: field `sem_perm' has incomplete type
gmake[3]: *** [pg_sema.o] Error 1
gmake[3]: Leaving directory `/users/dists/pgsql/builds/postgresql-8.1beta1/src/backend/port'
gmake[2]: *** [port-recursive] Error 2
gmake[2]: Leaving directory `/users/dists/pgsql/builds/postgresql-8.1beta1/src/backend'
gmake[1]: *** [all] Error 2
gmake[1]: Leaving directory `/users/dists/pgsql/builds/postgresql-8.1beta1/src'
gmake: *** [all] Error 2
*** Error exit code 2

Ed

Ed L. wrote:
> gcc -O2 -Wall -Wmissing-prototypes -Wpointer-arith -Wendif-labels
> -fno-strict-aliasing -I../../../src/include -D_XOPEN_SOURCE_EXTENDED
> -c -o pg_sema.o pg_sema.c In file included from pg_sema.c:22:
> /usr/include/sys/ipc.h:51: error: parse error before "cid_t"
> /usr/include/sys/ipc.h:56: error: parse error before '}' token

Evidently, you need to include another header before sys/ipc.h, namely
the one that defines the cid_t type. Care to find out which one that
is?

--
Peter Eisentraut
http://developer.postgresql.org/~petere/

"Ed L." <pgsql(at)bluepolka(dot)net> writes:
> In file included from pg_sema.c:22:
> /usr/include/sys/ipc.h:51: error: parse error before "cid_t"
> /usr/include/sys/ipc.h:56: error: parse error before '}' token

So where is cid_t defined?

regards, tom lane

On Thursday September 8 2005 2:18 pm, Tom Lane wrote:
> "Ed L." <pgsql(at)bluepolka(dot)net> writes:
> > In file included from pg_sema.c:22:
> > /usr/include/sys/ipc.h:51: error: parse error before "cid_t"
> > /usr/include/sys/ipc.h:56: error: parse error before '}'
> > token
>
> So where is cid_t defined?

/usr/include/sys/types.h: typedef int32_t cid_t; /* For compartment IDs */

This is 8.1beta1.

On Thursday September 8 2005 3:48 pm, Ed L. wrote:
> On Thursday September 8 2005 2:18 pm, Tom Lane wrote:
> > "Ed L." <pgsql(at)bluepolka(dot)net> writes:
> > > In file included from pg_sema.c:22:
> > > /usr/include/sys/ipc.h:51: error: parse error before
> > > "cid_t" /usr/include/sys/ipc.h:56: error: parse error
> > > before '}' token
> >
> > So where is cid_t defined?
>
> /usr/include/sys/types.h: typedef int32_t cid_t;
> /* For compartment IDs */
>
> This is 8.1beta1.

I see sys/ipc.h already #includes sys/types.h, though.

Ed

"Ed L." <pgsql(at)bluepolka(dot)net> writes:
> I see sys/ipc.h already #includes sys/types.h, though.

Possibly there's some magic #define needed to expose the cid_t
definition?

We have this in src/template/hpux
CPPFLAGS="-D_XOPEN_SOURCE_EXTENDED"
because it seems necessary on HPUX 10. It used to work on HPUX 11 too,
the last time I tried that ...

regards, tom lane

On Thursday September 8 2005 4:08 pm, Tom Lane wrote:
> "Ed L." <pgsql(at)bluepolka(dot)net> writes:
> > I see sys/ipc.h already #includes sys/types.h, though.
>
> Possibly there's some magic #define needed to expose the cid_t
> definition?
>
> We have this in src/template/hpux
> CPPFLAGS="-D_XOPEN_SOURCE_EXTENDED"
> because it seems necessary on HPUX 10. It used to work on
> HPUX 11 too, the last time I tried that ...

Here's my 11.23 compile line:
gcc -O2 -mlp64 -Wall -Wmissing-prototypes -Wpointer-arith
-Wendif-labels -fno-strict-aliasing -I../../../src/include
-D_XOPEN_SOURCE_EXTENDED -c -o pg_sema.o pg_sema.c

I've attached sys/types.h and sys/ipc.h. I'm happy to try any
patches you suggest...

Ed

"Ed L." <pgsql(at)bluepolka(dot)net> writes:
> Here's my 11.23 compile line:
> gcc -O2 -mlp64 -Wall -Wmissing-prototypes -Wpointer-arith
> -Wendif-labels -fno-strict-aliasing -I../../../src/include
> -D_XOPEN_SOURCE_EXTENDED -c -o pg_sema.o pg_sema.c

> I've attached sys/types.h and sys/ipc.h. I'm happy to try any
> patches you suggest...

The problem is obviously impossible given those headers ;-).

What seems most probable at this point is that gcc is not looking at
the headers you think it is. gcc keeps private "fixed" copies of
system headers that it thinks are broken. I suspect that your gcc
installation was copied from an earlier HPUX version that had different
contents of one or both of these headers.

I'm not sure if you need to completely reinstall gcc to fix this,
or if just rerunning its fixincludes script will do the job. Or
you could try manually updating the private copies, but that seems
pretty risky. In any case I think it's a gcc-doesn't-match-system
problem, not our bug.

regards, tom lane