On Thu, 8 Jan 2004, Tom Lane wrote:
> Holger Marzen <[EMAIL PROTECTED]> writes:
> > When I try to compile PostgreSQL 7.2.4 or 7.4.1 then I get
>
> > |gcc -O2 -Wall -Wmissing-prototypes -Wmissing-declarations
> > |-I../../../src/include -D_GNU_SOURCE -c hba.c -o hba.o
> > |hba.c: In function `parse_hba':
> > |hba.c:675: warning: implicit declaration of function `gai_strerror'
> > |hba.c:675: warning: format argument is not a pointer (arg 3)
> > |hba.c: In function `ident_unix':
> > |hba.c:1411: storage size of `peercred' isn't known
> > |hba.c:1411: warning: unused variable `peercred'
> > |make[3]: *** [hba.o] Error 1
>
> The configure process assumes that any platform that supplies
> getaddrinfo() will also supply gai_strerror(). Is that not true
> in your version of glibc? If not, one possible workaround is to
> force use of our own getaddrinfo replacement (#undef HAVE_GETADDRINFO
> and add getaddrinfo.o to LIBOBJS).
I changed the LIBOBJS in src/Makefile.global and put the #undef after
the #define in src/include/pg_config.h.
But unfortunately that's only half of the world domaination. I still
got:
hba.c: In function `ident_unix':
hba.c:1411: storage size of `peercred' isn't known
hba.c:1411: warning: unused variable `peercred'
make[3]: *** [hba.o] Error 1
> Probably the easiest way around the peercred problem is just to disable
> that chunk of code ("#elif defined(SO_PEERCRED)" to "#elif 0" at line
Yep. That's the other half!
> 1409 of hba.c will probably do it). You may need to dike out the
> corresponding code in client-side libpq as well.
Not neccessary. It compiled and linked.
Thanks, Tom.
--
PGP/GPG Key-ID:
http://blackhole.pca.dfn.de:11371/pks/lookup?op=get&search=0xB5A1AFE1
---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster
