I'm trying to compile an exim with GDBM on a system that has db1, db2,
db3, and gdbm installed (RH72).. (I need gdbm for compatibility with
something else that uses it)
Exim appears to ALWAYS be following the "#elif defined USE_DB" , because
I keep getting "undefined reference to `db_create'", and db_create is
ONLY referenced from within this #elif section.
I have tried:
USE_DB=no
I have tried:
#undef USE_DB
.. in Local/Makefile
It seems to be set in OS/Makefile-Linux
It doesnt seem to check if its defined positively, only if its defined
at all.
I don't seem to have a way to unset it without editing Makefile-Linux
itself.
Perhaps the test should have been more like
#elif USE_DB=yes
to allow for a setting of USE_DB=no (or even USE_DB=blah) preventing it
from obeying that section
On a related tangent - perhaps exim should support any/all DBM libraries
that are installed, and have a specification for which type it should
use for its own db files, and have a per-lookup way of specifying which
type to use for that lookup ??
Suppose two different closed-source apps, using two different db
libraries, were writing data files that the same instance of exim
needed to be able to access?
On Fri, 21 Dec 2001, Philip Hazel wrote:
> On Thu, 20 Dec 2001, Gabor Z. Papp wrote:
>
> > I have only LOOKUP_DBM=yes defined. Is that problem? Because
> > I still can't compile the latest exim4.
>
> Oh dear.
>
> > $ grep DB Makefile
> > #LOOKUP_DBM=yes
> > USE_GDBM=YES
> > DBMLIB=-R/usr/lib -L/usr/lib -lgdbm
> > # LOOKUP_CDB=yes
> > # LOOKUP_DNSDB=yes
> > # DB_DIRECTORY_MODE=0750
> > # DB_MODE=0640
> > # DB_LOCKFILE_MODE=0640
> > # The TESTDB lookup is for performing tests on the handling of lookup results,
> > # LOOKUP_TESTDB=yes
>
> That looks fine. I do not understand why it is trying to look at db.h in
> that case. The only mention of db.h in the entire source is like this:
>
> #elif defined USE_DB
>
> #include <db.h>
>
> which is why I asked the question about USE_DB. (This reference is in
> the source file dbstuff.h.) I was so sure this was the answer that I
> didn't keep your original message. If you can't resolve this, please
> send me the compile error messages again (off the list) and I'll see if
> I can think of anything else that might be happening.
>
> --
> Philip Hazel University of Cambridge Computing Service,
> ph10@??? Cambridge, England. Phone: +44 1223 334714.
>
>
>
> --
>
> ## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##
>
>
--
