Re: BUG #2608: dblink compile fails on AIX 5.3

In the last exciting episode, hoffmann(at)fh-trier(dot)de ("Michael Hoffmann") wrote:
> The following bug has been logged online:
>
> Bug reference: 2608
> Logged by: Michael Hoffmann
> Email address: hoffmann(at)fh-trier(dot)de
> PostgreSQL version: 8.1.2
> Operating system: AIX 5.3
> Description: dblink compile fails on AIX 5.3
> Details:
>
>>/usr/local/pgsql/src/postgresql-8.1.2/contrib/dblink>gmake
> ../../src/backend/port/aix/mkldexport.sh libdblink.a > libdblink.exp
> xlC -O2 -qmaxmem=16384 -qsrcmsg -qlonglong -Wl,-bnoentry -Wl,-H512
> -Wl,-bM:SRE -o libdblink.so libdblink.a -Wl,-bE:libdblink.exp
> -L../../src/interfaces/libpq -L../../src/port -lpq
> -Wl,-bI:../../src/backend/postgres.imp
> ld: 0711-317 ERROR: Undefined symbol: .bindtextdomain
> ld: 0711-317 ERROR: Undefined symbol: .dgettext
> ld: 0711-345 Use the -bloadmap or -bnoquiet option to obtain more
> information.
> gmake: *** [libdblink.so] Error 8

I mentioned earlier that changing SHLIB_LINK might be helpful.

The patch that turns out to be needful for crypto is the following:

Index: Makefile
===================================================================
RCS file: /projects/cvsroot/pgsql/contrib/dblink/Makefile,v
retrieving revision 1.11
diff -u -r1.11 Makefile
--- Makefile 27 Feb 2006 12:54:38 -0000 1.11
+++ Makefile 10 Sep 2006 21:01:45 -0000
@@ -20,3 +20,5 @@
include $(top_builddir)/src/Makefile.global
include $(top_srcdir)/contrib/contrib-global.mk
endif
+
+SHLIB_LINK += $(filter -lssl -lcrypto -lssleay32 -leay32, $(LIBS))

In effect, we need to take LIBS, which contains all the libraries
pulled in as "defaults" for builds of objects in the main tree, filter
that against the 4 crypto libraries, and add them to the set of libs
we need to link with (e.g. - SHLIB_LINK).

I'm not sure what libraries gettext draws in; supposing it adds, to a
build of psql, -lgettext, then you might also add the line:

SHLIB_LINK += $(filter -lgettext, $(LIBS))
--
wm(X,Y):-write(X),write('@'),write(Y). wm('cbbrowne','gmail.com').
http://cbbrowne.com/info/nonrdbms.html
"In computing, invariants are ephemeral." -- Alan Perlis