Re: OpenBSC compile error --> tests --> GTP hub

[Neels Hofmeyr]

On Sat, Nov 28, 2015 at 01:39:52PM +0100, Sipos Csaba wrote:
> Sorry, it was the missing libc-ares-dev package.
> 
> Its interesting because the configure script is recognising the missing 
> package but not complaining about it, and it causes compile time issues.

I can reproduce the problem. Thanks for the heads up!

I've fixed the build failure: the gtphub_test binary was being built, but
gtphub.o was not (ff252bb80).

However, I've noticed that the gtphub shows as FAILED instead of skipped,
just like the sgsn and oap tests. The cause is:
openbsc/tests/atlocal.in:
    enable_sgsn_test='@found_libgtp@'
    enable_gtphub_test='@found_libgtp@'

It should be something like

    enable_sgsn_test='@found_libgtp@ and @found_libcares@'

I'm not sure how best to achieve that. Maybe like this?

[[[
--- a/openbsc/configure.ac
+++ b/openbsc/configure.ac
@@ -84,6 +84,12 @@ PKG_CHECK_MODULES([LIBCARES], [libcares], [], 
[found_libcares=no])
 AM_CONDITIONAL(HAVE_LIBCARES, test "$found_libcares" = yes)
 AC_SUBST(found_libcares)
 
+found_libgtp_and_libcares=no
+if test "$found_libgtp" = "yes" -a "$found_libcares" = "yes"; then
+    found_libgtp_and_libcares=yes
+fi
+AC_SUBST(found_libgtp_and_libcares)
+
 dnl checks for header files
 AC_HEADER_STDC
 AC_CHECK_HEADERS(dahdi/user.h,,AC_MSG_WARN(DAHDI input driver will not be 
built))
diff --git a/openbsc/tests/atlocal.in b/openbsc/tests/atlocal.in
index 7475f63..362bfa9 100644
--- a/openbsc/tests/atlocal.in
+++ b/openbsc/tests/atlocal.in
@@ -2,5 +2,6 @@ enable_nat_test='@osmo_ac_build_nat@'
 enable_smpp_test='@osmo_ac_build_smpp@'
 enable_bsc_test='@osmo_ac_build_bsc@'
 enable_mgcp_transcoding_test='@osmo_ac_mgcp_transcoding@'
-enable_sgsn_test='@found_libgtp@'
-enable_gtphub_test='@found_libgtp@'
+enable_sgsn_test='@found_libgtp_and_libcares@'
+enable_oap_test='@found_libgtp_and_libcares@'
+enable_gtphub_test='@found_libgtp_and_libcares@'
]]]

And also, the OAP tests don't depend on GTP nor c-ares, but OAP compilation is
currently closely tied to the SGSN binary. Is it worthwhile to separate at this
point?

Thanks!
~Neels

-- 
- Neels Hofmeyr <nhofm...@sysmocom.de>          http://www.sysmocom.de/
=======================================================================
* sysmocom - systems for mobile communications GmbH
* Alt-Moabit 93
* 10559 Berlin, Germany
* Sitz / Registered office: Berlin, HRB 134158 B
* Geschäftsführer / Managing Directors: Holger Freyther, Harald Welte

signature.asc
Description: Digital signature