kannel 1.4.0 portability fixes

Hi,

I've been trying to compile kannel 1.4.0 with a non-gcc compiler, and
have a small pile of patches required to do so.

Anyway, here are the first two issues; is this the best place to
report them?


In radius/radius_pdu.c:

--- radius/radius_pdu.c 22 Jan 2004 14:08:25 -0000      1.1.1.1
+++ radius/radius_pdu.c 15 Feb 2005 15:02:32 -0000      1.1.1.1.2.1
@@ -296,7 +296,7 @@
                 int c = octstr_get_char(value, i);
                 Octstr *b = octstr_format("%d", c);
                 octstr_append(ret, b);
-                i < 3 ? octstr_append_cstr(ret, ".") : NULL;
+                if (i < 3) octstr_append_cstr(ret, ".");
                 octstr_destroy(b);
             }
             break;


This is a syntax error that GCC will complain about with "-pedantic
-Wall", but the fix (as you can see) is simple enough.   (It's a
syntax error because octstr_append_cstr has type "void", and NULL
isn't a void, but a deeply magic thing, so the two halfs of the ?:
have incompatible types).


Secondly, in gw/smsc/smsc_smpp.c:

--- gw/smsc/smsc_smpp.c 21 Oct 2004 12:33:40 -0000      1.1.1.2
+++ gw/smsc/smsc_smpp.c 16 Feb 2005 11:04:07 -0000      1.1.1.2.2.2
@@ -1548,7 +1548,7 @@
                     }
                 }
                 debug("bb.sms.smpp", 0, "SMPP[%s]: %s: break and shutting down",
-                      octstr_get_cstr(smpp->conn->id), __PRETTY_FUNCTION__);
+                      octstr_get_cstr(smpp->conn->id), __func__);

                 break;
             }

__PRETTY_FUNCTION__ is a GCCism.  gw-config.h already has a construct to 
define a replacement for the C99 __func__ macro if one is required, so 
using that seems sensible (and increases portability to older versions
of gcc for free).



-- 
Mike Bristow - really a very good driver