Don't follow the NULL pointer, not even in debug printouts.

Bug reported by Leonhard Knauff.

Closes RADSECPROXY-68.

AUTHORS

@@ -13,6 +13,7 @@ Fabian Mauchle
 Faidon Liambotis
 Hans Zandbelt
 Kolbjørn Barmen
+Leonhard Knauff
 Linus Nordberg
 Maja Wolniewicz
 Milan Sova

ChangeLog

@@ -3,6 +3,7 @@
 	- Use a listen(2) backlog of 128 (RADSECPROXY-72).
 
 	Bug fixes:
+	- Don't follow NULL the pointer at debug level 5 (RADSECPROXY-68).
 	- Completely reload CAs and CRLs with cacheExpiry (RADSECPROXY-50).
 	- Tie Access-Request log lines to response log lines (RADSECPROXY-60).
 	- Fix a couple of memory leaks and NULL ptr derefs in error cases.

hostport.c

@@ -172,7 +172,10 @@ int resolvehostport(struct hostportres *hp, int af, int socktype, uint8_t passiv
 	    }
 	}
     }
-    debug(DBG_DBG, "%s: %s -> %s", __func__, hp->host, addr2string(hp->addrinfo->ai_addr));
+    debug(DBG_DBG, "%s: %s -> %s", __func__,
+          (hp->host ? hp->host : "(src info not available)"),
+          ((hp->addrinfo && hp->addrinfo->ai_addr) ?
+           addr2string(hp->addrinfo->ai_addr) : "(dst info not available)"));
     return 1;
 
 errexit: