fix for www.gov.uk having iswwwdup bug.

because .gov.uk is a tld and so is .uk.
also added code to handle seg fault signals
better and run the default handler after saving
rather than calling abort(). hopefully a core
will be dumped all the time now.

Process.cpp

@@ -1666,9 +1666,16 @@ bool Process::shutdown2 ( ) {
 		if ( setrlimit(RLIMIT_CORE,&lim) )
 			log("gb: setrlimit: %s.", mstrerror(errno) );
 
+		// this is the trick: it will trigger the core dump by
+		// calling the original SIGSEGV handler.
+		int signum = SIGSEGV;
+		signal(signum, SIG_DFL);
+		kill(getpid(), signum);
+		// default handler should be called after we return now
+
 		// . force an abnormal termination which will cause a core dump
 		// . do not dump core on SIGHUP signals any more though
-		abort();
+		//abort();
 		// keep compiler happy
 		return true;
 	}

XmlDoc.cpp

@@ -15160,6 +15160,16 @@ char *XmlDoc::getIsWWWDup ( ) {
 		return &m_isWWWDup;
 	}
 
+	// watch out for idiot urls like www.gov.uk and www.gov.za
+	// treat them as though the TLD is uk/za and the domain
+	// is gov.uk and gov.za
+	if ( u->getDomain() &&
+	     strncmp ( u->getDomain() , "www." , 4 ) == 0 ) {
+		m_isWWWDup      = false;
+		m_isWWWDupValid = true;
+		return &m_isWWWDup;
+	}
+
 	// make it without the www
 	char withoutWWW[MAX_URL_LEN+1];
 	char *proto = "http";