librfc1035: use UTF8 for DNS queries

DNS lookups use libidn to convert UTF8 hostname to ACE.

DNS replies convert ACE to UTF8 hostnames.

5 files changed, 115 insertions, 36 deletions

diff --git a/rfc1035/Makefile.am b/rfc1035/Makefile.am
index 759ac78..ee18cbe 100644
--- a/rfc1035/Makefile.am
+++ b/rfc1035/Makefile.am
@@ -1,9 +1,10 @@
 #
-# Copyright 1998 - 2011 Double Precision, Inc.  See COPYING for
+# Copyright 1998 - 2018 Double Precision, Inc.  See COPYING for
 # distribution information.
 
 
 noinst_LIBRARIES=librfc1035.a
+AM_CFLAGS=@LIBIDN_CFLAGS@
 
 librfc1035_a_SOURCES= \
 		rfc1035.c rfc1035.h rfc1035an.c \
@@ -16,7 +17,6 @@ librfc1035_a_SOURCES= \
 		rfc1035str.c rfc1035tcp.c rfc1035udp.c rfc1035bindsource.c \
 \
 		spf.c spf.h
-
 noinst_PROGRAMS=testlookup testspf
 
 testlookup_SOURCES=testlookup.c
@@ -24,7 +24,7 @@ testlookup_DEPENDENCIES=librfc1035.a ../rfc822/libencode.la ../md5/libmd5.la \
 	../random128/librandom128.la ../soxwrap/libsoxwrap.a \
 	../soxwrap/soxlibs.dep
 testlookup_LDADD=librfc1035.a ../rfc822/libencode.la ../soxwrap/libsoxwrap.a ../md5/libmd5.la \
-	../random128/librandom128.la `cat ../soxwrap/soxlibs.dep`
+	../random128/librandom128.la `cat ../soxwrap/soxlibs.dep` @LIBIDN_LIBS@
 testlookup_LDFLAGS=-static
 
 EXTRA_DIST=testsuite.txt
@@ -35,7 +35,7 @@ testspf_DEPENDENCIES=librfc1035.a ../rfc822/libencode.la ../md5/libmd5.la \
 	../soxwrap/soxlibs.dep
 testspf_LDADD=librfc1035.a ../rfc822/libencode.la ../md5/libmd5.la \
 	../random128/librandom128.la ../soxwrap/libsoxwrap.a \
-	`cat ../soxwrap/soxlibs.dep`
+	`cat ../soxwrap/soxlibs.dep` @LIBIDN_LIBS@
 testspf_LDFLAGS=-static
 
 mycheck:
diff --git a/rfc1035/configure.ac b/rfc1035/configure.ac
index f4243f9..efae100 100644
--- a/rfc1035/configure.ac
+++ b/rfc1035/configure.ac
@@ -23,6 +23,12 @@ AC_PROG_LIBTOOL
 
 dnl Checks for libraries.
 
+PKG_CHECK_MODULES(LIBIDN, libidn >= 0.0.0, [libidn=yes], [libidn=no])
+
+if test "$libidn" = "no"
+then
+	AC_MSG_ERROR(libidn not found)
+fi
 dnl Checks for header files.
 AC_HEADER_STDC
 AC_CHECK_HEADERS(sys/types.h sys/time.h unistd.h arpa/inet.h netinet/in.h)
@@ -174,12 +180,12 @@ UINT16="$int16"
 
 AC_CACHE_CHECK([for socklen_t],
 	courier_cv_hassocklen_t,
- 
+
 AC_COMPILE_IFELSE([
 AC_LANG_SOURCE( [
 #include <sys/types.h>
 #include <sys/socket.h>
- 
+
 socklen_t sl_t;
 ],[
 	accept(0, 0, &sl_t);
@@ -187,9 +193,9 @@ socklen_t sl_t;
 	courier_cv_hassocklen_t=yes,
 	courier_cv_hassocklen_t=no)
 )
- 
+
 socklen_t="int"
- 
+
 if test $courier_cv_hassocklen_t = yes
 then
 	:
diff --git a/rfc1035/rfc1035dump.c b/rfc1035/rfc1035dump.c
index 6b62a5e..0ee8639 100644
--- a/rfc1035/rfc1035dump.c
+++ b/rfc1035/rfc1035dump.c
@@ -1,5 +1,5 @@
 /*
-** Copyright 1998 - 2000 Double Precision, Inc.
+** Copyright 1998 - 2018 Double Precision, Inc.
 ** See COPYING for distribution information.
 */
 
@@ -12,7 +12,7 @@ static void print_hostname(FILE *f, const char *p)
 {
 	for ( ; *p; p++)
 	{
-		if (*p < ' ' || *p >= 127)
+		if ((unsigned char)*p < ' ')
 		{
 			fprintf(f, "\\%03o", (int)(unsigned char)*p);
 			continue;
@@ -48,10 +48,10 @@ struct	rfc1035_reply *qr;
 
 	fprintf(f, " (server %s)", ipbuf);
 
-	fprintf(f, ":\n;;  Bytes: %ld\n", 
+	fprintf(f, ":\n;;  Bytes: %ld\n",
 		(long)r->replylen);
 
-	fprintf(f, ";;  Opcode: %s\n", 
+	fprintf(f, ";;  Opcode: %s\n",
 		rfc1035_opcode_itostr(r->opcode));
 	fprintf(f, ";;  Flags:");
 	if (r->qr)
diff --git a/rfc1035/rfc1035reply.c b/rfc1035/rfc1035reply.c
index 98c3766..ac1e87e 100644
--- a/rfc1035/rfc1035reply.c
+++ b/rfc1035/rfc1035reply.c
@@ -1,5 +1,5 @@
 /*
-** Copyright 1998 - 1999 Double Precision, Inc.
+** Copyright 1998 - 2018 Double Precision, Inc.
 ** See COPYING for distribution information.
 */
 
@@ -7,7 +7,7 @@
 #include	<stdlib.h>
 #include	<string.h>
 #include	<errno.h>
-
+#include	<idna.h>
 
 void rfc1035_replyfree(struct rfc1035_reply *p)
 {
@@ -71,7 +71,23 @@ int	cnt=0;
 		*ptr += l;
 	}
 
-	if (namebuf)	namebuf[i]=0;
+	if (namebuf)
+	{
+		char *p;
+
+		namebuf[i]=0;
+
+		if (idna_to_unicode_8z8z(namebuf, &p, 0) != IDNA_SUCCESS)
+			return (0);
+
+		if (strlen(p) >= RFC1035_MAXNAMESIZE)
+		{
+			free(p);
+			return (0);
+		}
+		strcpy(namebuf, p);
+		free(p);
+	}
 	return (namebuf ? namebuf:"");
 }
 
diff --git a/rfc1035/rfc1035resolve.c b/rfc1035/rfc1035resolve.c
index d772eb0..b8bb91f 100644
--- a/rfc1035/rfc1035resolve.c
+++ b/rfc1035/rfc1035resolve.c
@@ -1,5 +1,5 @@
 /*
-** Copyright 1998 - 2000 Double Precision, Inc.
+** Copyright 1998 - 2018 Double Precision, Inc.
 ** See COPYING for distribution information.
 */
 
@@ -13,7 +13,7 @@
 #endif
 #include	<stdlib.h>
 #include	<string.h>
-
+#include	<idna.h>
 
 struct querybuf {
 	char qbuf[512];
@@ -30,25 +30,82 @@ struct querybuf *qp=(struct querybuf *)q;
 	qp->qbuflen += l;
 }
 
-struct rfc1035_reply *rfc1035_resolve_multiple(
-			struct rfc1035_res *res,
-			int opcode,
-			const struct rfc1035_query *queries,
-			unsigned nqueries)
+
+static struct rfc1035_reply
+*rfc1035_resolve_multiple_idna(struct rfc1035_res *res,
+			       int opcode,
+			       const struct rfc1035_query *queries,
+			       unsigned nqueries);
+
+struct rfc1035_reply
+*rfc1035_resolve_multiple(struct rfc1035_res *res,
+			  int opcode,
+			  const struct rfc1035_query *queries,
+			  unsigned nqueries)
 {
-struct	querybuf qbuf;
-int	udpfd;
-int	attempt;
-const RFC1035_ADDR *ns;
-unsigned nscount;
-unsigned current_timeout, timeout_backoff;
-unsigned nbackoff, backoff_num;
-int	af;
-static const char fakereply[]={0, 0, 0, RFC1035_RCODE_SERVFAIL,
-		0, 0,
-		0, 0,
-		0, 0,
-		0, 0};
+	struct rfc1035_query *idna_queries=
+		(struct rfc1035_query *)malloc(nqueries * sizeof(*queries));
+	struct rfc1035_reply *r;
+	unsigned n;
+	char **buffers;
+
+	/*
+	** Translate each label from UTF8 to IDNA.
+	*/
+
+	if (!idna_queries)
+		return NULL;
+	if ((buffers=(char **)malloc(nqueries * sizeof(char *))) == NULL)
+	{
+		free(idna_queries);
+		return NULL;
+	}
+	for (n=0; n<nqueries; ++n)
+	{
+		idna_queries[n]=queries[n];
+		if (idna_to_ascii_8z(idna_queries[n].name, &buffers[n], 0)
+		    != IDNA_SUCCESS)
+		{
+			errno=EINVAL;
+			while (n)
+			{
+				free(buffers[--n]);
+			}
+			free(idna_queries);
+			free(buffers[n]);
+			return NULL;
+		}
+
+		idna_queries[n].name=buffers[n];
+	}
+
+	r=rfc1035_resolve_multiple_idna(res, opcode, idna_queries, nqueries);
+	for (n=0; n<nqueries; ++n)
+		free(buffers[n]);
+	free(idna_queries);
+	free(buffers);
+	return r;
+}
+
+static struct rfc1035_reply
+*rfc1035_resolve_multiple_idna(struct rfc1035_res *res,
+			       int opcode,
+			       const struct rfc1035_query *queries,
+			       unsigned nqueries)
+{
+	struct	querybuf qbuf;
+	int	udpfd;
+	int	attempt;
+	const RFC1035_ADDR *ns;
+	unsigned nscount;
+	unsigned current_timeout, timeout_backoff;
+	unsigned nbackoff, backoff_num;
+	int	af;
+	static const char fakereply[]={0, 0, 0, RFC1035_RCODE_SERVFAIL,
+				       0, 0,
+				       0, 0,
+				       0, 0,
+				       0, 0};
 
 	nscount=res->rfc1035_nnameservers;
 	ns=res->nameservers;
@@ -184,7 +241,7 @@ static const char fakereply[]={0, 0, 0, RFC1035_RCODE_SERVFAIL,
 				if ((reply=rfc1035_recv_tcp(res,
 					tcpfd, &nbytes, current_timeout))==0)
 					break;
-				
+
 				rfcreply=rfc1035_replyparse(reply, nbytes);
 				if (!rfcreply)
 				{