svn commit: r1695182 - in /spamassassin/trunk/lib/Mail/SpamAssassin/Plugin: ASN.pm AskDNS.pm URIDNSBL.pm

[mm...@apache.org]

Author: mmartinec
Date: Mon Aug 10 23:43:53 2015
New Revision: 1695182

URL: http://svn.apache.org/r1695182
Log:
Bug 7236: Net::DNS assumes strings in TXT resource records are in UTF-8 and gratuitously tries to decodes it

Modified:
    spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/ASN.pm
    spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/AskDNS.pm
    spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/URIDNSBL.pm

Modified: spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/ASN.pm
URL: http://svn.apache.org/viewvc/spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/ASN.pm?rev=1695182&r1=1695181&r2=1695182&view=diff
==============================================================================
--- spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/ASN.pm (original)
+++ spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/ASN.pm Mon Aug 10 23:43:53 2015
@@ -355,8 +355,10 @@ sub process_dns_result {
   foreach my $rr (@answer) {
     dbg("asn: %s: lookup result packet: %s", $zone, $rr->string);
     next if $rr->type ne 'TXT';
-    my @strings = $rr->char_str_list;
+    my @strings = Net::DNS->VERSION >= 0.69 ? $rr->txtdata
+                                            : $rr->char_str_list;
     next if !@strings;
+    for (@strings) { utf8::encode($_) if utf8::is_utf8($_) }
 
     my @items;
     if (@strings > 1 && join('',@strings) !~ m{\|}) {

Modified: spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/AskDNS.pm
URL: http://svn.apache.org/viewvc/spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/AskDNS.pm?rev=1695182&r1=1695181&r2=1695182&view=diff
==============================================================================
--- spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/AskDNS.pm (original)
+++ spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/AskDNS.pm Mon Aug 10 23:43:53 2015
@@ -574,6 +574,7 @@ sub process_response_packet {
         if ($rr_rdatastr =~ m/^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}\z/) {
           $rdatanum = Mail::SpamAssassin::Util::my_inet_aton($rr_rdatastr);
         }
+
       } elsif ($rr->UNIVERSAL::can('txtdata')) {
         # TXT, SPF: join with no intervening spaces, as per RFC 5518
         if ($txtdata_can_provide_a_list || $rr_type ne 'TXT') {
@@ -581,10 +582,22 @@ sub process_response_packet {
         } else {  # char_str_list() is only available for TXT records
           $rr_rdatastr = join('', $rr->char_str_list);  # historical
         }
+        # Net::DNS attempts to decode text strings in a TXT record as UTF-8,
+        # which is bad: octets failing the UTF-8 decoding are converted to
+        # three octets \x{EF}\x{BF}\x{BD} (i.e. to a Unicode "replacement
+        # character" U+FFFD encoded as UTF-8), and ASCII text is unnecessarily
+        # flagged as perl native characters (utf8 flag on), which can be
+        # disruptive on later processing, e.g. implicitly upgrading strings
+        # on concatenation. Unfortunately there is no way of legally bypassing
+        # the UTF-8 decoding by Net::DNS::RR::TXT in Net::DNS::RR::Text.
+        # Try to minimize damage by encoding back to UTF-8 octets:
+        utf8::encode($rr_rdatastr)  if utf8::is_utf8($rr_rdatastr);
+
       } else {
         # rdatastr() is historical, use rdstring() since Net::DNS 0.69
         $rr_rdatastr = $rr->UNIVERSAL::can('rdstring') ? $rr->rdstring
                                                        : $rr->rdatastr;
+        utf8::encode($rr_rdatastr)  if utf8::is_utf8($rr_rdatastr);
       }
     # dbg("askdns: received rr type %s, data: %s", $rr_type, $rr_rdatastr);
     }

Modified: spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/URIDNSBL.pm
URL: http://svn.apache.org/viewvc/spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/URIDNSBL.pm?rev=1695182&r1=1695181&r2=1695182&view=diff
==============================================================================
--- spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/URIDNSBL.pm (original)
+++ spamassassin/trunk/lib/Mail/SpamAssassin/Plugin/URIDNSBL.pm Mon Aug 10 23:43:53 2015
@@ -1108,6 +1108,7 @@ sub complete_dnsbl_lookup {
       # avoid space-separated RDATA <character-string> fields if possible;
       # txtdata provides a list of strings in list context since Net::DNS 0.69
       $rdatastr = join('',$rr->txtdata);
+      utf8::encode($rdatastr)  if utf8::is_utf8($rdatastr);
     } else {
       next;
     }