Comment the logic for database decode_utf8/is_utf8 checking

lib/RT/Record.pm

@@ -656,12 +656,16 @@ sub __Value {
 
     return undef if (!defined $value);
 
+    # Pg returns character columns as character strings; mysql and
+    # sqlite return them as bytes.  While mysql can be made to return
+    # characters, using the mysql_enable_utf8 flag, the "Content" column
+    # is bytes on mysql and characters on Postgres, making true
+    # consistency impossible.
     if ( $args{'decode_utf8'} ) {
-        if ( !utf8::is_utf8($value) ) {
+        if ( !utf8::is_utf8($value) ) { # mysql/sqlite
             utf8::decode($value);
         }
-    }
-    else {
+    } else {
         if ( utf8::is_utf8($value) ) {
             utf8::encode($value);
         }

lib/RT/Report/Tickets.pm

@@ -745,6 +745,9 @@ sub MapSubValues {
         my $dst = $item->{'values'}{ lc $to } = { };
         while (my ($k, $v) = each %{ $map } ) {
             $dst->{ $k } = delete $item->{'values'}{ lc $v->{'NAME'} };
+            # This mirrors the logic in RT::Record::__Value When that
+            # ceases tp use the UTF-8 flag as a character/byte
+            # distinction from the database, this can as well.
             utf8::decode( $dst->{ $k } )
                 if defined $dst->{ $k }
                and not utf8::is_utf8( $dst->{ $k } );