precheck: stricter checks including min length

We should reject values which are too short to be useful or sane.

2 files changed, 22 insertions, 6 deletions

diff --git a/lib/PublicInbox.pm b/lib/PublicInbox.pm
index b05fd8c5..cfa9d4bf 100644
--- a/lib/PublicInbox.pm
+++ b/lib/PublicInbox.pm
@@ -4,6 +4,7 @@ package PublicInbox;
 use strict;
 use warnings;
 use Email::Address;
+use Date::Parse qw(strptime);
 use constant MAX_SIZE => 1024 * 500; # same as spamc default
 
 # drop plus addressing for matching
@@ -17,13 +18,25 @@ sub __drop_plus {
 sub precheck {
         my ($klass, $filter, $recipient) = @_;
         my $simple = $filter->simple;
-        return 0 unless $simple->header("Message-ID");
-        return 0 unless defined($filter->from);
-        return 0 unless $simple->header("Subject");
+        my $mid = $simple->header("Message-ID");
+        return 0 unless usable_str(length('<m@h>'), $mid) && $mid =~ /\@/;
+        return 0 unless usable_str(length('u@h'), $filter->from);
+        return 0 unless usable_str(length(':o'), $simple->header("Subject"));
+        return 0 unless usable_date($simple->header("Date"));
         return 0 if length($simple->as_string) > MAX_SIZE;
         recipient_specified($filter, $recipient);
 }
 
+sub usable_str {
+        my ($len, $str) = @_;
+        defined($str) && length($str) >= $len;
+}
+
+sub usable_date {
+        my @t = eval { strptime(@_) };
+        scalar @t;
+}
+
 sub recipient_specified {
         my ($filter, $recipient) = @_;
         defined($recipient) or return 1; # for mass imports
diff --git a/t/precheck.t b/t/precheck.t
index 1bfa4c9f..acfd5e8b 100644
--- a/t/precheck.t
+++ b/t/precheck.t
@@ -36,7 +36,8 @@ sub do_checks {
                         Cc => 'c@example.com',
                         'Content-Type' => 'text/plain',
                         Subject => 'this is a subject',
-                        'Message-ID' => '<MID>',
+                        'Message-ID' => '<MID@host>',
+                        Date => 'Wed, 09 Apr 2014 01:28:34 +0000',
                 ],
                 body => "hello world\n",
         ));
@@ -50,7 +51,8 @@ sub do_checks {
                         Cc => 'John Doe <c@example.com>',
                         'Content-Type' => 'text/plain',
                         Subject => 'this is a subject',
-                        'Message-ID' => '<MID>',
+                        'Message-ID' => '<MID@host>',
+                        Date => 'Wed, 09 Apr 2014 01:28:34 +0000',
                 ],
                 body => "hello world\n",
         ));
@@ -64,7 +66,8 @@ sub do_checks {
                         Cc => 'c@example.com',
                         'Content-Type' => 'text/plain',
                         Subject => 'this is a subject',
-                        'Message-ID' => '<MID>',
+                        'Message-ID' => '<MID@host>',
+                        Date => 'Wed, 09 Apr 2014 01:28:34 +0000',
                 ],
                 body => "hello world\n",
         );