filter: use regexp to check multipart bodies

This should be safer than running file(1), which has had its share
of vulnerabilities this year (early 2014) We really only care about
diffs and maybe short log files, here.

1 files changed, 5 insertions, 15 deletions

diff --git a/lib/PublicInbox/Filter.pm b/lib/PublicInbox/Filter.pm
index 0b4dce87..4a348217 100644
--- a/lib/PublicInbox/Filter.pm
+++ b/lib/PublicInbox/Filter.pm
@@ -215,23 +215,13 @@ sub replace_body {
         mark_changed($simple);
 }
 
-# run the file(1) command to detect mime type
-# Not using File::MMagic for now since that requires extra configuration
-# Note: we do not rewrite the message with the detected mime type
+# Check for display-able text, no messed up binaries
+# Note: we can not rewrite the message with the detected mime type
 sub recheck_type_ok {
         my ($part) = @_;
-        my $cmd = "file --mime-type -b -";
-        my $pid = open2(my $out, my $in, $cmd);
-        print $in $part->body;
-        close $in;
-        my $type = eval {
-                local $/;
-                <$out>;
-        };
-        waitpid($pid, 0);
-        chomp $type;
-
-        (($type =~ $MIME_TEXT_ANY) && ($type !~ $MIME_HTML))
+        my $s = $part->body;
+        ((bytes::length($s) < 0x10000) &&
+                ($s =~ /\A([\P{XPosixPrint}\f\n\r\t]+)\z/))
 }
 
 1;