From 118013befdcdcaf96c64657439c441f0108fbdcc Mon Sep 17 00:00:00 2001
From: Chris Dyer <cdyer@allegro.clab.cs.cmu.edu>
Date: Mon, 15 Sep 2014 23:00:01 -0400
Subject: migrate to new Cython version

---
 corpus/support/tokenizer.pl | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'corpus/support/tokenizer.pl')

diff --git a/corpus/support/tokenizer.pl b/corpus/support/tokenizer.pl
index f57bc87a..aa285be4 100755
--- a/corpus/support/tokenizer.pl
+++ b/corpus/support/tokenizer.pl
@@ -386,7 +386,7 @@ sub deep_proc_token {
     }
 
     ##### step 1: separate by punct T2 on the boundary
-    my $t2 = '\`|\!|\@|\+|\=|\[|\]|\<|\>|\||\(|\)|\{|\}|\?|\"|;';
+    my $t2 = '\`|\!|\@|\+|\=|\[|\]|\<|\>|\||\(|\)|\{|\}|\?|\"|;|●|○';
     if($line =~ s/^(($t2)+)/$1 /){
 	return proc_line($line);
     }
-- 
cgit v1.2.3