From ecc4d056754a528706f7bf43f8754c448ea2dd4b Mon Sep 17 00:00:00 2001
From: speedplane <mes65@cornell.edu>
Date: Tue, 11 Nov 2014 23:34:33 -0500
Subject: [PATCH] Fix a unicode conversion bug.

See https://github.com/euske/pdfminer/issues/75
---
 pdfminer/psparser.py | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/pdfminer/psparser.py b/pdfminer/psparser.py
index c1ebe93..be715af 100644
--- a/pdfminer/psparser.py
+++ b/pdfminer/psparser.py
@@ -343,7 +343,15 @@ class PSBaseParser(object):
             self.hex = b''
             self._parse1 = self._parse_literal_hex
             return j+1
-        self._add_token(LIT(unicode(self._curtoken)))
+        
+        try:
+            # Try to interpret the token as a utf-8 string
+            utoken = self._curtoken.decode('utf-8')
+        except UnicodeDecodeError:
+            # We failed, there is possibly a corrupt PDF here.
+            if STRICT: raise
+            utoken = ""
+        self._add_token(LIT(utoken))
         self._parse1 = self._parse_main
         return j