From 44653071c3de365b9ee4bf951d00e47c5edd686c Mon Sep 17 00:00:00 2001
From: Jordan Reiter <jordan@aace.org>
Date: Wed, 27 Mar 2013 13:05:29 -0400
Subject: [PATCH 1/2] Fixes for LZW error (see
 https://bitbucket.org/hsoft/pdfminer3k/commits/ae9a4ca0691a/)

---
 pdfminer/lzw.py       | 23 +++++++++++++++++------
 pdfminer/pdfinterp.py |  5 ++++-
 2 files changed, 21 insertions(+), 7 deletions(-)

diff --git a/pdfminer/lzw.py b/pdfminer/lzw.py
index 6d0eaf1..5a1a668 100644
--- a/pdfminer/lzw.py
+++ b/pdfminer/lzw.py
@@ -6,6 +6,10 @@ except ImportError:
     from StringIO import StringIO
 
 
+class CorruptDataError(Exception):
+    pass
+
+
 ##  LZWDecoder
 ##
 class LZWDecoder(object):
@@ -46,12 +50,12 @@ class LZWDecoder(object):
         return v
 
     def feed(self, code):
-        x = ''
+        x = b''
         if code == 256:
-            self.table = [ chr(c) for c in xrange(256) ] # 0-255
+            self.table = [bytes([i]) for i in range(256)] # 0-255
             self.table.append(None) # 256
             self.table.append(None) # 257
-            self.prevbuf = ''
+            self.prevbuf = b''
             self.nbits = 9
         elif code == 257:
             pass
@@ -60,10 +64,12 @@ class LZWDecoder(object):
         else:
             if code < len(self.table):
                 x = self.table[code]
-                self.table.append(self.prevbuf+x[0])
-            else:
-                self.table.append(self.prevbuf+self.prevbuf[0])
+                self.table.append(self.prevbuf+x[:1])
+            elif code == len(self.table):
+                self.table.append(self.prevbuf+self.prevbuf[:1])
                 x = self.table[code]
+            else:
+                raise CorruptDataError()
             l = len(self.table)
             if l == 511:
                 self.nbits = 10
@@ -81,6 +87,11 @@ class LZWDecoder(object):
             except EOFError:
                 break
             x = self.feed(code)
+            try:
+                x = self.feed(code)
+            except CorruptDataError:
+                # just ignore corrupt data and stop yielding there
+                break
             yield x
             if self.debug:
                 print >>sys.stderr, ('nbits=%d, code=%d, output=%r, table=%r' %
diff --git a/pdfminer/pdfinterp.py b/pdfminer/pdfinterp.py
index 3b0c789..c0c7ed2 100644
--- a/pdfminer/pdfinterp.py
+++ b/pdfminer/pdfinterp.py
@@ -217,7 +217,10 @@ class PDFContentParser(PSStackParser):
                 self.istream += 1
             else:
                 raise PSEOF('Unexpected EOF, file truncated?')
-            self.fp = StringIO(strm.get_data())
+            data = strm.get_data()
+            if isinstance(data, bytes):
+                data = data.decode('latin-1')
+            self.fp = io.StringIO(data)
         return
 
     def seek(self, pos):

From e28b75a462ccc189c93ef8a7a9a806966fb51583 Mon Sep 17 00:00:00 2001
From: Jordan Reiter <jordan@aace.org>
Date: Wed, 27 Mar 2013 13:14:58 -0400
Subject: [PATCH 2/2] StringIO

---
 pdfminer/pdfinterp.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/pdfminer/pdfinterp.py b/pdfminer/pdfinterp.py
index c0c7ed2..15f3ee0 100644
--- a/pdfminer/pdfinterp.py
+++ b/pdfminer/pdfinterp.py
@@ -220,7 +220,7 @@ class PDFContentParser(PSStackParser):
             data = strm.get_data()
             if isinstance(data, bytes):
                 data = data.decode('latin-1')
-            self.fp = io.StringIO(data)
+            self.fp = StringIO(data)
         return
 
     def seek(self, pos):