Add a timeout to the PDF metadata writer as it hangs on some PDF files

src/calibre/ebooks/metadata/pdf.py

@@ -2,7 +2,8 @@ __license__   = 'GPL v3'
 __copyright__ = '2008, Kovid Goyal <kovid at kovidgoyal.net>'
 '''Read meta information from PDF files'''
 
-import sys, os, StringIO
+import sys, os, cStringIO
+from threading import Thread
 
 from calibre.ebooks.metadata import MetaInformation, authors_to_string, get_parser
 from pyPdf import PdfFileReader, PdfFileWriter
@@ -29,24 +30,45 @@ def get_metadata(stream):
         print >>sys.stderr, msg.encode('utf8')
     return mi
 
+class MetadataWriter(Thread):
+
+    def __init__(self, out_pdf, buf):
+        self.out_pdf = out_pdf
+        self.buf = buf
+        Thread.__init__(self)
+        self.daemon = True
+
+    def run(self):
+        try:
+            self.out_pdf.write(self.buf)
+        except RuntimeError:
+            pass
+
 def set_metadata(stream, mi):
     stream.seek(0)
 
     # Use a StringIO object for the pdf because we will want to over
     # write it later and if we are working on the stream directly it
     # could cause some issues.
-    raw = StringIO.StringIO(stream.read())
+    raw = cStringIO.StringIO(stream.read())
     orig_pdf = PdfFileReader(raw)
 
     title = mi.title if mi.title else orig_pdf.documentInfo.title
     author = authors_to_string(mi.authors) if mi.authors else orig_pdf.documentInfo.author
 
     out_pdf = PdfFileWriter(title=title, author=author)
+    out_str = cStringIO.StringIO()
+    writer = MetadataWriter(out_pdf, out_str)
     for page in orig_pdf.pages:
         out_pdf.addPage(page)
 
-    out_str = StringIO.StringIO()
-    out_pdf.write(out_str)
+    writer.start()
+    writer.join(10) # Wait 10 secs for writing to complete
+    out_pdf.killed = True
+    writer.join()
+    if out_pdf.killed:
+        print 'Failed to set metadata: took too long'
+        return
 
     stream.seek(0)
     stream.truncate()

src/pyPdf/generic.py

@@ -299,7 +299,7 @@ def readStringFromStream(stream):
             elif tok == "t":
                 tok = "\t"
             elif tok == "b":
-                tok == "\b"
+                tok = "\b"
             elif tok == "f":
                 tok = "\f"
             elif tok == "(":

src/pyPdf/pdf.py

@@ -39,15 +39,12 @@ __author__ = "Mathieu Fenniak"
 __author_email__ = "biziqe@mathieu.fenniak.net"
 
 import struct
-try:
 from cStringIO import StringIO
-except ImportError:
-    from StringIO import StringIO
 
-import filters
-import utils
-import warnings
-from generic import *
+from generic import DictionaryObject, NameObject, NumberObject, \
+createStringObject, ArrayObject, ByteStringObject, StreamObject, \
+IndirectObject, utils, readObject, TextStringObject, BooleanObject, \
+RectangleObject, DecodedStreamObject
 from utils import readNonWhitespace, readUntilWhitespace, ConvertFunctionsToVirtualList
 
 
@@ -56,6 +53,7 @@ from utils import readNonWhitespace, readUntilWhitespace, ConvertFunctionsToVirt
 # class (typically {@link #PdfFileReader PdfFileReader}).
 class PdfFileWriter(object):
     def __init__(self,title=u"Unknown",author=u"Unknown"):
+        self.killed = False
         self._header = "%PDF-1.3"
         self._objects = []  # array of indirect objects
 
@@ -162,7 +160,7 @@ class PdfFileWriter(object):
     # @param stream An object to write the file to.  The object must support
     # the write method, and the tell method, similar to a file object.
     def write(self, stream):
-        import struct, md5
+        import md5
 
         externalReferenceMap = {}
         self.stack = []
@@ -214,6 +212,8 @@ class PdfFileWriter(object):
         stream.write("\nstartxref\n%s\n%%%%EOF\n" % (xref_location))
 
     def _sweepIndirectReferences(self, externMap, data):
+        if self.killed:
+            raise RuntimeError('Writer killed')
         if isinstance(data, DictionaryObject):
             for key, value in data.items():
                 origvalue = value