Properly handle the case of missing CCR dependency.

The new Chakra website does not return an HTTP error, but some
error page, which chaser tries to interpret as the package tarfile,
raising an exception. Catch the exception and print a more helpful
error message.

chaser/chaser.py

@@ -33,10 +33,17 @@ def get_source_files(args, workingdir=None):
         os.mkdir(workingdir)
 
     for pkgname in pkgnames:
+        print('Downloading %s' % pkgname)
         r = requests.get(ccr.pkg_url(pkgname))
         r.raise_for_status()
-        tar = tarfile.open(mode='r', fileobj=io.BytesIO(r.content))
-        tar.extractall(workingdir)
+
+        try:
+            tar = tarfile.open(mode='r', fileobj=io.BytesIO(r.content))
+            tar.extractall(workingdir)
+        except tarfile.ReadError as re:
+            print('Unable to open the tar file. Either the package does not '
+                  'exist in CCR or it is malformed: %s' % str(re))
+
 
 def recurse_depends(pkgname, workingdir=None, graph=None):
     """Build a dependency graph"""