[ck2cti] Fix reading some input files containing non-unicode characters

In Python 3, attempting to convert some mechanisms would give errors like the
following:

UnicodeDecodeError: 'utf-8' codec can't decode byte 0x96 in position 348:
invalid start byte

interfaces/cython/cantera/ck2cti.py

@@ -64,15 +64,24 @@ ENERGY_UNITS = {'CAL/': 'cal/mol',
                 'KJOULES/MOL': 'kJ/mol',
                 'KJOULES/MOLE': 'kJ/mol'}
 
+_open = open
 if sys.version_info[0] == 2:
     string_types = (str, unicode)
     def strip_nonascii(s):
         return s.decode('ascii', 'ignore')
+
+    def open(filename, *args):
+        return _open(filename, *args)
+
 else:
     string_types = (str,)
     def strip_nonascii(s):
         return s.encode('ascii', 'ignore').decode()
 
+    def open(filename, *args):
+        return _open(filename, *args, errors='ignore')
+
+
 def compatible_quantities(quantity_basis, units):
     if quantity_basis == 'mol':
         return 'molec' not in units

test/data/pdep-test.inp

@@ -1,4 +1,5 @@
 ! Generic mechanism header
+! This line contains a non-unicode character (–) that should just be ignored.
 ELEMENTS
 H  C
 END