Py3: Encode strings when compressing

[TomAugspurger]

Another drive by report. Maybe I'll contribute someday :) This one should be easy

In [1]: import odo

In [2]: with open('tst.csv', 'w') as f:
    f.write('a,1\nb,2\nc,3')
   ...:

In [3]: !cat tst.csv
a,1
b,2
c,3
In [4]: odo.odo('tst.csv', 'tst.csv.gz')

raises

...
/usr/local/Cellar/python3/3.4.3/Frameworks/Python.framework/Versions/3.4/lib/python3.4/gzip.py in write(self, data)
    341         if len(data) > 0:
    342             self.size = self.size + len(data)
--> 343             self.crc = zlib.crc32(data, self.crc) & 0xffffffff
    344             self.fileobj.write( self.compress.compress(data) )
    345             self.offset += len(data)

TypeError: 'str' does not support the buffer interface

I believe that

ipdb> zlib.crc32(data.encode('utf-8'), self.crc) & 0xffffffff
2324740828

Or something like sys.defaultencoding instead of utf-8 is the fix.

In [5]: odo.__version__
Out[5]: '0.3.2'

[cpcloud]

Hm actually, this looks like a bug.

[cpcloud]

Thanks for the report!