Str to byte fix for DTM wrapper.

[bhargavvader]

@piskvorky , this is with respect to #698.

You were right, all that had to be done was use utils.to_utf8 before writing to file.
Works fine with both python2 and 3.

[bhargavvader]

Note: would ideally want to convert to bytes from the list itself instead of converting to string and then to byte.

[bhargavvader]

@tmylk , I can't directly convert from integer to bytes while writing to file, it has to be to string first.
For e.g, bytes(4) is b'\x00\x00\x00\x00', but str(4).encode('utf-8') is b'4' as expected.

Is there any idea for a workaround to this?

[piskvorky]

I don't get this. Doesn't six.u return unicode? How does this work @tmylk ?

[piskvorky]

@bhargavvader we want to store binary strings, not "bytes" as such. We don't want to be storing numbers as raw bytes here, as in b'\x00\x00\x00\x04'. We want to store their (binary) string representation, as in b'4'. We just don't want to be storing these strings as unicode, as in u'4'.

Integers are always representable in ASCII encoding (subset of utf8), so the bytes/unicode conversion is no problem.

[bhargavvader]

@piskvorky , here it says that six.u returns unicode in python 2 and string in python 3.

I understand we'd want to store it in binary strings, yeah.
So can the six.u be removed? Can it just be utils.to_utf8(str()) in all places?

Edit: Put in a PR and removed six.u. Works fine.

[piskvorky]

I don't understand how this version with storing unicode to binary files even worked. It means our unit tests must be faulty / incomplete.

[klvbdmh]

Were the changes pushed to PyPI? Pip installs version 0.13.1. and the bug still persists.