make it easier to patch-out "six"

[a-detiste]

No description provided.

[avylove]

Can you provide a description and rationale? I'm confused because this is targeting 3 of the 9 files that import six and they are all test files.

[a-detiste]

Hi, I wrote up this rationale two years ago. https://wiki.debian.org/Python3-six-removal

This does not really apply to your project: it is alive and well maintained.

I use it myself and consider patching-out the usage of six, as it is the very last package doing so.

This current PR upstream some of that work, so I & you don't have to maintain it anymore.

The later fix would be:

-from six import StringIO
+from io import StringIO

I would gladly send a PR that removes all Python 2 compatibility if that were your wish

[avylove]

Appreciate the use and PR.

I don't think we're looking at removing Python 2.7 support at this point since it doesn't really have much of a benefit and we had 15.5k downloads for 2.7 in the last month from PyPI.

I'm not sure I agree with your rational. It seems to be using a dependency on six as a proxy for unrelated things. That said, we don't really have a hard dependency on six. We use it for maybe 5 things that could easily be moved to version guards or a util/compat module.

But again, this PR has no description and only touches 3 of 9 files that use six, so I wasn't sure if you meant this to be draft PR or not.

[a-detiste]

But again, this PR has no description and only touches 3 of 9 files that use six, so I wasn't sure if you meant this to be draft PR or not.

I want to patch-out six, but keep patch as small as possible.

I also always propose patches to remove six.print_() six.u('constant') six.u('constant') because these are not anymore needed if 2.7 & >3.2 are combined

[avylove]

six was dropped in #282