add separate want output for py2 vs. py3

py2 generates different output than py3 for some tests (see main_test.go notes for details) -- for full testing to work properly, a separate want output is required for each. for now, only doing py3 testing.

Start by reading the notes in main_test.go to identify which tests produce different output under Python 2 and Python 3. Add separate expected output for each version, then verify that the tests use the appropriate output for the interpreter being tested.

Written by the indexing model from the issue text.