READY FOR REVIEW - Fix #171. Pass newbytes's native value to its parent constructor.

[posita]

Fix #171. Pass newbytes's native value to its parent constructor.

CAVEAT: without monkey-patching Python 2's native str's constructor, I do not know how to handle this case:

Python 2.7.10 (default, Sep 24 2015, 10:13:45)
[GCC 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> nativebytes = bytes ; nativestr = str ; from builtins import *
>>> nativebytes(bytes(b'asdf'))
"b'asdf'" # Whoops!
>>> # This means you can't pass newbytes in many contexts, such as:
>>> from urllib import urlencode
>>> urlencode({ bytes(b'a'): 1, bytes(b'b'): 2 })
'b%27a%27=1&b%27b%27=2'
>>> # :o(

But this works now:

Python 2.7.10 (default, Sep 24 2015, 10:13:45)
[GCC 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.57)] on darwin
>>> nativebytes = bytes ; nativestr = str ; from builtins import *
>>> bytes(str(u'asdf'), u'utf_8')
b'asdf'

Some things just won't work (and are not addressed by this PR). Consider:

Python 2.7.10 (default, Sep 24 2015, 10:13:45)
[GCC 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> nativebytes = bytes ; nativestr = str ; from builtins import *
>>> from urllib import urlencode
>>> urlencode({ bytes(b'a'): 1, bytes(b'b'): 2 }) # the bytes get double-encoded ...
'b%27a%27=1&b%27b%27=2'
>>> urlencode({ str(b'a'): 1, str(b'b'): 2 }) # ... but the strs do okay ...
'a=1&b=2'
>>> str(b'a') # ... because of this
'a'
>>> urlencode({ u'asdf': u'1' }) # this works though
'asdf=1'
>>> urlencode({ bytes(str(b'a')): 1, bytes(str(b'b')): 2 }) # as expected
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/.../python-future/src/future/types/newbytes.py", line 85, in __new__
    raise TypeError('unicode string argument without an encoding')
TypeError: unicode string argument without an encoding
>>> urlencode({ str(bytes(b'a')): 1, str(bytes(b'b')): 2 }) # weird, but same as Py 3
'b%27b%27=2&b%27a%27=1'
>>> str(bytes(b'a')) # this explains the immediately previous output
"b'a'"

Now consider:

Python 3.4.3 (default, Sep  3 2015, 05:14:09)
[GCC 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> from urllib.parse import urlencode
>>> urlencode({ bytes(b'a'): 1, bytes(b'b'): 2 }) # now the bytes work ...
'b=2&a=1'
>>> urlencode({ str(b'a'): 1, str(b'b'): 2 }) # ... but the strs don't ...
'b%27b%27=2&b%27a%27=1'
>>> str(b'a') # ... because of this
"b'a'"
>>> urlencode({ u'asdf': u'1' }) # this works though
'asdf=1'
>>> urlencode({ bytes(str(b'a')): 1, bytes(str(b'b')): 2 }) # yup
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: string argument without an encoding
>>> urlencode({ str(bytes(b'a')): 1, str(bytes(b'b')): 2 }) # same as in Py 2
'b%27b%27=2&b%27a%27=1'
>>> str(bytes(b'a')) # same as in Py 2
"b'a'"

The way to fix one of those inconsistencies is to change newstr's constructor to treat Python 2's native strs as bytes, which probably means more checking at future/types/newstr.py, line 95, but this might break other things.

[posita]

By the way, I'm not sure if anyone cares, but it looks like there are some debug-by-print(f) statements being output when running tests:

% python setup.py \test -q
running test
running egg_info
writing src/future.egg-info/PKG-INFO
writing top-level names to src/future.egg-info/top_level.txt
writing dependency_links to src/future.egg-info/dependency_links.txt
writing entry points to src/future.egg-info/entry_points.txt
reading manifest template 'MANIFEST.in'
warning: no files found matching '*.au' under directory 'tests'
warning: no files found matching '*.gif' under directory 'tests'
warning: no files found matching '*.txt' under directory 'tests'
writing manifest file 'src/future.egg-info/SOURCES.txt'
running build_ext
............x....x.......x.....................................s..........x.........xxx.....x......E.s..x......x........sAdding an item
..x....x..................................xx....xx.x...........x......x..........x.x..xx.xsss.....s...x.x.x..x.......s..x....x...x.....x.....x.RefactoringTool: Skipping implicit fixer: idioms
RefactoringTool: Skipping implicit fixer: ws_comma
RefactoringTool: No files need to be modified.
..............................................................................................................................................s................................x.......x.x....object
xx.......x....................................................x...x.....<super: <class 'Singleton'>, <Singleton object>>
............x....xxRefactoringTool: No files need to be modified.
..x...x.......................Creating /var/folders/vx/pghvp8k129s2xwlx3g43t6ym0000gn/T/tmpraUC_F/test_imports_future_stdlib.py ...
Requests doesn't seem to be available. Skipping requests test ...
.Creating /var/folders/vx/pghvp8k129s2xwlx3g43t6ym0000gn/T/tmpr5nNmx/test_imports_future_stdlib.py ...
Requests doesn't seem to be available. Skipping requests test ...
...........x..x./.../lib/python2.7/tempfile.pyc
/.../lib/python2.7/copy.pyc
/.../lib/python2.7/textwrap.pyc
....Import succeeded!

.......x...x..s.................x.....s............................s..xxsss.........x...x.x...x.......s......................................................................................................................................s.......................................................s.xxxxsxxs......xx......s...s.........................................................................................................................s...........................Succeeded!
.Succeeded!
.Succeeded!
.Succeeded!
.Succeeded!
.Succeeded!
.Hello from a Python 2-style print statement!
Succeeded!
.Succeeded!
.
======================================================================
ERROR: test_open_default_encoding (test_future.test_builtins.BuiltinTest)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/.../python-future/tests/test_future/test_builtins.py", line 1217, in test_open_default_encoding
    self.write_testfile()
  File "/.../python-future/tests/test_future/test_builtins.py", line 1182, in write_testfile
    fp = open(TESTFN, 'w')
LookupError: unknown encoding:

----------------------------------------------------------------------
Ran 1075 tests in 45.515s

FAILED (errors=1, skipped=23, expected failures=65)

Regarding that failed test, I did not see that on Travis, but I do see it locally when running on OS X. It was present both before and after my changes.

[posita]

Closing this in light of #171 (comment) et seq.