Issue 30721: Show expected input for right shift operator usage in custom "print" error message

While working on issue: http://bugs.python.org/issue30597 to enhance the custom error message by showing expected output in Python3 syntax when someone uses Python2 syntax, a PR was raised: https://github.com/python/cpython/pull/2009, where we just handled the case for print with soft-space and excessive white-space.

In the implementation discussion, an issue was raised to handle the case with right shift operator by Nick as in here:
http://bugs.python.org/issue30597#msg295484

Nick suggested here about the possible patch: https://github.com/python/cpython/pull/2009#issuecomment-307539241

The specific error in question here is the one where Python 3 reads the old Python 2 stream redirection syntax as a combination of the right shift operator and tuple creation:

```
>>> print >> sys.stderr, "message"
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: unsupported operand type(s) for >>: 'builtin_function_or_method' and '_io.TextIOWrapper'
```

Searching for that error message indicates that people are hitting it reasonably frequently, so we may be able to save them a trip to Google by adding a custom 'Did you mean "print(<output>, file={:100!r})'.format(rhs)"? message when the right-shift operand dispatch is about to report the default "no implementation" type error, and the LHS is the print builtin.

The test added in the PR passes on linux (Travis) and Mac (op's machine), fails on Windows (Appveyor).  The patch itself modifies abstract.c.  Could some Windows expert take a look?

test_string_with_stream_redirection (test.test_print.TestPy2MigrationHint) ... FAIL
FAIL: test_string_with_stream_redirection (test.test_print.TestPy2MigrationHint)

Traceback (most recent call last):
File "C:\projects\cpython\lib\test\test_print.py", line 165, in test_string_with_stream_redirection
'file=<output_stream>)', str(context.exception))
AssertionError: 'Did you mean "print(, file=<output_stream>)' not found in "unsupported operand type(s) for >>: 'builtin_function_or_method' and '_io.StringIO'"

The "op_slot == 96" looks suspicious - how is this value actually supposed to be calculated?

Checking how we do it elsewhere, `NB_SLOT(nb_rshift)` looks like the right replacement.

That's a compiler-dependent struct field offset calculation, so a discrepancy there could easily be the cause of a Windows-only failure.

New changeset 5e2eb35bbed3e84079165e576cdb50ef36e13493 by Serhiy Storchaka (Sanyam Khurana) in branch 'master':
bpo-30721: Show correct syntax hint in Py3 when using Py2 redirection syntax (#2345)
https://github.com/python/cpython/commit/5e2eb35bbed3e84079165e576cdb50ef36e13493

Checking the merged implementation locally, I belatedly noticed that we forgot the trailing question mark on the question:

```
>>> import sys
>>> print >> sys.stderr
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: unsupported operand type(s) for >>: 'builtin_function_or_method' and '_io.TextIOWrapper'. Did you mean "print(<message>, file=<output_stream>)"
```

So putting this back to "needs patch" (no NEWS entry needed, just the missing question marks in the patch and the tests).

Ah, sorry for that. I'll just fix that right away in a few mins :)

Note that I filed a separate issue to ask Ned about potentially backporting this to 3.6: https://bugs.python.org/issue31232

New changeset a7c449b8c08933deabcf329fb74ed1336f6db34f by Nick Coghlan (Sanyam Khurana) in branch 'master':
bpo-30721: Add missing '?' to new error message (GH-3131)
https://github.com/python/cpython/commit/a7c449b8c08933deabcf329fb74ed1336f6db34f

And done - thanks folks!

New changeset 1a05e87ec75436d818f05a5dabcecaea67334cbd by Nick Coghlan in branch '3.6':
[3.6] bpo-31232: Backport custom print rshift message (GH-3155)
https://github.com/python/cpython/commit/1a05e87ec75436d818f05a5dabcecaea67334cbd

stage: needs patch -> resolved

messages: + msg299128
stage: patch review