Issue 23583: IDLE: printing unicode subclasses broken (again)

This is a regression or recurrence of issue #19481. 

To reproduce, create a subclass of unicode and try and print an instance of that class:

class Foo(unicode): pass

print Foo()

results in a traceback:

Traceback (most recent call last):
  File "<pyshell#4>", line 1, in <module>
    print Foo()
  File "/usr/local/Cellar/python/2.7.9/Frameworks/Python.framework/Versions/2.7/lib/python2.7/idlelib/PyShell.py", line 1353, in write
    s = unicode.__getslice__(s, None, None)
TypeError: an integer is required

because unicode.__getslice__ does not accept None for the start and end indices.

Proposed fix, replace line 1352-1353 in PseudoOutputFile.write():

if isinstance(s, unicode):
    s = unicode.__getslice__(s, None, None)

with

if isinstance(s, unicode):
    s = unicode.__getslice__(s, 0, len(s))

I can propose two ways to fix this:

    s = unicode.__getslice__(s, 0, unicode.__len__(s))

and

    s = unicode.__getitem__(s, slice(None))

Just len() doesn't work when the subclass overrides __len__.

I like the unicode.__getitem__(s, slice(None)) approach, it has the advantage of not having to rely on len(s).

What about just 's = unicode(s)'?  The doc says

"object.__unicode__(self)

    Called to implement unicode() built-in; should return a Unicode object. When this method is not defined, string conversion is attempted, and the result of string conversion is converted to Unicode using the system default encoding."

but the latter part seems not to be true for unicode subclasses.

>>> class F(unicode):
	def __str__(self): raise TypeError()

>>> f = F()
>>> u = unicode(f)
>>> type(u)
<type 'unicode'>

See example in msg205740. The same works with overridden __unicode__.

I'm writing tests now.

I don't see what you have in mind.  Translating my example

class S(str):
    def __str__(self):
        return 'S: ' + str.__str__(self)

to

class U(unicode):
    def __unicode__(self):
        return u'U: ' + unicode.__unicode__(self)

does not work because there is no unicode.__unicode__.

> does not work because there is no unicode.__unicode__.

Yes, this is why unicode.__getslice__ or unicode.__getitem__ should be used.

Here is a patch for 2.7 that fixes this issue and adds tests for PseudoOutputFile and PseudoInputFile. Tests should be ported also to 3.x.

New changeset cd9e4f73b5ce by Serhiy Storchaka in branch '2.7':
Issue #23583: Fixed writing unicode to standard output stream in IDLE.
https://hg.python.org/cpython/rev/cd9e4f73b5ce

New changeset 0c72cdf3ff22 by Serhiy Storchaka in branch '2.7':
Issue #23583: Fixed writing unicode to standard output stream in IDLE.
https://hg.python.org/cpython/rev/0c72cdf3ff22

New changeset 2c680d7d8ca6 by Serhiy Storchaka in branch '3.4':
Issue #23583: Added tests for standard IO streams in IDLE.
https://hg.python.org/cpython/rev/2c680d7d8ca6

New changeset f860915e5705 by Serhiy Storchaka in branch 'default':
Issue #23583: Added tests for standard IO streams in IDLE.
https://hg.python.org/cpython/rev/f860915e5705

assignee: serhiy.storchaka
keywords: + patch
stage: needs patch -> patch review