type name in various spots.

the result of a call to PyObject_Repr() into the string. This makes
it possible to simplify many repr implementations.

PyUnicode_FromFormat() uses two steps to create the final string: A first
pass through the format string determines the size of the final string and
a second pass creates the string. To avoid calling PyObject_Repr() twice
for each %R specifier, PyObject_Repr() is called during the size
calculation step and the results are stored in an array (whose size is
determined at the start by counting %R specifiers).

This does nothing to fix the tests though...

hostname was passed.

PyString_Concat() and PyString_ConcatAndDel() (the name PyUnicode_Concat()
was already taken).

Change PyObject_Repr() to always return a unicode object.

Update all repr implementations to return unicode objects.

Add a function PyObject_ReprStr8() that calls PyObject_Repr() and converts
the result to an 8bit string.

Use PyObject_ReprStr8() where using PyObject_Repr() can't be done
straightforward.

to PyObject_CallFunction()) that take a char * (and a size
in the case of 'U#') and create a unicode object out of it.

Add functions PyUnicode_FromFormat() and PyUnicode_FromFormatV()
that work similar to PyString_FromFormat(), but create a unicode
object (also a %U format character has been added, that takes
a PyObject *, which must point to a unicode object).

Change the encoding and reason attributes of UnicodeEncodeError,
UnicodeDecodeError and UnicodeTranslateError to be unicode
objects.

for one more character anyway.

nullbyte can be copied.

This is undoubtedly insufficient and in some cases just as broken as before.

Changes to io.py, necessary to make this work:
- Redid io.StringIO as a TextIOWrapper on top of a BytesIO instance.
- Got rid of _MemoryIOMixin, folding it into BytesIO instead.
- The read() functions that take -1 to mean "eveything" now also take None.
- Added readline() support to BufferedIOBase. :-(

their encoded input/output.

XXX Should we just rip out this BSD Rune compatibility hack?

svn+ssh://pythondev@svn.python.org/python/branches/p3yk

........
  r55360 | guido.van.rossum | 2007-05-15 14:57:59 -0700 (Tue, 15 May 2007) | 2 lines

  obcheckin.
........
  r55361 | guido.van.rossum | 2007-05-15 14:59:18 -0700 (Tue, 15 May 2007) | 2 lines

  Get rid of strop module.
........
  r55367 | brett.cannon | 2007-05-15 21:06:28 -0700 (Tue, 15 May 2007) | 2 lines

  Remove the 'pure' module.
........
  r55369 | brett.cannon | 2007-05-15 21:07:31 -0700 (Tue, 15 May 2007) | 2 lines

  Remove the lib-old directory (already empty).
........
  r55370 | neal.norwitz | 2007-05-15 21:30:40 -0700 (Tue, 15 May 2007) | 1 line

  Get rid of a bunch more references to strop
........
  r55374 | brett.cannon | 2007-05-15 21:39:00 -0700 (Tue, 15 May 2007) | 2 lines

  Complete the removal of IRIX-specific modules.
........
  r55379 | brett.cannon | 2007-05-15 22:31:54 -0700 (Tue, 15 May 2007) | 2 lines

  Update removed IRIX modules based on what is gone from removing plat-irix6.
........
  r55388 | brett.cannon | 2007-05-16 14:34:52 -0700 (Wed, 16 May 2007) | 2 lines

  Clean up the docstring for the compiler resource.
........
  r55406 | brett.cannon | 2007-05-17 11:05:37 -0700 (Thu, 17 May 2007) | 2 lines

  Remove BaseException.message (deprecated in Python 2.6).
........

encoding) which makes the buffer mutable. Strings are encoded on the way in
and decoded on the way out.

Use io.StringIO in test_codecs.py.

Fix the base64_codec test in test_codecs.py.

bytes instead of strings.  We'll see how long that lasts.

Don't insist that float('1'*10000) raises an exception.

a fixed-length buffer of 256 bytes.

fails.