different eol when writing to fp = codecs.open(.., 'w', 'cp866')
(windows, python-2.6.4)

def write(fp):
    fp.write("""\
a
""")

# eol=0d0a (windows, python-2.6.4)
with open('0d0a.tmp', 'w') as fp:
    write(fp)

# eol=0d0a (windows, python-2.6.4)
with codecs.open('0d0a-codecs.tmp', 'w') as fp:
    write(fp)

# --- BUG ---
# eol=0a (windows, python-2.6.4)
with codecs.open('0a-codecs.tmp', 'w', 'cp866') as fp:
    write(fp)

The docs say::
  Files are always opened in binary mode, even if no binary mode was
  specified. This is done to avoid data loss due to encodings using
  8-bit values. This means that no automatic conversion of '\n' is done
  on reading and writing.

But this does not match the code of codecs.open()::
    if encoding is not None and \
       'b' not in mode:
        # Force opening of the file in binary mode
        mode = mode + 'b'

When the encoding is None, the file is opened in text mode.
Marc-Andre, what do you think? is it a documentation bug instead?

Amaury Forgeot d'Arc wrote:
> 
> Amaury Forgeot d'Arc <amauryfa@gmail.com> added the comment:
> 
> The docs say::
>   Files are always opened in binary mode, even if no binary mode was
>   specified. This is done to avoid data loss due to encodings using
>   8-bit values. This means that no automatic conversion of '\n' is done
>   on reading and writing.
> 
> But this does not match the code of codecs.open()::
>     if encoding is not None and \
>        'b' not in mode:
>         # Force opening of the file in binary mode
>         mode = mode + 'b'
> 
> When the encoding is None, the file is opened in text mode.
> Marc-Andre, what do you think? is it a documentation bug instead?

Agreed, it's a documentation bug. If no encoding is specified,
codecs.open() works just like the standard open() (except for
the default value of mode, but that's documented).

The idea was to provide a drop-in replacement for open() - with
the added encoding parameter support.

Perhaps the default mode should have been 'r' to match the
regular open() - I guess it's too late to change that, though.

The docs still read the same.  Would someone in the know like to propose a doc patch please.