DOC: Add note about unicode layout

[sinhrks]

Follow-up for #11102. Added note because doc output may look incorrect depending on font.

[jreback]

hmm maybe better to put this in a code block which has a mono spaced font?

[sinhrks]

OK, let me check code block results.

[jreback]

@sinhrks ?

[jreback]

@sinhrks ?

[sinhrks]

I may misunderstand something, but code block doesn't use monospaced font always? I tried both :: and .. code-block::, but the results look same.

Markdown

::

    >>> df = pd.DataFrame({'a': ['xxx', u'¡¡'], 'b': ['yyy', u'¡¡']})
    >>> df
         a    b
    0  xxx  yyy
    1   ¡¡   ¡¡

.. code-block:: python

   >>> df = pd.DataFrame({'a': ['xxx', u'¡¡'], 'b': ['yyy', u'¡¡']})
   >>> df
        a    b
   0  xxx  yyy
   1   ¡¡   ¡¡

Output

Do you know any way to force all characters to be monospaced? Alternative idea is attach the screenshot of the console...

CC: @jorisvandenbossche

[jorisvandenbossche]

Hmm, don't think I can help here. So the 'easy' solution is indeed to use a picture that shows the effect (I think that is better than the note in any case, as the user can then visually see what is meant)

Some possible things: the output during the doc built is always decoded to utf-8 in a hack in the ipython directive (https://github.com/pydata/pandas/blob/master/doc/sphinxext/ipython_sphinxext/ipython_directive.py#L255) -> your example above does not use the ipython directive
It happens with the ambiguous width characters? And you needed to change something (the previous PR) to ensure the console output shows correcty? So maybe a similar issue is the case here with the sphinx directive

[sinhrks]

@jorisvandenbossche Thanks to confirm. The screen capture using ipython directive is attached on the bottom of #11102. Affected characters seems to depend on environment. At a quick glance, CSS doesn't specify monospaced font for highlight blocks.

As a workaround, updated to attach screen captures of each output as below.

[jreback]

thanks @sinhrks