Expand unicodedata module documentation

[ezio.melotti]

http://bugs.python.org/review/10665/diff/1830/3423
File Doc/library/unicodedata.rst (left):

http://bugs.python.org/review/10665/diff/1830/3423#oldcode66
Doc/library/unicodedata.rst:66: 
You can also use the simplified syntax for the tables.

http://bugs.python.org/review/10665/diff/1830/3423
File Doc/library/unicodedata.rst (right):

http://bugs.python.org/review/10665/diff/1830/3423#newcode64
Doc/library/unicodedata.rst:64: -1
This example is not clear IMHO.  Do you want to show how to use the default
value or just that the "superscript nine" char has no decimal value?

http://bugs.python.org/review/10665/diff/1830/3423#newcode77
Doc/library/unicodedata.rst:77: -1
See previous comment.

http://bugs.python.org/review/10665/diff/1830/3423#newcode103
Doc/library/unicodedata.rst:103: | **General Categories**                       
                           |
Headers spanning multiple columns are not supported by `make latex`.

http://bugs.python.org/review/10665/diff/1830/3423#newcode166
Doc/library/unicodedata.rst:166:
+----+-------------+------------------+------------------------------------+
All the characters not in range 00-FF should be replaced, either by their name
(possibly suggesting the use of lookup() to see how they look like) or with
their codepoint.
Another option is to remove the example column altogether

http://bugs.python.org/review/10665/diff/1830/3423#newcode170
Doc/library/unicodedata.rst:170: .. literalinclude:: ../includes/unistat.py
You forgot to include this in the latest patch.

http://bugs.python.org/review/10665/diff/1830/3423#newcode174
Doc/library/unicodedata.rst:174: Counter({'C': 1004868, 'L': 100520, 'S': 5508,
'M': 1498, 'N': 1100, 'P': 598, 'Z': 20})
How will this be updated when the unicode db is updated?

http://bugs.python.org/review/10665/diff/1830/3423#newcode231
Doc/library/unicodedata.rst:231: ====      ============
There are also trailing spaces here and in another places that should be
removed.

http://bugs.python.org/review/10665/diff/1830/3423#newcode250
Doc/library/unicodedata.rst:250: '0065 0300'
Maybe explain shortly what decomposition is and what is it useful for?
Showing the name of the two codepoints in the example is also a good idea.

[belopolsky_users.sourceforge.net]

http://bugs.python.org/review/10665/diff/1830/3423
File Doc/library/unicodedata.rst (right):

http://bugs.python.org/review/10665/diff/1830/3423#newcode64
Doc/library/unicodedata.rst:64: -1
On 2011/03/26 14:31:35, ezio.melotti wrote:
> This example is not clear IMHO.  Do you want to show how to use the default
> value or just that the "superscript nine" char has no decimal value?

Both, but mostly the latter.  I did not want to show an example with an
exception traceback.  Maybe we can make this clearer by adding some explanatory
prose?  Can you suggest a specific change?

http://bugs.python.org/review/10665/diff/1830/3423#newcode77
Doc/library/unicodedata.rst:77: -1
On 2011/03/26 14:31:35, ezio.melotti wrote:
> See previous comment.
Is  unicodedata.name() example above OK?  I would like to use consistent format
throughout.

http://bugs.python.org/review/10665/diff/1830/3423#newcode103
Doc/library/unicodedata.rst:103: | **General Categories**                       
                           |
On 2011/03/26 14:31:35, ezio.melotti wrote:
> Headers spanning multiple columns are not supported by `make latex`.

Hmm.  I am sure I've seen these used elsewhere in python docs.  How badly does
this break LaTeX?

http://bugs.python.org/review/10665/diff/1830/3423#newcode174
Doc/library/unicodedata.rst:174: Counter({'C': 1004868, 'L': 100520, 'S': 5508,
'M': 1498, 'N': 1100, 'P': 598, 'Z': 20})
On 2011/03/26 14:31:35, ezio.melotti wrote:
> How will this be updated when the unicode db is updated?

Hopefully by the time UCD is updated we will start validating the tests. 
Otherwise, someone will report a doc-bug and we'll fix it.  I don't see this as
a critical issue if the counts are slightly off.

[ezio.melotti]

http://bugs.python.org/review/10665/diff/1830/3423
File Doc/library/unicodedata.rst (right):

http://bugs.python.org/review/10665/diff/1830/3423#newcode64
Doc/library/unicodedata.rst:64: -1
The confusing parts are:
 1) the char *seems* to have a decimal value (9), but it doesn't;
 2) the default value is a decimal, but it's "wrong".
Maybe something like unicodedata.decimal('a', None) might be better.

http://bugs.python.org/review/10665/diff/1830/3423#newcode77
Doc/library/unicodedata.rst:77: -1
Using None or 'no digit' or something similar is OK.
FWIW I've seen and used '(undefined)' in these cases, I'm not sure if it's
something commonly used though.

http://bugs.python.org/review/10665/diff/1830/3423#newcode103
Doc/library/unicodedata.rst:103: | **General Categories**                       
                           |
This badly:
Markup is unsupported in LaTeX:
library/unicodedata:: column or row spanning cells are not yet implemented.
make: *** [build] Error 1

i.e. it doesn't even pass the `make latex` step.