bpo-32285: Add unicodedata.is_normalized to check the current norma…

[maxbelanger]

Introduces unicodedata.is_normalized, which can check whether a unistr is in a given normal form.

This makes use of the internal helper (also called is_normalized) that can "quick check" normalization, but falls back on creating a normalized copy and comparing when necessary.

https://bugs.python.org/issue32285

[maxbelanger]

This was tested locally using ./python.exe -E -Wd -m test -u urlfetch -v test_normalization on macOS 10.13.

[vstinner]

I like the new function, but it should be documented in Doc/library/unicodedata.rst.

You may also add it to Doc/whatsnew/3.7.rst, in a "unicodedata" section of Improved Modules.

[vstinner]

Please add also a NEWS entry for the Changelog using the "blurb" tool:
https://devguide.python.org/committing/#what-s-new-and-news-entries

[maxbelanger]

@vstinner any other changes you'd like to see here? Just made a tiny signature change to ensure consistency with the rest of the module, otherwise I think this is good to go.

[maxbelanger]

@vstinner should I rebase this patch for 3.8?

[benjaminp]

There should be some negative cases, too. Make sure the MAYBE case is being exercised.

[maxbelanger]

Increased coverage + confirmed that this is exercising the MAYBE path.

[vstinner]

Maybe add also tests when it returns False. If the function always returns True, the test still pass ;-)

[benjaminp]

These could be bool.

[maxbelanger]

This is meant to conform to the existing implementation of is_normalized, which takes in ints. Could change is_normalized, but I preferred to avoid making changes outside the scope of my own.