bpo-36778: cp65001 encoding becomes an alias to utf_8

[vstinner]

https://bugs.python.org/issue36778

[vstinner]

I reproduced #13110 (comment) benchmark:

Mean +- std dev: [ref] 156 ns +- 3 ns -> [remove] 105 ns +- 3 ns: 1.48x faster (-32%)

[vstinner]

@methane: Are you ok to simply remove cp65001.py?

[paulmon]

I can verify that this change fixes the issue with test_startup_imports I found on Windows IoT Core ARM32 as expected.

Also, all test_site and test_codec tests pass on Windows IoT Core with a default codepage of 65001

[methane]

@methane: Are you ok to simply remove cp65001.py?

Yes.

[serhiy-storchaka]

Add the versionchanged directive.

I think it is better to remove this row add add cp65001 to the list of utf-8 aliases.

[vstinner]

I created PR #13240 for your doc change proposal.

[serhiy-storchaka]

Does utf-8 pass these tests?

[methane]

I confirmed utf-8 passes this test.

[vstinner]

Does utf-8 pass these tests?

Sorry, I forgot to specify before I merged my PR that yes: I tested on my Windows 10 and the test still passed. But CP65001Test is now redundant with UTF8Ttest.

I confirmed utf-8 passes this test.

Thanks for checking :-)

[taleinat]

Shouldn't this comment in CodePageTest in Lib/test/test_codecs.py also be removed? It no longer makes any sense.

class CodePageTest(unittest.TestCase):
    # CP_UTF8 is already tested by CP65001Test
    CP_UTF8 = 65001

[vstinner]

Shouldn't this comment in CodePageTest in Lib/test/test_codecs.py also be removed? It no longer makes any sense.

Right. I wrote PR #13807 to remove the comment.