Implement incremental decoder for cp65001

[vstinner]

BPO	20574
Nosy	@loewis, @vstinner, @larryhastings, @serhiy-storchaka
Files	incremental_cp_utf8.patch

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2015-03-18.13:22:37.172>
created_at = <Date 2014-02-09.13:18:25.151>
labels = ['type-feature']
title = 'Implement incremental decoder for cp65001'
updated_at = <Date 2015-03-18.13:22:37.171>
user = 'https://github.com/vstinner'

bugs.python.org fields:

activity = <Date 2015-03-18.13:22:37.171>
actor = 'vstinner'
assignee = 'none'
closed = True
closed_date = <Date 2015-03-18.13:22:37.172>
closer = 'vstinner'
components = []
creation = <Date 2014-02-09.13:18:25.151>
creator = 'vstinner'
dependencies = []
files = ['34008']
hgrepos = []
issue_num = 20574
keywords = ['patch']
message_count = 9.0
messages = ['210759', '210764', '210783', '213905', '213906', '213907', '213908', '213923', '213926']
nosy_count = 5.0
nosy_names = ['loewis', 'vstinner', 'larry', 'python-dev', 'serhiy.storchaka']
pr_nums = []
priority = 'normal'
resolution = 'fixed'
stage = None
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue20574'
versions = ['Python 3.5']

[vstinner]

(Follow up of issue bpo-20538 and bpo-20571.) Attached patch implements incremental decoders for multibyte code pages (on Windows), especially for CP_UTF8 aka "cp65001" in Python.

Code pages 932, 936, 949, 950 and 1361 already have an incremental decoder since:
---
changeset: 38817:549c547700af
branch: legacy-trunk
user: Martin v. Löwis <martin@v.loewis.de>
date: Wed Jun 14 05:21:04 2006 +0000
files: Doc/api/concrete.tex Include/unicodeobject.h Lib/encodings/mbcs.py Misc/NEWS Modules/_codecsmodule.c Objects/unicodeobject.c
description:
Patch bpo-1455898: Incremental mode for "mbcs" codec.
---

Python currently uses IsDBCSLeadByteEx():
http://msdn.microsoft.com/en-us/library/windows/desktop/dd318667%28v=vs.85%29.aspx

And CharPrevA():
http://msdn.microsoft.com/en-us/library/windows/desktop/ms647471%28v=vs.85%29.aspx

But IsDBCSLeadByteEx() only supports code pages 932, 936, 949, 950 and 1361.

Python supports the code page 65001 (codec "cp65001") since Python 3.3. New tests on incremental decoders were added in Python 3.4: I addedd a skip for cp65001 since it was not supported (bpo-20571). This issue implements the incremental decoder and so removes the skip.

I prefer to wait for Python 3.5 (not rush for add this new feature after 3.4 beta 3). cp65001 is mostly used for output (sys.stdout/sys.stderr) on Windows, not for input.

[serhiy-storchaka]

Nice.

Could you please also add test_partial for CP65001 (if this will make sense)?

What is performance regression of this patch?

I considered this issue as a bug. And if performance regression is not too big, I think it can be applied to 3.3+. Otherwise a warning should be added that CP65001 doesn't not work with input text streams.

[vstinner]

It might be faster, or (more likely) has zero impact on performances.

[python-dev]

New changeset 08f9b881f78c by Victor Stinner in branch 'default':
Issue bpo-20574: Implement incremental decoder for cp65001 code
http://hg.python.org/cpython/rev/08f9b881f78c

[python-dev]

New changeset 85b87789f048 by Victor Stinner in branch 'default':
Issue bpo-20574: Add more tests for cp65001
http://hg.python.org/cpython/rev/85b87789f048

[vstinner]

Could you please also add test_partial for CP65001 (if this will make sense)?

I added CP65001Test which inherit from UTF8Test and so runs all UTF-8 tests on cp65001 codec. I'm surprised that the test pass.

[vstinner]

I don't feel the need to backport the new feature, so I'm closing the issue.

[python-dev]

New changeset f6794a0fb2b3 by Victor Stinner in branch 'default':
Issue bpo-20574: Remove duplicated test failing on Windows XP
http://hg.python.org/cpython/rev/f6794a0fb2b3

[vstinner]

I removed the test because there were two classes tesing the same codec and that tests were failing. I need to refactor tests, and so I reopen the issue.

http://buildbot.python.org/all/builders/x86%20XP-4%203.x/builds/10291/steps/test/logs/stdio

======================================================================
FAIL: test_lone_surrogates (test.test_codecs.CP65001Test)
----------------------------------------------------------------------

Traceback (most recent call last):
  File "D:\cygwin\home\db3l\buildarea\3.x.bolen-windows\build\lib\test\test_codecs.py", line 773, in test_lone_surrogates
    super().test_lone_surrogates()
  File "D:\cygwin\home\db3l\buildarea\3.x.bolen-windows\build\lib\test\test_codecs.py", line 349, in test_lone_surrogates
    self.assertRaises(UnicodeEncodeError, "\ud800".encode, self.encoding)
AssertionError: UnicodeEncodeError not raised by encode