[Zope-dev] [Checkins] SVN: zope.app.testing/trunk/ Switch doctests to use the stdlib ``doctest`` module.

[Adam GROSZER]

Hello Tres,

Thursday, June 3, 2010, 5:29:41 PM, you wrote:

TS> -----BEGIN PGP SIGNED MESSAGE-----
TS> Hash: SHA1

TS> Adam GROSZER wrote:

>>> Log message for revision 110723:
>>>  Switch doctests to use the stdlib ``doctest`` module.
>>>
>>>  o Disuse than the deprecated ``zope.testing.doctest`` variant.
>>>
>>>  Remove a duplicated import.
>>>
>>>
>>> Changed:
>>>  U   zope.app.testing/trunk/CHANGES.txt
>>>  U   zope.app.testing/trunk/src/zope/app/testing/functional.py
>>>  U   zope.app.testing/trunk/src/zope/app/testing/tests.py
>> 
>> The problem with this is that it just breaks tests that have accented
>> unicode characters.
>> 
>> zope.app.testing.functional.FunctionalDocFileSuite breaks, while
>> zope.testing.doctest.DocFileSuite works.
>> I don't remember what was the final decision on deprecating
>> zope.testing.doctest, but this definitely s*cks.
>> 
>> Thoughts?

TS> What tests are breaking?

Our apps's tests.
Excerpt is:

# -*- coding: UTF-8 -*-
<snip>
  >>> mail = mailer.prepareEmail(u"Some subject",
  ...     u"Some body éáÉÁ",
  ...     u"from at me.com",
  ...     u"to at you.com")
<snip>

With the stdlib doctest prepareEmail gets
u'Some body \xc3\xa9\xc3\xa1\xc3\x89\xc3\x81'
Note, it's unicode() and utf-8 encoded at the same time.
I guess stdlib doctest does not care about # -*- coding: UTF-8 -*-

-- 
Best regards,
 Adam GROSZER                            mailto:agroszer at gmail.com
--
Quote of the day:
A physicist is an atom's way of knowing about atoms. -- George Wald