This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author eli.bendersky
Recipients Arfrever, asvetlov, danielsh, eli.bendersky, ezio.melotti, flox, ncoghlan, python-dev, serhiy.storchaka, tshepang
Date 2013-02-27.04:40:58
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <>
In-reply-to <>
On Tue, Feb 26, 2013 at 8:08 PM, Ezio Melotti <>wrote:

> Ezio Melotti added the comment:
> > The next logical step is to make all test classes in test_xml_etree
> > accept the ET module in some way and store it, using it to get classes
> > & function. I.e. no more global "ET" and "pyET" at all.
> The idiom suggested by PEP 399 has the two modules (cmod and pymod) as
> globals, and then simply sets them as class attributes.  Is there any
> reason why this should be avoided in this case?

I'm just not sure how the globals help except for saving 2-3 lines of code.
I do see how they can cause problems because even when I just want to run
the pure Python code, the C module gets imported. Why should it be? I
really don't want it to, I want to isolate things as much as possible
(after all, testing the pure Python module actually tests a scenario where
there is no C module). Pickle is one concrete place that can cause problems
with this.

We talked about related things in Issue #15083 and AFAIR Eric's and Brett's
proposals move these modules away from the global namespace.
Date User Action Args
2013-02-27 04:40:59eli.benderskysetrecipients: + eli.bendersky, ncoghlan, ezio.melotti, Arfrever, asvetlov, flox, tshepang, python-dev, serhiy.storchaka, danielsh
2013-02-27 04:40:59eli.benderskylinkissue15083 messages
2013-02-27 04:40:58eli.benderskycreate