[XML-SIG] Finding _xmlplus in Python 2.3a2

Martin v. L÷wis martin@v.loewis.de
04 Mar 2003 08:10:12 +0100


John J Lee <jjl@pobox.com> writes:

> > I bring up the bug reporting simply because I don't recall anyone reporting
> > that PyXML broke their Python installation until now.  The cases you mention
> > are abstract, but without concrete details, I cannot reply otherwise.
> [...]
> 
> That argument makes sense only when applied to a feature to be introduced
> in the future, in which case we can't know what effect it will have
> without considering concrete details.  This discussion is about a feature
> already introduced (the _xmlplus hack), which we *know* causes problems
> (unless you don't believe your users).  

I think this is the case: We don't believe the users. I know Martijn
has coworkers who are confused by this (in the sense that they are
unaware of the entire issue, or have the wrong mental model). This, in
itself, is no problem: Everybody of us certainly is unaware of some
things, and has wrong mental models of other things.

The question is whether this has created real problems, in the sense
of wasted time and energy. Martijn has reported that it did, in the
case of using undocumented API for constructing nodes. I find this
example inconvincing, as the next Python release will make the same
change, and thus break this code even without PyXML being installed.

I haven't heard any other practical examples of real problems caused
by the current setup.

Regards,
Martin