[Python-Dev] please back out changeset f903cf864191 before alpha-2

Stefan Behnel stefan_ml at behnel.de
Sat Aug 24 07:32:34 CEST 2013

Antoine Pitrou, 24.08.2013 01:26:
> On Sat, 24 Aug 2013 00:57:48 +0200
> Stefan Behnel wrote:
>> ticket 17741 has introduced a new feature in the xml.etree.ElementTree
>> module that was added without any major review.
>> http://bugs.python.org/issue17741
> As I've already indicated on the tracker, I'm completely saturated
> with Stefan's qualms about a minor API addition and I'm not willing to
> process anymore of them. Hence I won't respond to the bulk of his
> e-mail.
> But I still want to clarify that claiming that the feature was "added
> without any major review" is outrageous and manipulative. Perhaps
> Stefan thinks that an ElementTree code review not by him (but, for
> example, by Eli, who currently maintains ElementTree) is not "major".

The reason why I'm saying this is that the way the change came in is rather
- unorthodox. As Antoine noted in the ticket, he proposed the change on the
tulip mailing list. That is a completely wrong place to discuss a new XML
API, as can be seen from the replies.


Specifically, no-one noticed the major overlap with the existing API and
functionality at that point, nor the contradictions between the existing
API and the new one.

In the ticket, Eli stated that he didn't have time for the review ATM, and
then, two days later, commented that the patch looks good. To me (and I'm
really only interpreting here), this indicates that the review was mostly
at the patch level. Note that he didn't comment on the API overlap either,
so my guess is that he just didn't notice it. In my experience, reviewing
design and thinking about alternatives takes more time than that,
especially when you're "swamped", as he put it. I mean, it took *me* almost
a day to dig into the implications and into the patch (as can be seen by my
incremental comments), and I have the background of having written a
complete implementation of that library myself.

So, to put it more nicely, I think this feature was added without the
amount of review that it needs, and now that I've given it that review, I'm
asking for removal of the feature and a proper redesign that fits into the
existing library.


More information about the Python-Dev mailing list