[Catalog-sig] egg_info in PyPI

"Martin v. Löwis" martin at v.loewis.de
Mon Sep 20 23:49:36 CEST 2010

Am 20.09.2010 23:20, schrieb P.J. Eby:
> At 09:26 PM 9/20/2010 +0200, Martin v. Löwis wrote:
>> That's how I would have implemented it. Alas, this very
>> idea was just shot down.
> Hm.  When the issue was adding comments to PyPI, you mostly ignored the
> strong opposition from people who didn't want to use the feature, in
> order to support those who did.

Back then, I firmly believed that the feature was absolutely important
to have. Now, I admit that people can get the information if they
download all sdists. If there is demand for it, someone could also
provide a service that extracts all sdists, and provides the files
off-site, to reduce bandwidth usage for end-users.

So unlike with the comment system, the world is not going to end without
the feature.

> While I'm only +0 (at most) on adding egg_info, I'm not sure why the
> objections of people who won't be using the feature should carry more
> weight in the present discussion than in the previous one.  ;-)

If I understand correctly, it's more than that, though (and I would
easily continue ignoring people who object to a feature just because
they are not going to use it).

I feel that the core of the objection (from Tarek in particular) is
that it is viewed as directly competing to PEP 345. While I personally
think that this objection is flawed (i.e. exposing the data will not
have any significant impact on the PEP 345 adoption rate), I also
value Tarek's recent contributions to PyPI, and accept his doubts
(even though I don't share them).


More information about the Catalog-SIG mailing list