On Wed, 2009-03-11 at 17:27 +0100, Martijn Faassen wrote:
Hi there,
Stephan Richter wrote: [snip]
In my opinion going for an extra here just to avoid this is speculating a bit too much right now. Do we really have users that want to use zope.password and really don't want zope.component and zope.schema? If so, we'll hear from them when they speak up and *then* declare an extra or take some other action.
+1. I want more of our decisions to go into this direction. It is a sign that we turn the # of packages knob as well.
I agree with you in the case against extras.
It appears though that Dan has a concrete use case for using zope.password in a Pylons app where he isn't interested in zope.component, so I'm +1 on the extra in this case. We'll see whether this leads to difficulties. Luckily the zope.component and zope.schema libraries are typically around anyway so it doesn't make reasoning about the graph that much harder.
I remember that at the sprint we used to identify packages which are "always good". E.g. zope.interface is a declared no-brainer to add to your dependencies. The other two that keep popping up that we *might* wanna white-list are zope.schema and zope.component. In addition this would need documentation as well. Christian -- Christian Theune · ct@gocept.com gocept gmbh & co. kg · forsterstraße 29 · 06112 halle (saale) · germany http://gocept.com · tel +49 345 1229889 7 · fax +49 345 1229889 1 Zope and Plone consulting and development