On 03/15/2010 06:51 PM, Chris Calloway wrote:
On 3/15/2010 6:28 PM, Cristopher Ewing wrote:
A lot of the issues that jaroel opens seem to fall into this
territory. In fact, this whole process of doing svn checkins while
creating your code skeleton seem to point to a bunch of assumptions
that are hard-coded in paster. If that means that paster is a tool
that is too specific to a particular use case to fit as our base, I'm
not totally surprised. I think we've edged around the discussion of
how suitable paster is as a core for zopeskel for a while now. Let's
see where these issues lead us, and that's a discussion we can have
more of going forward.


I was going to suggest that would be one more reason to keep upstream
issues open; but stopped short thinking I'd said enough.

As the upstream open tickets accumulate, I'm sure they are going to
provide more evidence that we are just abusing the hell out of paster,
as we've discussed many times.

Having these open tickets might just eventually provide us with some
clue about what kind of tool we need to implement ZopeSkel properly.

The tickets will surely continue to accumulate. I was curious and looked at the code for this particular bug. It is definitely paster assuming that the template will be an egg and not a buildout when checking the code in.

Before the sprint, we went to Ian for commit access to paster in case
there was anything we felt we urgently needed to fix in order to get our
work done. Not only was he glad to grant us to that access, he kind of
begged us to take paster over. We decided not to get stuck with that, as
the larger issue seemed to be should we even be using paster in the long
term.

When we were sprinting after the Plone Conference in DC, Ian stopped by for a bit. He was looking for a maintainer then, and it seems he hasn't found one.

Paster has served us well for quite a while now. At Six Feet Up, we extend ZopeSkel to our particular way of doing things. While working on what I call SixieSkel, I have seen some of the limitations in PasteScript templates but have been able to work around them.

If we get tired of workarounds then we have to re-write something from scratch. Try and run pep8.py against any Paste* package and you will get glimpse into the future ;) There are some good ideas in there, but it needs a new vision.

Of course, having trac instead of Poi is just one more reason why we
want ZopeSkel to become a peer with Plone and Archetypes with its own
repository instead of a Collective product susceptible to drive-by code.

There is a lot of drive by code in the collective, this is for sure. ZopeSkel seems to be more of a mainstay than a lot of that code though. And now it has a mailing list and a maintainer. That puts it in the top 5% in the collective :)

Let's not forget that other communities are using paster as well. There are django and repoze.bfg folks using it for sure. Maybe we can all work towards the next generation set of templates.

Clayton
--
clay...@sixfeetup.com | +1 (317) 861-5948 x603
six feet up presents INDIGO : The Help Line for Plone
More info at http://sixfeetup.com/indigo or call +1 (866) 749-3338


_______________________________________________
ZopeSkel mailing list
ZopeSkel@lists.plone.org
http://lists.plone.org/mailman/listinfo/zopeskel

Reply via email to