On Mar 13, 2006, at 10:59 AM, Martijn Faassen wrote:

Thanks for looking into this Martijn. We have some internal versions of a sortable zc.table factory that take a different approach, and are not using the client-side part of the zc.table code much at all. If you're willing to look at it and consider integrating it into zc.table, one of us will send you the files privately. You up for that?

BTW, I think I listed in the release announcement that both of these packages could use some TLC. Thank you generally for considering giving them some!

I would love to see resourcelibrary developed in these directions, fwiw:

- pluggable insertion, to address one of the concerns in your original message. I don't remember exactly what Benji and I talked about. Maybe the resources would not be snippets but callable transformations. Maybe we could provide a factory for one kind of transformation which could be combined to effectively get the same behavior as our current transformation? A piece of code could request a transformation, which might depend on other transformations? Maybe this is too general.

- order-aware dependency insertion: IIRC, the algorithm for gathering the necessary dependencies doesn't honor order, which may be a problem in some cases. order-aware could get fragile, though--for instance, if one resource says it depends on A and then B, and another says it wants B then A, what to do? Maybe it's necessary to have one spelling for order-unaware dependencies, and another for order-aware dependencies; maybe that itself is insufficient...

- lighter-weight insertion into the publishing framework. No one (including myself) has followed up materially on Jim's old discussion of in-Zope publication pipelines. The request approach is heavy, and maybe a WSGI pipeline element is too heavy also. It would be nice to resolve this, though: replacing the request factory is a bit crazy.

It's not a priority for us to work on these things right now, but we open-sourced some of the less-polished packages in hopes that others would see their possibilities, as you did, and be intrigued. ;-)

Zope3-dev mailing list
Unsub: http://mail.zope.org/mailman/options/zope3-dev/archive%40mail-archive.com

Reply via email to