Re: [z3-five] RFC: The future of Products.Five
--On 30. Oktober 2006 00:01:33 +0100 Philipp von Weitershausen <philipp@weitershausen.de> wrote:
* It is unclear to me at this point what Zope 2's egg story will be. I *hope* that 2.11 will get the same egg story as the Zope 3.4 that ships with it does, though noone has talked about eggifying Zope 2 yet. We should probably do that.
What are our benefits from a Z2 Eggification. Eggification is basically about packaging and distribution of components with little dependencies in order to re-use them in other contexts. However most components of Zope 2 are heavily dependent on others. Or are we talking about using Zope 3 eggs within Zope 2 as part of the Zope 2 distribution process? Andreas
Andreas Jung wrote:
--On 30. Oktober 2006 00:01:33 +0100 Philipp von Weitershausen <philipp@weitershausen.de> wrote:
* It is unclear to me at this point what Zope 2's egg story will be. I *hope* that 2.11 will get the same egg story as the Zope 3.4 that ships with it does, though noone has talked about eggifying Zope 2 yet. We should probably do that.
What are our benefits from a Z2 Eggification. Eggification is basically about packaging and distribution of components with little dependencies in order to re-use them in other contexts. However most components of Zope 2 are heavily dependent on others. Or are we talking about using Zope 3 eggs within Zope 2 as part of the Zope 2 distribution process?
I would argue a bit of both. Zope 2 can be a giant egg without much difficultly enabling use of easy_install and dependency management in relation to that egg(think of how products depend on a version or point version of zope2). As an egg friendly installable, it can then be used in egg recipes etc. it also lets us start teasing parts of zope2 out of the monolith of zope2 their own eggs(ZTC comes to mind) so they can evolve at a rate not controlled by the zope2 release process. Since it is possible to match dependencies, releases can be built from egg recipes dynamically rather than cumbersome bundle linking. I'm guessing this would make release management ultimately easier and more consistent; a recipe would simply match all stable and compatible eggs to the zope version that needed them. -w
Andreas Jung wrote:
What are our benefits from a Z2 Eggification. Eggification is basically about packaging and distribution of components with little dependencies in order to re-use them in other contexts. However most components of Zope 2 are heavily dependent on others.
Yes, that's true. But eggs don't necessarily have to have few dependencies ;). Eggs just have well-defined dependencies. And like Whit said, there might be a desire to have certain Zope 2 components have a faster life cycle than others, e.g. ZopeTestCase in Whit's case.
Or are we talking about using Zope 3 eggs within Zope 2 as part of the Zope 2 distribution process?
That too. As Zope 3.4 will hopefully be embracing eggs for its packaging story, it would be a shame if we wouldn't get that story somehow in Zope 2.11 at the same time.
--On 31. Oktober 2006 02:56:39 +0100 Philipp von Weitershausen <philipp@weitershausen.de> wrote:
Andreas Jung wrote:
What are our benefits from a Z2 Eggification. Eggification is basically about packaging and distribution of components with little dependencies in order to re-use them in other contexts. However most components of Zope 2 are heavily dependent on others.
Yes, that's true. But eggs don't necessarily have to have few dependencies ;). Eggs just have well-defined dependencies. And like Whit said, there might be a desire to have certain Zope 2 components have a faster life cycle than others, e.g. ZopeTestCase in Whit's case.
Yeah, it would make sense for particular components but not for all components just for the sake of eggs. -aj
Andreas Jung wrote:
--On 31. Oktober 2006 02:56:39 +0100 Philipp von Weitershausen <philipp@weitershausen.de> wrote:
Andreas Jung wrote:
What are our benefits from a Z2 Eggification. Eggification is basically about packaging and distribution of components with little dependencies in order to re-use them in other contexts. However most components of Zope 2 are heavily dependent on others.
Yes, that's true. But eggs don't necessarily have to have few dependencies ;). Eggs just have well-defined dependencies. And like Whit said, there might be a desire to have certain Zope 2 components have a faster life cycle than others, e.g. ZopeTestCase in Whit's case.
Yeah, it would make sense for particular components but not for all components just for the sake of eggs.
-aj
I would expect things to move as needed rather than just for the sake of it. -w
Andreas Jung wrote:
Andreas Jung wrote:
What are our benefits from a Z2 Eggification. Eggification is basically about packaging and distribution of components with little dependencies in order to re-use them in other contexts. However most components of Zope 2 are heavily dependent on others.
Yes, that's true. But eggs don't necessarily have to have few dependencies ;). Eggs just have well-defined dependencies. And like Whit said, there might be a desire to have certain Zope 2 components have a faster life cycle than others, e.g. ZopeTestCase in Whit's case.
Yeah, it would make sense for particular components but not for all components just for the sake of eggs.
Well, it may be that we decide to embrace eggs for our packaging story, at least in Zope 3 (as a replacement for zpkg). At that point we should ask ourselves if it wouldn't be a good idea to do the same for Zope 2. In Zope 2, eggs could really help with the add-on story (as a more general replacement for Products).
participants (3)
-
Andreas Jung -
Philipp von Weitershausen -
whit