zope.app.module, zope.app.interface and zodb in KGS?
zope.app.module provides support for persistemt modules. zope.app.interface provides support for persistent interfaces. Both rely on a highly experimental zodbcode. I don't have a problem with these being projects, but I don't think they should be in the KGS. I don't think we should commit to supporting them, at least not as part of the KGS. Does anyone disagree? Jim -- Jim Fulton Zope Corporation
On Thu, Jul 2, 2009 at 12:30 PM, Jim Fulton<jim@zope.com> wrote:
zope.app.module provides support for persistemt modules. zope.app.interface provides support for persistent interfaces. Both rely on a highly experimental zodbcode. I don't have a problem with these being projects, but I don't think they should be in the KGS. I don't think we should commit to supporting them, at least not as part of the KGS. Does anyone disagree?
+10 on removing those from the KGS. Right now both zope.app.zapi and zope.app.zcmlfiles still have a hard dependency on zope.app.interface though. Hanno
On Jul 2, 2009, at 7:58 AM, Hanno Schlichting wrote:
On Thu, Jul 2, 2009 at 12:30 PM, Jim Fulton<jim@zope.com> wrote:
zope.app.module provides support for persistemt modules. zope.app.interface provides support for persistent interfaces. Both rely on a highly experimental zodbcode. I don't have a problem with these being projects, but I don't think they should be in the KGS. I don't think we should commit to supporting them, at least not as part of the KGS. Does anyone disagree?
+10 on removing those from the KGS.
Right now both zope.app.zapi
This was deprecated a long time ago. I think we should remove it from the kgs. Is anyone using it?
and zope.app.zcmlfiles still have a hard dependency on zope.app.interface though.
I don't think zope.app.zcml files should be in it either. It was intended as a bridge to get us from teh monolith world to the new world. Perhaps zope.app.zcml files should become a sort of meta package for some working set of the things it depends on, at some revision at which it worked, perhaps corresponding to the 3.4.0 release/kgs. Jim -- Jim Fulton Zope Corporation
On Thu, Jul 2, 2009 at 3:41 PM, Jim Fulton<jim@zope.com> wrote:
On Jul 2, 2009, at 7:58 AM, Hanno Schlichting wrote:
and zope.app.zcmlfiles still have a hard dependency on zope.app.interface though.
I don't think zope.app.zcml files should be in it either. It was intended as a bridge to get us from teh monolith world to the new world. Perhaps zope.app.zcml files should become a sort of meta package for some working set of the things it depends on, at some revision at which it worked, perhaps corresponding to the 3.4.0 release/kgs.
I think zcmlfiles is still a testing dependency of quite a bit of stuff. Basically anything that tries to do functional testing sets up some kind of complete test fixture based on this IIRC. And we haven't yet gotten rid of all zope.app.* packages that do this. I'm not saying this should be that case, but it might be something that needs to be resolved, before we can remove this thing from the KGS. Removing it when still various tests depend on it, doesn't strike me as the best way to do things. Hanno
On Jul 2, 2009, at 12:11 PM, Hanno Schlichting wrote:
On Thu, Jul 2, 2009 at 3:41 PM, Jim Fulton<jim@zope.com> wrote:
On Jul 2, 2009, at 7:58 AM, Hanno Schlichting wrote:
and zope.app.zcmlfiles still have a hard dependency on zope.app.interface though.
I don't think zope.app.zcml files should be in it either. It was intended as a bridge to get us from teh monolith world to the new world. Perhaps zope.app.zcml files should become a sort of meta package for some working set of the things it depends on, at some revision at which it worked, perhaps corresponding to the 3.4.0 release/kgs.
I think zcmlfiles is still a testing dependency of quite a bit of stuff. Basically anything that tries to do functional testing sets up some kind of complete test fixture based on this IIRC. And we haven't yet gotten rid of all zope.app.* packages that do this.
You're probably right
I'm not saying this should be that case, but it might be something that needs to be resolved, before we can remove this thing from the KGS. Removing it when still various tests depend on it, doesn't strike me as the best way to do things.
probably not Then the question is whether the dependence of zope.app.zcmlfiles on zope.app.interface is needed. I'll look to see what that's about. I need to resist the urge to clean things up and leave them in if they aren't causing test failures, I'd really like to get to something more flexible that zope.app.testing.functional, to make it simpler for applications to use only as much set up as they need. zope.app.testing.function was great in its time, but I think we can do better. Jim -- Jim Fulton Zope Corporation
Hey, Jim Fulton wrote: [snip]
Then the question is whether the dependence of zope.app.zcmlfiles on zope.app.interface is needed. I'll look to see what that's about.
For the record, I'd be happy to see zope.app.interface and zope.app.module gone from our dependencies. zope.app.zcmlfiles too, but that's going to be a harder nut to crack, I expect.
I'd really like to get to something more flexible that zope.app.testing.functional, to make it simpler for applications to use only as much set up as they need. zope.app.testing.function was great in its time, but I think we can do better.
+1. It's keeping a lot of (testing) dependencies alive. In general it's nice if a package's testing dependencies are equivalent to its normal dependencies, i.e. (almost) no testing dependencies. Regards, Martijn
participants (3)
-
Hanno Schlichting -
Jim Fulton -
Martijn Faassen