Re: SVN: zope.location/trunk/s - moved IPossibleSite and ISite from zope.app.component to zope.location
Philipp von Weitershausen <[email protected]>
| Newsgroups | gmane.comp.web.zope.zope3 |
|---|---|
| Message-ID | <[email protected]> |
Thomas Lotze wrote: > Philipp von Weitershausen wrote: > >> This is an interesting move. I can settle with the idea, but I wonder this >> was discussed... Last time I remember we were discussing the idea of a >> package a la zope.site or zope.componentsite... (Also, as long as we're >> moving things, I wouldn't mind renaming ISite to IComponentSite). > > There had been some discussion on this list a week ago when we discovered, > during some dependency clean-up at gocept, that ISite being defined in a > zope.app package caused a lot of non-zope.app packages to indirectly > depend on a wagon load full of zope.app stuff by way of zope.location. > > So the idea to solve this was two-fold: > > - Step 1: Move ISite to zope.location to kill the dependency of > zope.location on zope.app.component. This would not affect anything > aside from the dependencies, but that would be a gain of itself. > > - Step 2: Something should be done about the naming of ISite and the > non-separation of the concepts of a component registration context and > locality in the zope.location sense. This requires further discussion > and is a more long-term goal than cleaning up the spurious dependency on > zope.app.component. That discussion is, however, not affected by ISite > being moved to zope.location. > > I hope this makes things a bit clearer. Ok, I can live with that. Thanks for the clarification! -- http://worldcookery.com -- Professional Zope documentation and training