Hi, Am Samstag, den 05.04.2008, 18:49 +0200 schrieb Alexander Saar: > Hi all, > very interesting discussion and I see that this issues raises more > possibilities and questions than I expected. > > I agree that the final checkin state of a node (or a set of nodes) is > something that is application specific and therefore it should be > possible for a bundle developer to decide what the final state should > be. To be honest I started with a special flag as part of the content > definition (didn't know jcr:isCheckedOut before), but as said in my > first mail I was not satisfied with that solution because it was > something that was not part of the content but rather a piece of meta > information defining how to handle the content import. > > But with jcr:isCheckedOut property this should be solvable in a way > that prevents such a mix.
Well not completely because it requires special casing for the property no matter the name because jcr:isCheckedOut is protected. Nevertheless, the property has the advantage, that it may already be set in the repository when you extract it to package it in the bundle. > I also like the idea of a global flag as > part of the bundle, because this would allow to control the behaviour > at one point, which is useful for big content definition files. > > I will try to extend my patch with a global flag and to respect > jcr:isCheckedOut. After that I will add an according JIRA issue so > that Sling experts can have a look at it. Cool. Thanks. Regards Felix
