[Building Sakai] BaseSiteService: is this a bug?

Zach A. Thomas zach at aeroplanesoftware.com
Tue Mar 17 14:58:08 PDT 2009


On Mar 17, 2009, at 2:45 PM, Stephen Marquard wrote:

> Generally speaking I'd say this is a dubious approach.
>
> In particular, it's not a good idea to introduce dependence on  
> current tool context into the service implementation, as it can  
> break use from non-tool contexts like web services, quartz jobs,  
> etc. s a dubious approach.
>
> If the only context you want people to be able to create a site is  
> from within a specific tool, rather add a security advisor there.

The SecurityAdvisor approach works, but the BaseSiteService code still  
suggests it's doing something which it doesn't do.

And as long as we're talking about SecurityAdvisor: isn't it odd that  
I have no way of controlling whether another inhabitant of my thread  
removes my advisors? In other words, I generally want to be able to  
push an advisor, do some work, then pop my advisor, but I can't know  
that the state of the SecurityAdvisors stack is the same when I pop as  
when I pushed.

Shrug, my current problem is solved anyway.

thanks, Stephen.

Zach


More information about the sakai-dev mailing list