ServiceLoader.load* take 2

[Alan Bateman]

On 12/06/2012 16:03, Paul Sandoz wrote:
> Hi,
>
> A set of slides has been created [1] which presents a set of terms, an overview of modular services, describes the current state of affairs with service instance creation and the use of ServiceLoader.load*, then proposes an alternative solution for ServiceLoader, the first part of which is implemented by the following webrev, which changes service instance creation so that it is scoped to the configuration rather than the class loader:
>
>    http://cr.openjdk.java.net/~psandoz/jigsaw/global-services/webrev/
>
> Summary: in module mode any module class loader can be used with ServiceLoader to create the service instances for a service interface. Previously the class loader of the service consumer module, that consumes the service interface, was required to be used.
>
I skimmed through the webrev and the changes look reasonable to me. It's 
good to have the hacks removed from ServiceLoader, as you know these 
were needed to avoid a pit stop to change the gear box. I'm also happy 
with that the proposal addresses the permits issue, that was a left-over 
open issue from the first installation. With the proposal then it's much 
simpler and is conceptually equivalent to have all the service providers 
on the class path.

One thing I didn't quite get was the reason for moving LazyIterator1 out 
of ServiceLoader to a new org.openjdk.jigsaw.ModuleServiceLoader. My 
comment is really just wondering about the iteration/instantiation in 
two places rather than one.

You listed in the "next steps" that the configuration needs to be 
updated to include a map from the service interface to the list of 
service provider modules that supply implementations and it would be 
good to do that (as it's regressing to have to check every context for 
service implementations).

On the new test then it reminds me that we may be able to consolidate 
the tests in the services directory into a more complete set of unit 
tests (just thinking that there is a bit of duplication between the new 
test the existing hello.sh, many.sh and optional.sh).

-Alan.