Hi, thanks all for your very helpful comments.
So I see that I cannot rely on my Service implementation at an early stage. But I can imagine that I query for the service provider and if it's not there I have to be able to live without it. The question that remains is, when I query the ServiceLoader for my impl at a later time, then I would find it registered, right? I assume that if my module specifies "provides <my service> with <my provider>", the provider will automatically be registered once the module is loaded. Can you confirm that? Thanks Christoph > -----Original Message----- > From: Alan Bateman [mailto:alan.bate...@oracle.com] > Sent: Mittwoch, 15. Februar 2017 13:22 > To: David Holmes <david.hol...@oracle.com>; Langer, Christoph > <christoph.lan...@sap.com>; Chris Hegarty <chris.hega...@oracle.com>; > Weijun Wang <weijun.w...@oracle.com> > Cc: jigsaw-dev@openjdk.java.net > Subject: Re: Extending java.base module > > > > On 15/02/2017 12:10, David Holmes wrote: > > On 15/02/2017 8:03 PM, Langer, Christoph wrote: > >> Hi Chris, Max, > >> > >> thanks for your quick answers. So the service approach seems to fit > >> quite well. > >> > >> But can I assume that my service implementation will be available > >> already at "bootstrap time" of the JDK? E.g. if I need to > >> register/reach my service already at the early stages of JVM > >> initialization, e.g. when a class java.lang.Thread gets initialized, > >> can I assume a service from my extension module would be available? > > > > I'm pretty sure the answer to that will be No! Thread is one of the > > earlier classes to be initialized, the module system is initialized > > much later. > That's right as only classes in java.base can be loaded during startup. > I don't know what services that Christoph is thinking of but hopefully > they can be deferred until the VM is initialized. > > -Alan