I think that it makes a lot of sense to follow the rules from type-safe
resolution here, i.e.
https://jakarta.ee/specifications/cdi/4.1/jakarta-cdi-spec-4.1#primitive_types_and_null_values.

Martin

On 17. 10. 24 9:39, Matej Novotny wrote:
> Hello,
>
> this is a good candidate for spec issue. I've created one -
> https://github.com/jakartaee/cdi/issues/829
> <https://github.com/jakartaee/cdi/issues/829>
> My personal take is that Weld behavior is a way to go as otherwise you'd
> have no use for observing primitive types.
> Strictly speaking, it is ATM Weld value-add but I don't see why the spec
> couldn't cover this or at least have some TCK for it (I found none).
>
> Matej
>
> On Thu, Oct 17, 2024 at 4:06 AM Laird Nelson <ljnelson@gmail.com
> <mailto:ljnelson@gmail.com>> wrote:
>
>     The CDI 4.0 spec says in section 2.8.1
>     (https://jakarta.ee/specifications/cdi/4.0/jakarta-cdi-spec-4.0#event_types_and_qualifier_types <https://jakarta.ee/specifications/cdi/4.0/jakarta-cdi-spec-4.0#event_types_and_qualifier_types>):
>
>     "Any Java type may be an observed event type."
>
>     That means I can do this:
>
>     private void primitive(@Observes int i) { // any type, including
>     int, can be an observed event type
>        System.out.println("primitive: " + i);
>     }
>
>     And this:
>
>     private void wrapper(@Observes Integer i) {
>        System.out.println("wrapper: " + i);
>     }
>
>     Then of course the only way to fire an event is by using the Event
>     object and specifying the type of even you want to fire as its sole
>     type argument:
>
>     // for example
>     @Inject
>     private Event<Integer> e;
>
>     (Type arguments of course cannot be primitive.)
>
>     No mention in the specification's section on observer resolution
>     talks about primitive types and/or wrapper types and how they
>     interact. (Obviously there are unrelated sections regarding typesafe
>     resolution in the context of dependency injection that do talk about
>     this, but they do not seem to apply here without a biiiiiig stretch
>     (observed event parameters are not injection points; event types are
>     not bean types).)
>
>     If I do e.fire(Integer.valueOf(42)), the primitive(int) observer
>     method above is called (in Weld 5.1.3.Final). So is the
>     wrapper(Integer) observer method. This certainly is convenient and
>     I'm kind of glad it works that way, but is it correct? Is it a Weld
>     value-add? Or should the spec have said something about
>     interoperability of primitives and wrapper types in observer resolution?
>
>     Best,
>     Laird
>     _______________________________________________
>     weld-dev mailing list -- weld-dev@lists.jboss.org
>     <mailto:weld-dev@lists.jboss.org>
>     To unsubscribe send an email to weld-dev-leave@lists.jboss.org
>     <mailto:weld-dev-leave@lists.jboss.org>
>     Privacy Statement: https://www.redhat.com/en/about/privacy-policy
>     <https://www.redhat.com/en/about/privacy-policy>
>     List Archives:
>     https://lists.jboss.org/archives/list/weld-dev@lists.jboss.org/message/IDWNDAJLY57OE6MJNNRVW5HSOV3MJLPJ/ <https://lists.jboss.org/archives/list/weld-dev@lists.jboss.org/message/IDWNDAJLY57OE6MJNNRVW5HSOV3MJLPJ/>
>
>
> _______________________________________________
> weld-dev mailing list -- weld-dev@lists.jboss.org
> To unsubscribe send an email to weld-dev-leave@lists.jboss.org
> Privacy Statement: https://www.redhat.com/en/about/privacy-policy
> List Archives: https://lists.jboss.org/archives/list/weld-dev@lists.jboss.org/message/FZQQCKWI6JJZ4A7I5MU7DNEVXGOXVERI/

--
Martin Kouba
Principal Software Engineer
Red Hat, Czech Republic