8202471: Resolves generic receiver type for types with generic signatures
David Holmes
david.holmes at oracle.com
Tue Jul 23 05:29:19 UTC 2019
On 23/07/2019 3:08 pm, Rafael Winterhalter wrote:
> Thanks, I'll send an inline version this evening. I have written a
> couple of reproducers for these issues. Should I add them to jtreg and
> also send them as an inline patch?
Yes, please do.
Thanks,
David
> I'll submit the CSR tonight, too.
>
> Thanks for the pointers!
> Best regards, Rafael
>
> David Holmes <david.holmes at oracle.com <mailto:david.holmes at oracle.com>>
> schrieb am Di., 23. Juli 2019, 06:10:
>
> Hi Rafael,
>
> A couple of comments on process here ...
>
> On 23/07/2019 6:48 am, Rafael Winterhalter wrote:
> > Hello,
> >
> > I have created a patch such that getReceiverType() returns a
> parameterized
> > type if the receiver type declaration is itself generic.
> Currently, the
> > receiver type is always a type representation of Class such that
> > annotations on the type variables or the receiver type's owner
> type cannot
> > be resolved:
> https://gist.github.com/raphw/a155d5ef66d11e5fb131b7e6b8fb10e5
>
> All OpenJDK contributions must be provided via OpenJDK infrastructure,
> so either direct code in an email to a mailing list (attachments
> usually
> get stripped), or a webrev hosted on cr.openjdk.java.net
> <http://cr.openjdk.java.net> (available
> directly to Authors else find someone to host for you). Links to github
> are not acceptable at this time.
>
> > Note that this change can potentially break existing code if
> callers of the
> > method expect this behavior for parameterized receiver types.
> However,
> > without this change, the type information is lost and I would
> argue that
> > the current behavior can be considered to be incorrect.
>
> That argument needs to be made via a Compatibility and Specification
> Review (CSR) request:
>
> https://wiki.openjdk.java.net/display/csr/Main
>
> Cheers,
> David
>
> > Best regards, Rafael
> >
>
More information about the core-libs-dev
mailing list