[vector] specification update to Vector.random
Paul Sandoz
paul.sandoz at oracle.com
Tue Oct 29 17:44:32 UTC 2019
Thanks.
I’ll push this, then send another patch later today to remove the following:
- Vector.random
- Vector.toList
- Vector.single
Paul.
> On Oct 29, 2019, at 2:50 AM, Yang Zhang (Arm Technology China) <Yang.Zhang at arm.com> wrote:
>
> Hi Paul
>
> The patch looks good to me.
>
> I think Vector.single can be removed too. It just reuses zero and withLane API. Developers can implement this by themselves. https://bugs.openjdk.java.net/browse/JDK-8226369
> How about creating such a list? So that we can discuss and remove them one by one.
>
> Regards
> Yang
>
> -----Original Message-----
> From: panama-dev <panama-dev-bounces at openjdk.java.net> On Behalf Of Paul Sandoz
> Sent: Tuesday, October 29, 2019 1:57 AM
> To: panama-dev at openjdk.java.net
> Subject: [vector] specification update to Vector.random
>
> Hi,
>
> Here is an update to specification of Vector.random to avoid direct binding to ThreadLocalRandom:
>
> http://cr.openjdk.java.net/~psandoz/panama/vector-random/webrev/ <http://cr.openjdk.java.net/~psandoz/panama/vector-random/webrev/>
>
> —
>
> Separately, I question whether this operation is needed in the first revision of the API. My inclination is to update the spec and push, then remove it (so we have a history if we want ti revive later on).
>
> Expanding further, there may be other methods that might be for the chop too, has anyone considered such a list?
>
> I am inclined to remove Vector.toList. I suspect it’s use will be very limited, given toArray and toString can cover debugging cases, and may be more of a distraction until we can, for example, express List<int>.
>
> Paul.
More information about the panama-dev
mailing list