RFR: 8174039: (ch) DefaultAsynchronousChannelProvider should be split into platform specific versions

[Claes Redestad]

+1

Thanks for doing this!

/Claes

On 02/07/2017 01:30 PM, Langer, Christoph wrote:
>
> Thanks Alan for the quick review.
>
> I’ll push it to JDK10 then.
>
> Best regards
>
> Christoph
>
> *From:*Alan Bateman [mailto:Alan.Bateman at oracle.com]
> *Sent:* Dienstag, 7. Februar 2017 13:00
> *To:* Langer, Christoph <christoph.langer at sap.com>; 
> nio-dev at openjdk.java.net
> *Subject:* Re: RFR: 8174039: (ch) DefaultAsynchronousChannelProvider 
> should be split into platform specific versions
>
> On 07/02/2017 11:54, Langer, Christoph wrote:
>
>     Hi,
>
>     as discussed yesterday in
>     http://mail.openjdk.java.net/pipermail/nio-dev/2017-February/004107.html,
>     there is another potential for improvement like
>     https://bugs.openjdk.java.net/browse/JDK-8173982.
>
>     Bug: https://bugs.openjdk.java.net/browse/JDK-8174039
>     <https://bugs.openjdk.java.net/browse/JDK-8174039>
>
>     Webrev: http://cr.openjdk.java.net/~clanger/webrevs/8174039.0/
>     <http://cr.openjdk.java.net/%7Eclanger/webrevs/8174039.0/>
>
>     I also made a little cleanup in DefaultSelectorProvider.java.
>
>     The webrev is based on JDK9. Please let me know if this could
>     still go there or if I should do it in JDK10. I think the change
>     is rather explicit and clear – so maybe there’s a chance for 9, still?
>
> This looks okay to me.
>
> At this point in the release when P4 and P5 issues are meant to be 
> deferred. So it seems like something for JDK 10 to me. The reason that 
> the FileSystemProvider change was important for JDK 9 is because it's 
> related to a small regression in startup performance.
>
> -Alan
>