[EXTERNAL] Re: Backporting Keepalive Windows Extended Socket Options (JDK-8308593) to 11u and 17u

[Terry Chow (SIMBA TECHNOLOGIES INC)]

Thanks Severin for your feedback. I'll start on the approval process for the backport since all looks ok.
________________________________
From: Severin Gehwolf <sgehwolf at redhat.com>
Sent: November 7, 2023 2:30 AM
To: Terry Chow (SIMBA TECHNOLOGIES INC) <v-terrychow at microsoft.com>; jdk-updates-dev at openjdk.org <jdk-updates-dev at openjdk.org>
Subject: Re: [EXTERNAL] Re: Backporting Keepalive Windows Extended Socket Options (JDK-8308593) to 11u and 17u

[You don't often get email from sgehwolf at redhat.com. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]

Hi,

Thanks for reaching out.

On Fri, 2023-11-03 at 23:33 +0000, Terry Chow (SIMBA TECHNOLOGIES INC)
wrote:
> Hello Severin and JDK maintainers,
>
> The backports for 21u and 17u have been completely merged, and so I
> want to move forward with discussing on getting the backport into
> 11u. I've already done some initial work/investigating for
> backporting to 11u. The following are necessary changes to ensure a
> working backport:
>    1. Lib-jdk.net.gmk will need to be updated to support building
> extnet for Windows.
>    2. jdk.net.ExtendedSocketOption will need to add a case to support
> returning a Windows instance of PlatformSocketOptions.
>    3. The Windows PlainSocketImpl class will need to have similar
> setOption/getOption methods as the Unix PlainSocketImpl class to
> support setting these extended socket options for Windows.
> For points 1 and 2, those changes were also required for the 17u
> backport and so I believe they should also be acceptable for 11u. For
> point 3, that is mostly copying what's already in the Unix
> PlainSocketImpl class over to the Windows version of the class.
> Overall, the changes seem innocuous, but let me know if I missed
> anything that could be disruptive for 11u.

While the JDK 11u updates tree is a very stable bugfix-only release at
this point, it seems OK to backport this feature for Windows as it
would mostly affect Windows code (one platform). The risk seems
tolerable. Personally, for the JDK 11u backport I would refrain from
those renaming changes, though.

Thanks,
Severin

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/jdk-updates-dev/attachments/20231109/50d7c1d5/attachment.htm>