Backporting JDK-8217264 on JDK11u

[Langer, Christoph]

Hi Brice,

sounds good. For the javadoc change we might think about going through the CSR process to be on the safe side (although the original change doesn’t have a CSR attached). But in my opinion the javadoc change is worthwhile doing, too.

Cheers
Christoph

From: Brice Dutheil <brice.dutheil at gmail.com>
Sent: Dienstag, 11. Oktober 2022 10:39
To: Langer, Christoph <christoph.langer at sap.com>
Cc: jdk-updates-dev at openjdk.org
Subject: Re: Backporting JDK-8217264 on JDK11u

You don't often get email from brice.dutheil at gmail.com<mailto:brice.dutheil at gmail.com>. Learn why this is important<https://aka.ms/LearnAboutSenderIdentification>
Hi Chrstopher,

Thank you for the feedback.

I will go ahead with maybe two PR: one for the javadoc and one for the deadlock prevention fix.

On Mon 10 Oct 2022 at 18:07 Langer, Christoph <christoph.langer at sap.com<mailto:christoph.langer at sap.com>> wrote:
Hi Brice,

yes, the backport you suggest seems appropriate. Please go ahead with a PR if you like.

Cheers
Christoph

From: jdk-updates-dev <jdk-updates-dev-retn at openjdk.org<mailto:jdk-updates-dev-retn at openjdk.org>> On Behalf Of Brice Dutheil
Sent: Samstag, 1. Oktober 2022 14:01
To: jdk-updates-dev at openjdk.org<mailto:jdk-updates-dev at openjdk.org>
Subject: Backporting JDK-8217264 on JDK11u

Some people who received this message don't often get email from brice.dutheil at gmail.com<mailto:brice.dutheil at gmail.com>. Learn why this is important<https://aka.ms/LearnAboutSenderIdentification>
Hi everyone,

I was redirected there, to ask about backporting some fixes made to `HttpClient` to JDK11u.

Indeed, some API usage around the BodySubscribers with a common need of handling GZipped HTTP payload, are problematic (results in a deadlock) on JDK11 because,

1. The API is improperly documented ; this was fixed in JDK 13 with https://bugs.openjdk.org/browse/JDK-8217627
2. Daniel fuchs contributed a mechanism that prevents the deadlock, also contributed in JDK 13 with https://bugs.openjdk.org/browse/JDK-8217264

While I would prefer users to upgrade to a more recent JDK, I believe fixing this deadlock on an LTS seems appropriate.

Thank you in advance for considering the backport of these fixes.

Cheers,
-- Brice
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.openjdk.org/pipermail/jdk-updates-dev/attachments/20221011/f0e6d397/attachment-0001.htm>