[jdk17u] RFR: 8263364: sun/net/www/http/KeepAliveStream/KeepAliveStreamCloseWithWrongContentLength.java wedged in getInputStream

[Daniel Fuchs]

On Wed, 24 Nov 2021 12:23:35 GMT, Evan Whelan <ewhelan at openjdk.org> wrote:

> Hi all,
> 
> Please review my clean backport of JDK-8263364.
> 
> This adds some test stability to the `sun/net/www/http/KeepAliveStream/KeepAliveStreamCloseWithWrongContentLength.java` test.
> 
> I verified the fix by running a test-repeat of 25 on this test.
> 
> Kind regards,
> Evan

LGTM

test/jdk/sun/net/www/http/KeepAliveStream/KeepAliveStreamCloseWithWrongContentLength.java line 27:

> 25:  * @test
> 26:  * @bug 4533243 8263364
> 27:  * @summary Closing a keep alive stream should not give NullPointerException and should accept a connection from  a

The summary should preferably tell what functionality/API/behavior the test is testing.
Sorry I missed that in the original fix to the main line. 
Since this is a backport I agree to keep the `@summary` in sync with what is in the mainline.

test/jdk/sun/net/www/http/KeepAliveStream/KeepAliveStreamCloseWithWrongContentLength.java line 67:

> 65:                 clientSocket = null;
> 66: 
> 67:                 // in a concurrent test environment it can happen that other rouge clients connect to this server

I guess clients are rogue and not rouge ;-)
Maybe we should clean that up in the mainline next time we revisit this test...

-------------

Marked as reviewed by dfuchs (Reviewer).

PR: https://git.openjdk.java.net/jdk17u/pull/309