Re: A new proposal to add methods to HttpsURLConnection to access SSLSession

2018-11-08 Thread Xuelei Fan
To make sure the SecureCacheResponse class work, two new tests are added (DefaultCacheResponse for default implementation, DummyCacheResponse for a test implementation): http://cr.openjdk.java.net/~xuelei/8212261/webrev.04/ Thanks, Xuelei On 11/8/2018 7:03 AM, Sean Mullan wrote: On 11/7/1

Re: A new proposal to add methods to HttpsURLConnection to access SSLSession

2018-11-08 Thread Daniel Fuchs
On 08/11/2018 15:03, Sean Mullan wrote: Ah, I see. I am sure there is a good reason, but why doesn't it have an implementation? IIRC there was an implementation in the deploy code. best regards, -- daniel

Re: A new proposal to add methods to HttpsURLConnection to access SSLSession

2018-11-08 Thread Sean Mullan
On 11/7/18 7:22 PM, Xuelei Fan wrote: On 11/7/2018 1:30 PM, Sean Mullan wrote:    https://bugs.openjdk.java.net/browse/JDK-8213161    http://cr.openjdk.java.net/~xuelei/8212261/webrev.03/ I didn't see a test for SecureCacheResponse - is it possible? JDK does not have the reference implementat

Re: RFR [12] 8213418: Socket/ServerSocket supportedOptions does not work with custom SocketImpl

2018-11-08 Thread Alan Bateman
On 08/11/2018 13:03, Chris Hegarty wrote: On 08/11/18 11:53, Chris Hegarty wrote: ... Received some offline comments about the use of VarHandle. Removed the superfluous use of VarHandle, we now implicitly trust the socket impl ( even more ). http://cr.openjdk.java.net/~chegar/8213418/webr

Re: RFR [12] 8213418: Socket/ServerSocket supportedOptions does not work with custom SocketImpl

2018-11-08 Thread Chris Hegarty
On 08/11/18 11:53, Chris Hegarty wrote: ... Received some offline comments about the use of VarHandle. Removed the superfluous use of VarHandle, we now implicitly trust the socket impl ( even more ). http://cr.openjdk.java.net/~chegar/8213418/webrev.02/ -Chris.

Re: RFR:8046500 GetIpAddrTable function failed on Pure Ipv6 environment

2018-11-08 Thread Chris Hegarty
Hi Vyom, On 08/11/18 09:22, vyom tewari wrote: Hi, Please review the below code change. Webrev: http://cr.openjdk.java.net/~vtewari/8046500/webrev0.0/index.html enumInterfaces may return -1 for error conditions other than IP Helper Library GetIfTable function failed. I wonder if this specifi

RE: JDK-7014531

2018-11-08 Thread MUELLER-SCHRAMM Gerd
Wildcards work if you set the system property "http.nonProxyHosts" but not if you set "java.net.useSystemProxies". System proxy is evaluated in http://hg.openjdk.java.net/jdk/jdk/file/4547f8303f2d/src/java.base/windows/native/libnet/DefaultProxySelector.c, line 272 and this is a simple compare,

Re: JDK-7014531

2018-11-08 Thread Chris Hegarty
Gerd, On 07/11/18 15:24, MUELLER-SCHRAMM Gerd wrote: Hi, I’ve got a question regarding JDK-7014531 . This bug is still open and after checking it against Java 11 still valid. Will it ever be fixed? > I’ve took a look at the source code and t

Re: RFR [12] 8213418: Socket/ServerSocket supportedOptions does not work with custom SocketImpl

2018-11-08 Thread Chris Hegarty
On 08/11/18 08:22, Alan Bateman wrote: ... Right, the caching, if any, has to be per instance in order to work with different SocketImpl implementations. The change drops the synchronization and so brings up the question as to whether this method is thread safe or not. One simple approach (a

RFR:8046500 GetIpAddrTable function failed on Pure Ipv6 environment

2018-11-08 Thread vyom tewari
Hi, Please review the below code change. Webrev: http://cr.openjdk.java.net/~vtewari/8046500/webrev0.0/index.html BugId:  https://bugs.openjdk.java.net/browse/JDK-8046500 This issue is specific to pure ipv6 environment, In pure ipv6 environment  if "enumInterface" fails,  i am clearing th

Re: RFR [12] 8213418: Socket/ServerSocket supportedOptions does not work with custom SocketImpl

2018-11-08 Thread Alan Bateman
On 07/11/2018 15:16, Chris Hegarty wrote: Socket and ServerSocket supportedOptions() caches the supported options in a static field on the assumption that all implementations support the same set of options. This assumption is incorrect. The cache can be on a per-socket basis ( which has been im