RFR [9] 8044773: Refactor jdk.net API so that it can be moved out of the base module

Chris Hegarty chris.hegarty at oracle.com
Tue Apr 26 09:51:38 UTC 2016

On 26 Apr 2016, at 10:35, Erik Joelsson <erik.joelsson at oracle.com> wrote:

> Hello Chris,
> In general it looks good.

Thanks for the review Erik.

> Just a couple style [1] nits that I would like to get sorted. In Lib-jdk.net.gmk, the arguments to SetupNativeCompilation should be indented 4 spaces relative to the call (continuation). Also line 32 and 45 needs a space after comma.

Got it. I updated the webrev in-place:


> /Erik
> [1] http://openjdk.java.net/groups/build/doc/code-conventions.html
> On 2016-04-25 23:01, Chris Hegarty wrote:
>> One of the remaining open issues in JEP 200 [1] is that the base module
>> exports the jdk.net package, thereby violating Principle 4 of JEP 200:
>> a Java SE module must not export any non-SE API packages without
>> qualification.
>> http://cr.openjdk.java.net/~chegar/8044773/
>> https://bugs.openjdk.java.net/browse/JDK-8044773
>> Summary:
>>  - The jdk.net package has been moved to the jdk.net module. The native
>>  code and implementation that provides support for extended socket
>>  options, currently just SO_FLOW_SLA, has been moved too. A private
>>  interface between the base module and the jdk.net module now supports
>>  registration and operation of extended options.
>>  - The native code has been simplified and cleaned up.
>>  - Test coverage has been updated to exercise socket options, both with
>>  the jdk.net module present, and absent. The use of the
>>  jdk.launcher.limitmods system property is temporary until jtreg with
>>  support for -limitmods is more widely available.
>>  - Thanks to Erik for helping with build support that now cooperates
>>  with the intention of the GenModuleInfoSource build tool, to support
>>  unqualified export additions from non-open code.
>> -Chris.
>> [1] https://bugs.openjdk.java.net/browse/JDK-8051618

More information about the build-dev mailing list