Thanks, Peter.

I now realized where that Release Note list in the doc distribution is, so
it is easy to refer to it.

Cheers,
Zsolt

On Sat, Jan 7, 2017 at 11:57 PM, Peter <j...@zeus.net.au> wrote:

> Hi Zsolt,
>
> See inline below...
>
> On 7/01/2017 10:16 PM, Zsolt Kúti wrote:
>
>> Hi Peter,
>>
>> I remember some of your mails having summarized developments for 3.0.0. I
>> have been browsing the mailing list for an hour now, but I am yet to find
>> any of them.
>> Do you have one of those by any chance?
>>
>> I am aware of the followings:
>>
>> RIVER-431: Java Memory Model Compliance
>>
>> https://issues.apache.org/jira/browse/RIVER-431
>>
>>
>> RIVER-420: Export during construction
>>
>> https://issues.apache.org/jira/browse/RIVER-420
>>
>>
>> RIVER-418: Service server implementations start threads before
>> construction
>> is complete allow "this" to escape
>>
>> https://issues.apache.org/jira/browse/RIVER-418
>>
>>
>> Java 9 compliant
>>
> Not yet, the Java 9 changes aren't included in the River 3.0 release,
> although the jsk-policy.jar doesn't need to be in the extension ClassLoader
> like earlier releases.  Java 9 doesn't allow extensions.
>
>
>> permission tool?
>>
>> http://mail-archives.apache.org/mod_mbox/river-dev/201604.mb
>> ox/%3c5704fd9d.2030...@zeus.net.au%3e
>>
> The security policy file generation tool was created in response to
> requests to develop tools to make configuring River security easier,
> however it was rejected by a committer citing security concerns.  The tool
> grants AllPermission, while it determines required permissions and writes
> them to a policy file.
>
> The original intent was to advise the djinn network be air gapped, or use
> a test network environment, while using the tool to establish / generate
> security policy files, once the policy files are generated, have been
> reviewed and are in force, the network could be connected to outside
> networks.
>
> It's possible that the tool could be modified to use an interactive GUI,
> where an administrator can "ok" new permissions as they are requested,
> however doing so would be complex:
>
>    * A number of jvm's may be generating their policy files at the same
>      time, they may or may not be headless, which may require a remote
>      service, in order to communicate with an administrator
>      interactively, which creates other security problems.
>    * During a SecurityManager.checkPermission call, it's important to
>      avoid recursive calls where additional permission checks are
>      required in order to interact with an administrator using a
>      service / GUI.
>
> It's something we could look into again.
>
>>
>> Has this patch been revoked?
>>
>> http://mail-archives.apache.org/mod_mbox/river-dev/201604.mb
>> ox/%3c570cccaa.5090...@zeus.net.au%3e
>>
> No, this patch is included.
>
>>
>>
>> Cheers,
>> Zsolt
>>
>
>
>  Release Notes - River 3.0.0
>
>
>    Sub-task
>
> Release Notes - River - Version River_3.0.0
>
>
>    Sub-task
>
>    * [RIVER-319 <https://issues.apache.org/jira/browse/RIVER-319>] -
>      Change River Build Dist structure to support jtreg test automation
>    * [RIVER-344 <https://issues.apache.org/jira/browse/RIVER-344>] -
>      com.sun.jini.thread.TaskManager scalability and concurrency.
>
>
>    Bug
>
>    * [RIVER-19 <https://issues.apache.org/jira/browse/RIVER-19>] -
>      PreferredClassLoader doesn't implement preferred semantics for
>      getResources(String)
>    * [RIVER-113 <https://issues.apache.org/jira/browse/RIVER-113>] -
>      JoinManager synchronization on each proxyReg should be reviewed,
>      doc'd and fixed where appropriate
>    * [RIVER-145 <https://issues.apache.org/jira/browse/RIVER-145>] -
>      JoinManager synchronization on serviceItem should be reviewed,
>      doc'd and fixed where appropriate
>    * [RIVER-148 <https://issues.apache.org/jira/browse/RIVER-148>] -
>      JoinManager.ProxyReg.fail synchronization may be wrong or may be
>      able to simplify it
>    * [RIVER-265 <https://issues.apache.org/jira/browse/RIVER-265>] -
>      PreferredClassProvider performs 'unlucky' caching
>    * [RIVER-282 <https://issues.apache.org/jira/browse/RIVER-282>] -
>      Suspect exception cast
>    * [RIVER-335 <https://issues.apache.org/jira/browse/RIVER-335>] -
>      com.sun.jini.phoenix.ConstrainableAID missing from phoenix.jar
>    * [RIVER-337 <https://issues.apache.org/jira/browse/RIVER-337>] -
>      Attempted discard of unknown registrar kills
>      LookupLocatorDiscovery thread
>    * [RIVER-345 <https://issues.apache.org/jira/browse/RIVER-345>] -
>      SDM LookupCache multi-LUS stale proxy/discard problems
>    * [RIVER-348 <https://issues.apache.org/jira/browse/RIVER-348>] -
>      Possible race condition in net.jini.lookup.ServiceDiscoveryManager
>      addProxyReg
>    * [RIVER-367 <https://issues.apache.org/jira/browse/RIVER-367>] -
>      com.sun.jini.mahalo.TxnManagerImpl fails to abort a Transaction
>      when notified of its lease expiration.
>    * [RIVER-387 <https://issues.apache.org/jira/browse/RIVER-387>] -
>      KerberosServerEndpoint calls com.sun.security methods,
>      animal-sniffer warns
>    * [RIVER-395 <https://issues.apache.org/jira/browse/RIVER-395>] -
>      Ill-behaved DiscoveryListener can terminate discovery notifier
>      threads
>    * [RIVER-402 <https://issues.apache.org/jira/browse/RIVER-402>] -
>      NullPointerException in LookupCacheImpl.notifyServiceMap
>    * [RIVER-418 <https://issues.apache.org/jira/browse/RIVER-418>] -
>      Service server implementations start threads before construction
>      is complete allow "this" to escape
>    * [RIVER-420 <https://issues.apache.org/jira/browse/RIVER-420>] -
>      Export during construction.
>    * [RIVER-422 <https://issues.apache.org/jira/browse/RIVER-422>] -
>      Missing reference-collections and high-scale-lib in Manifest for
>      jsk-platform
>    * [RIVER-431 <https://issues.apache.org/jira/browse/RIVER-431>] -
>      Java Memory Model Compliance
>    * [RIVER-433 <https://issues.apache.org/jira/browse/RIVER-433>] -
>      Test suite freeze while testing service discovery category
>
>
>    Improvement
>
>    * [RIVER-26 <https://issues.apache.org/jira/browse/RIVER-26>] - Make
>      UmbrellaGrantPermission work with DynamicPolicy
>    * [RIVER-107 <https://issues.apache.org/jira/browse/RIVER-107>] -
>      DynamicPolicyProvider could use finer grained locking
>    * [RIVER-123 <https://issues.apache.org/jira/browse/RIVER-123>] -
>      ConfigurationFile should support arithmetic operations
>    * [RIVER-140 <https://issues.apache.org/jira/browse/RIVER-140>] -
>      JoinManager synchronization strategy should be reviewed,
>      documented, and fixed where appropriate
>    * [RIVER-193 <https://issues.apache.org/jira/browse/RIVER-193>] -
>      support declaring entries in a "common" configuration source for
>      use in other configuration sources
>    * [RIVER-249 <https://issues.apache.org/jira/browse/RIVER-249>] -
>      DynamicPolicy providers do not support UmbrellaGrantPermission
>    * [RIVER-274 <https://issues.apache.org/jira/browse/RIVER-274>] -
>      Improve logging of diagnostic messages in ServiceDiscoveryManager
>    * [RIVER-343 <https://issues.apache.org/jira/browse/RIVER-343>] -
>      Private class extends java.lang.Thread, causing synchronization
>      bottleneck.
>    * [RIVER-386 <https://issues.apache.org/jira/browse/RIVER-386>] -
>      Refactor of FastList inside of Outrigger
>    * [RIVER-401 <https://issues.apache.org/jira/browse/RIVER-401>] -
>      PreferredClassProvider using URL as key in map
>    * [RIVER-412 <https://issues.apache.org/jira/browse/RIVER-412>] -
>      rename com.sun.jini packages to org.apache.river.impl
>    * [RIVER-439 <https://issues.apache.org/jira/browse/RIVER-439>] -
>      River only builds on Sun's JVM, add support for other JVM's
>
>
>    New Feature
>
>    * [RIVER-313 <https://issues.apache.org/jira/browse/RIVER-313>] -
>      Provide mechanism to swap in alternatives to Java DSL for service
>      configuration
>    * [RIVER-340 <https://issues.apache.org/jira/browse/RIVER-340>] -
>      Additional Dynamic Grants and Revokeable Permissions
>
>
>    Question
>
>    * [RIVER-365 <https://issues.apache.org/jira/browse/RIVER-365>] -
>      main build.xml contains remarks about deprecated (and to be
>      removed) targets, needs clarification
>
>
>    TCK Challenge
>
>    * [RIVER-419 <https://issues.apache.org/jira/browse/RIVER-419>] -
>      ServiceDiscoveryManager lookup qa TCK tests need to be reviewed
>
>
>    Task
>
>    * [RIVER-261 <https://issues.apache.org/jira/browse/RIVER-261>] -
>      update com.sun.* namespace to org.apache.river.*
>
>
>    Test
>
>    * [RIVER-304 <https://issues.apache.org/jira/browse/RIVER-304>] -
>      Reactivate River jtreg tests
>
>
> On Thu, Jan 5, 2017 at 11:58 PM, Peter<j...@zeus.net.au>  wrote:
>>
>> BTW I'm happy for you to announce 3.0 on the new website.
>>>
>>> Cheers,
>>>
>>> Peter.
>>>
>>> Sent from my Samsung device.
>>>
>>>    Include original message
>>> ---- Original message ----
>>> From: Peter<j...@zeus.net.au>
>>> Sent: 06/01/2017 07:35:52 am
>>> To: dev@river.apache.org<dev@river.apache.org>
>>> Subject: Re: about 3.0 artifacts and announcement
>>>
>>> Hi Zsolt,
>>>
>>> The release process looks up to date, it doesn't contain
>>> any detail on publishing jars to Maven, as a binary build
>>> isn't currently implemented.  Due to a bug in ClassDep that
>>> causes class files to be omitted from jars on occassion, we must run the
>>> full test suite to validate the build.  I also have
>>> Apache jar signer certs for signing release jars.
>>>
>>> As you're probably aware I'm working on / experimenting
>>> with a Maven build on git.  I'm finding the build process
>>> on Maven much simpler, modularity should lower the
>>> bar for new developers as it clearly demarcates components,
>>>   allowing them to focus on one component. With the
>>> recent discussion surrounding OSGi, I'm also creating
>>> bundles to allow us to explore the use of OSGi to
>>> manage the modules at runtime.  Merging&  donating this
>>>
>>> work back to River can occur as soon as River is tranisitioned to git
>>>
>>> Cheers,
>>>
>>> Peter.
>>>
>>>
>>>
>>> Sent from my Samsung device.
>>>
>>>    Include original message
>>> ---- Original message ----
>>> From: Zsolt Kúti<la.ti...@gmail.com>
>>> Sent: 06/01/2017 06:51:07 am
>>> To: dev@river.apache.org
>>> Subject: about 3.0 artifacts and announcement
>>>
>>> Hi,
>>>
>>> Can somebody tell if our release process documentation is up-to date:
>>> http://river.apache.org/dev-doc/building-a-release.html
>>>
>>> As to the release, the last mail was:
>>> http://mail-archives.apache.org/mod_mbox/river-dev/201610.
>>> mbox/%3cCAK_o9cH7JPsfd_CK4-pOGb3nswH4R8jB1Kh6=
>>> UTWF2c0Ge9V=w...@mail.gmail.com%3e
>>>
>>> The 3.0.0 release artifacts (no binary) are available from here:
>>> http://www.apache.org/dyn/closer.cgi/river/
>>> So nothing is against a release announcement on our website, isn't  it?
>>>
>>>
>>> If nobody else is willing to, I can take a look into how to
>>> add our jars to
>>> maven repo,.
>>>
>>> Cheers,
>>> Zsolt
>>>
>>>
>>>
>>>
>

Reply via email to