Jena, the codebase, passes the Java9 inspection tool so a release shoudl
run find with Java9.
It's just the maven build system that's lagging.
Working out where that is at would be a great help.
Andy
On 20/10/17 22:18, Claude Warren wrote:
Not really an immediate need so much as just wondering how close our code
is to working under Java 9. I think it would also be nice to know when the
various tools we use are Java 9 ready and perhaps lend them a hand if need
be. More curiosity than anything else.
Claude
On Fri, Oct 20, 2017 at 3:47 PM, Andy Seaborne <[email protected]> wrote:
Claude - you can see branches that exist via the GH interface.
And, no, theer isn't one.
There is a jenkins job - it does not work, waiting in updates to roll
through. Taking over version mgt of the plugins from the ASF parent seems
to me like extra work for little benefit.
Unless there is an immediate need?
Andy
On 19/10/17 08:19, Claude Warren wrote:
Did we get a Java 9 branch started? Seems like most of the issues are
around tooling not functionality of the product. If this is the case I
would expect the differences between the java9 branch and the master to be
contained in the pom.xml files.
On Wed, Oct 18, 2017 at 2:20 AM, Andy Seaborne <[email protected]> wrote:
That would be good to see.
Personally, I think that ways to use modules in term good practice and
patterns, and also frameworks, in the java ecosystem will emerge but
anything we can do to reduce barriers seems like a good thing.
On Java9 generally:
The build itself doesn't work with Java9 because it needs updated
versions
of some plugins, and those are inherited from the Apache parent POM. To
take over the version control and override the std settings just seems
like
much work to get ahead by a short period of time.
I'm assuming we stay on java8 as the requirement for applications for a
while yet.
Andy
On 17 October 2017 at 11:20, Aaron Coburn <[email protected]>
wrote:
Would it make sense to add an Automatic-Module-Name header to the
manifest
files so that Jena is easier to use in a JDK9 context?
I could even volunteer to do this.
Aaron
On Oct 17, 2017, at 9:56 AM, [email protected] wrote:
Claude--
I see some updates available for the contract test machinery:
org.xenei:contract-test-maven-plugin .................. 0.1.5 -> 0.1.7
org.xenei:junit-contracts ............................. 0.1.5 -> 0.1.7
Worth doing before a release?
ajs6f
Andy Seaborne wrote on 10/16/17 6:32 PM:
The tick is approaching.
Are we ready to go? JIRA to be marked resolved?
If so, I'll sort out a release soon.
Andy
Here's a list of changes of note that I gathered:
==== Release changes
Introducing TDB2:
http://jena.staging.apache.org/documentation/tdb2/
*TDB2 is not compatible with TDB1*
Compared to TDB1:
* No size limits on transactions : bulk uploads into a live Fuseki
can e 100's of millions of triples.
* Models and Graphs can be passed across transactions
* No queue of delayed updates, no transaction backlog problems.
* "Writer pays" - readers don't
All work for update is done on the writer thread.
* Datatypes of numerics preserved; xsd:doubles supported.
TDB2 is subject to change.
We solicit any and all feedback (good and bad!) about TDB2 to help
advance it to deployment-ready.
JENA-1390 : Add StmtIterator.toModel :
JENA-1392 : Add dynamic dataset support to SDB.
JENA-1395 : "--output RDF/XML" now prints using the basic
block-oriented
writer, which uses less memory. Use "--formatted" (same as
"--pretty")
for pretty printed RDF/XML.
JENA-1398 :
Upgrade FOAF to add new spelling and deprecation of old for archaic
FOAF
properties
== Dependency changes:
No license changes.
Upgrade jsonld-java to 0.11
jackson to 2.9.0
commons-fileuploader to 1.3.2->1.3.3
commons-io 2.5 in jena-base
(was pulled in anyway by jsonld-java)
