I did some initial testing of the current master branch, and generally it
looks great. It builds fine with JDK 8 and 11; all tests pass. I tried
using those artifacts in some downstream code, and all looks good
(including with Java 11 JPMS).

The only possible issue I encountered was with the OSGi deployment. In
Karaf 4.2.6 and later, the `feature:install jena` command works if I
manually add a javax.annotation bundle (either from jakarta or the regular
javax dependency). Given that this is a platform api and downstream users
might have an opinion about which artifact to use for this, it might make
sense to leave this as-is. It is trivial for downstream users to work
around this.

The one thing that should probably be fixed before a release is that, in
Karaf 4.2.7 and later, the log4j2 dependency has to be manually added in
order to start up the Jena feature. Since log4j2 is a particular logging
implementation, I suspect that is really should be excluded from the OSGi
bundle (as `org.apache.log` currently is). If that seems right, I could
supply a PR.

Aaron



On Sun, 19 Apr 2020 at 16:40, Andy Seaborne <[email protected]> wrote:

> I'm sending this out a little early because of wanting to give warning
> of the logging change before release, and generally get people to try
> dev builds before we formally put out a release.
>
> Draft message for logging to come after this email.
>
> Outstanding:
> * JENA-1837: MaxBasicQueries for Lucene
>
> For JENA-1837, I don't see any reason not to merge it but it would be
> good for one of the reviewers to approve it.
>
> * Documentation for experimental RDF*
>
> Tickets:
> https://s.apache.org/jena-3.15.0-jira
>
>      Andy
>
>
> Major changes:
>
> While logging should be largely an invisible change, there are possible
> impacts.
>
> API-using code has had to have its own slf4j setup for sometime (we can
> recommend moving to log4j2). If it uses Jena's LogCtl, there might be a
> need to adjust.
>
> Fuseki or command line might be using a custom log4j1 setup (maybe
> custom log4j.properties or log4j1 log routing).
>
> I think sending a message user@ now to alter people is worth doing.
>
>
> Features:
>
> RDF* - experimental, in-memory only.
>
> (Doing TDB may happen, may not - there are two places to change, node
> storage and the basic graph pattern solver. The latter in the same way
> that StageGeneratorGenericStar works.)
>
> There have been quite a few contributions.
>
>
> This release:
>
>
> -- Upgrades
>
> JENA-1812: Update Commons Codec 1.13 -> 1.14
>
> JENA-1883: Commons Lang3 : 3.9 -> 3.10
>
> JENA-1839: Update Commons CSV 1.7 -> 1.8
>
> JENA-1005: Update slf4j 1.7.26 -> 1.7.30
>
> JENA-1850: Commons compress 1.19 -> 1.20
>
> JENA-1836:  Shiro: 1.2.6 -> 1.5.0
> CVE-2019-12422
>
> JENA-1850: Jetty: 9.4.12.v20180830 -> 9.4.26.v20200117
>
> JENA-1851: Lucene: 7.4.0 -> 7.7.2
> (Not -> 8.4.1)
>
> jena-es: ElasticSearch 6.4.2 -> ES 6.8.6 for Lucene 7.7.x
> (Not -> ES 7.6.0 for Lucene 8.4.x)
>
> JENA-1829: Apache Parent POM to v23.
>
> -- Jena-1005: Logging
>
> Separate message
>
> -- Website
>
> With the help of Roy Lenferink, the production of the Apache Jena
> website has moved from using svn and Apache's custom CMS to being stored
> in git and using Hugo.
>
> The website source is https://github.com/apache/jena-site
>
> One change is that "Improve this page" has become "Edit this page".
>
> Proposed changes to the website can be made with git pull requests.
>
> Many thanks to Roy for nudging us to make this migration.
>
> -- Contributions and Notable JIRA
>
> JENA-1881: Experimental support for RDF*
> This is limited to in-memory storage.
>
> JENA-1005: Logging
> (separate message)
>
> JENA-1884: Removed JenaUUID
>
> JENA-1827: Use ConcurrentHashMap in TypeMapper
> Simon Fell
>
> JENA-1835:
> Capture and expose lexicalForm and dataType in DatatypeFormatException
> Simon Fell
>
> JENA-1826: use plain RDF/XML in Fuseki RDF responses
> CONSTRUCT/DESCRIBE RDF/XML output change
>
> JENA-1831:
> Fix SDB transform issue with using UNION in restricting a query
> Graham Triggs
>
> JENA-1855 : TriG parser fix
> Claus Stadler
>
> JENA-1858 : SERVICE blocks after 5 attempts
> Claus Stadler
>
>
> JENA-1862: Improvements to Query.cloneQuery
> Claus Stadler
>
> JENA-1868, JENA-1869 : TDB2 issues
> Bernhard Stiftner
>
> JENA-1846: Fuseki CORS in Fuseki main
>
> JENA-1864:
> Fix for @base output in TURTLE_BLOCKS and TURTLE_FLAT formats.
> Viresh Gupta
>
> JENA-1878: Switch to using https://prefix.cc
> David Fichtmueller
>
>

Reply via email to