Release Candidate 2.0.0

[rzo1]

Hi folks,

a new version of JWPL is in the starting blocks.

This version contains many improvements and bug fixes, e.g. the new release supports the Jakarta namespace and sets the Java language level to Java 17.
Please note, that the packages and artifact ids also changed, so this release is considered breaking.

An overview of the changes can be found at https://github.com/dkpro/dkpro-jwpl/milestone/14?closed=1

This version should work with the Wikipedia dumps from 2022. Since there was also a modernization of the Mediawiki schema, we will bring the corresponding support to Swebble in a next step. Unfortunately, development has also fallen asleep here, so we have to do a bit more ;-) But help and PRs are always welcome!

Here is the staging repository:

https://oss.sonatype.org/content/repositories/orgdkprojwpl-1003/

Here is the tag:

https://github.com/dkpro/dkpro-jwpl/releases/tag/dkpro-jwpl-2.0.0

Please provide feedback via GitHub discussions or on the related mail thread.

Best regards
Richard

[mawiesne]

+1 to publish this release candidate.

[mawiesne]

FYI @tgalery

[reckart]

I locally upgraded the dkpro-core-io-jwpl-asl code to build against the RC. Looks like the code compiles. Unfortunately, the unit tests do not work anymore since the test DB is no longer available :(

[reckart]

I'm still planning to have a look at the artifacts themselves, but didn't get to it yet...

[reckart]

• ✅ Upgraded DKPro Core JWPL Reader against the RC - looks good
• ⚠️ Notice file in util module contains a strange class name: parser.templates.util.org.dkpro.jwpl.SectionExtractor:
• ⚠️ swc-engine-shade module does not remap the packages - this can lead to problems in scenarios where JWPL is used together with another library that uses a non-shaded sweble. Can we remap the packages?
• ⚠️ swc-engine-shade NOTICE file does not describe the origin and original version of the artifact
• ⚠️ revision-machine module contains a "package" called configSamples - does this need to be part of the JAR?
• ⚠️ jwpl-api module contains a packaged called org.dkpro.jwpl.util - there should only be a single top-level package in each JAR - this package overlaps with a package of the same name in the jwpl-util module. Can the package in jwpl-api be renamed or moved to jwpl-util?
• ⚠️ the jwpl-api module contains a file jwpl-env-configuration.properties in the root packages. The root package should not be used for files or classes. Is the file needed? Can it be moved to a proper package?
• ⚠️ the jwpl-api module contains a log4j2.xml file - such files may be acceptable in far-jar modules that are not used as libraries by other application - but the jwpl-api module seems to be a library module and thus should not contain such a file.
• ⚠️ the jwpl-util module contains files jwpl-env-configuration.properties in the root packages. Are they needed? Can they be moved to a proper package?

[rzo1]

#283

[rzo1]

@reckart @mawiesne I re-rolled the release with the fixes from #283 and updated the announcement to point to the new staging repository. Please check ;-)

[reckart]

I locally checked out the 2.0.0 tag and built.

I can still see the inflating: create_tpltables_simplegenerator.sql in the dkpro-jwpl-util-2.0.0.jar

At least DKPro Core is using dkpro-jwpl-revisionmachine-2.0.0.jar as a library which means that it is running into a problem with the log4j2.xml file contained in that JAR. It seems not to be a fat JAR, so maybe having that file in there is not necessary?

There are tons of warnings in the build log from the build.

That said... I wouldn't -1 a release either at this point. So +0 from me. There is still some stuff that could be improved (in particular the log4j2.xml)... So +0 from me.

[rzo1]

@reckart See #285 - this should remove log4j2 from the jars distributed to mvn central, moves the sql files to README and fixes some shade issues.

There are a lot of Javadoc warnings / issues inside the build log, but that is something for the next minor version of JWPL (imho). The Javadoc plugin seems to be configured really harsh forcing to document even useless things ;-)

[rzo1]

After #285 is included, I will build another RC and we should be good to go (for now), imho.

[mawiesne]

+1