Open Source Web Crawler for Java (crawler4j) dependency issues in the Solr module
In Magnolia 6.2.23, the Tika libraries were updated (BUILD-684). However, the magnolia-content-indexer was still using the old Tika version, which created a conflict.
To eliminate this issue, the older tika-parsers artifact has been excluded from the magnolia-content-indexer in version 6.1.1 of the Solr module, followed by a crawler4j dependency update (4.4.0-magnolia > 4.4.1-magnolia) in version 6.1.2.
Therefore, version 6.1.2 of the Solr module is recommended for all users of Magnolia 6.2.23 or higher. For more details, see MGNLEESOLR-172Exclude dependency on tika-parsers `1.x`.
Old magnolia-content-indexer artifactId
The info.magnolia.solr groupId libraries keep bringing the old magnolia-content-indexer artifactId. This happens when the DX-core’s parent POM is used to manage Magnolia Solr module dependencies. The 6.2.26 parent POM manages Solr module in version 6.1.1, which is affected by the crawler4j dependency issue described above.
Until this old magnolia-content-indexer issue is fixed in the next Magnolia release, you can use the following workaround.
Workaround
When adding the magnolia-solr-search-provider artifactId dependency, exclude the magnolia-content-indexer.
The exclusion must be done for every library whose groupId is info.magnolia.solr (magnolia-solr-search-provider, magnolia-solr-templating, magnolia-solr-workbench, magnolia-solr-ui).