Java: Use the models packs in the query pack and promote experimental models.

[michaelnebel]

In this PR we

• Update the Java dependencies
• Apply the Java models in the model packs to the queries in the java query pack.
• Promote experimental models to production models (note that I have not reviewed the accuracy of these).

DCA was used to check that this didn't cause any massive discrepancies: https://github.com/github/codeql-dca-main/issues/24959
The experiment shows no changes in alerts or performance.
For the curious reader, here are the tuple counts for the queries in the default query suite for the java DCA projects: https://github.com/github/codeql-dca-main/blob/data/michaelnebel/auto/community-pack/e876bc7/1732286194857/reports/summaries/csv/Query_result_tuple_counts.csv

The experiment was started using the DCA command (for this to be reproducible https://github.com/github/codeql-dca/pull/2324 needs to be merged):

./dca generate --language java 416266523758ea327e414257e07fe5e0e1b8fd48@00ea65f151921841cfb5a102959fc867c41e3502 416266523758ea327e414257e07fe5e0e1b8fd48@6dd34ce0f226d96652fa9ef748ec9f2b1cca43eb --extra-config-file dca-configs/community-packs.yml

[pwntester]

Thanks @michaelnebel What were all those ActivateXXX classes for?

[michaelnebel]

Thanks @michaelnebel What were all those ActivateXXX classes for?

That was a way of "activating" (only include specific) experimental Models as Data models for a given query. Since the experimental models are converted to "ordinary" models (which means that they also apply to all queries), we no longer need to "activate" anything.