This implements the new mimetype range indexer only. The previous
indexers (base class ContentIndexer, and ContentMimetypeIndexer) are
still present.
The base class because it's still needed for the other not yet
migrated indexers (fossology license is next, language, ctags).
The ContentMimetypeIndexer because i'm not 100% certain we really want
to drop it entirely. That could be used if we wanted to make a journal
client just for the content indexing on new contents.
Note:
- The tests need to be rewritten to reuse in-memory storages (and ideally data generation). It's somewhat partially done but it's not good enough.I did not want to start this right now because i want to deploy! It needs to be done nonethelesss at some point (i'll open a task).
- I created a new type of tasks named swh.indexer.tasks.StatusTask who returns the task's status. That's more consistent with all other swh tasks (outside swh-indexer). We need to align on this behavior.
Related T991