Retest .clear

[NickColley]

We've got a skipped .clear test that was blocked on the fact that PouchDB does not reinitilize the database, but now downstream we've worked around this with a .reset method in https://github.com/hoodiehq/hoodie-client-store

What's the plan of action to fix this test? Some options:

• See if PouchDB reiniting the database makes sense
• Fix clear in a similar way to what we've done for .reset
• Finally just test the fact that database no longer exists

[gr2m]

I would re-initialise the PouchDB db instance in .clear() as we do with .reset() in https://github.com/hoodiehq/hoodie-client-store.

A problem I see is that we expose the db instance at store.db which will be replaced on clear, but we can add that information to the README with a "use at your own risk" note

[gr2m]

strike that, I think we should no re-initialse anything on .clear(), instead do what PouchDB itself does: become unusable :) so the user has to create another instance manually if they want to add data again

[NickColley]

Could we rename clear to destroy to better reflect what's going on? That conflicts with PouchDB's api though, is it necessary to have the events firing for this use case?

[gr2m]

yeah I don’t mind renaming that method to .destroy(), and renaming the clear event, too. We still need a clear event in https://github.com/hoodiehq/hoodie-client-store, but that is out of scope here