WIP: Manual schema management API

[davidmartos96]

Here is a proposed API for the manual schema management in the raw tables feature.

A new flag is added to the PowerSyncDatabase factories manualManagement, and a new function to mark the schema as ready db.markSchemaAsReady()

When using this manual mode, the developer is responsible for calling db.updateSchema(schema). Alternatively we could call automatically the updateSchema function in the markSchemaAsReady() function, but we believe it makes sense to be all manual and not mix.

While the schema is not ready, the public functions that are not related to executing SQL are guarded by _assertSchemaIsReady(), to prevent calling connect for example.

A working example with this is the branch example-raw-tables in https://github.com/davidmartos96/powersync.dart

Pending changes here are:

• Docs
• Pass the schema down to syncIteration() in the web worker implementation. We are not familiar with the communication protocol it uses.
• Tests

[simolus3]

I'm not super happy with this design, but I don't see a better option either.

One thing that looks cleaner might be to have users open their own databases (from sqlite_async), initialize the schema there and then pass the database to PowerSyncDatabase.withDatabase. But that would make the database tricky to use with wrappers like drift because we can't swap out the earlier database with the other one later.
So this might be the best option we have here, but I'll ping @rkistner for API design advice as well.

[davidmartos96]

One issue I can see with that is that the developer wouldn't be able to call powersyncDB.updateSchema() because they can only use the sqlite_async API at that point

[davidmartos96]

One thing we noticed while doing this PR is the logger param in PowerSyncDatabase.withDatabase actually says loggers. Technically it's a breaking change, but maybe you guys want to fix it in a separate commit