The `sqlite` feature changes behavior of many tests but tests should be additive

[cactusdualcore]

The Cargo book states about features

A consequence of this is that features should be additive. That is, enabling a feature should not disable functionality, and it should usually be safe to enable any combination of features. A feature should not introduce a SemVer-incompatible change.

A lot of code branches on whether the "sqlite" or "sqlite-dynlib" features are enabled. This is not necessarily semver breaking, but most of that code that doesn't have a hard dependency the SqliteBackedHistory. Examples and even more importantly tests shouldn't alter behaviour like this (e.g. in history/cursor.rs).