feature: lockable mixin

[psomhorst]

While working on tests for ContinuousData, I was fixing some locking mechanisms. Since these will probably be reused in the future, I decided to write a mixin instead.

This mixin has been implemented and tested with ContinuousData only.

[psomhorst]

On naming: the Lockable mixin class and the lock(), unlock(), islocked() and islockable() methods might suggest that the entire object is made lockable. This is not the case. Attributes in the object are lockable. We could rename this to:

• lock_attr()
• unlock_attr()
• attr_is_locked()
• attr_is_lockable()

but that becomes quite verbose.

[DaniBodor]

On naming: the Lockable mixin class and the lock(), unlock(), islocked() and islockable() methods might suggest that the entire object is made lockable. This is not the case.

I think it's fine to use this nomenclature. I don't think the alternatives you suggest make it that much clearer. While it is formally speaking more accurate, I don't think it would be interpreted differently by a reader. We need to document well what locked (etc) means, or find a way to make the entire object lockable.