It doesn't replace finalize, but its the recommended way of handling resources. IMO, anyone with java experience would understand this, and treat this class as such. we can make it clear on the release notes and the comment

requested changes

Forgot to add this last minute change

make requested changes and implement Autocloseable into the class

Make requested changes

I responded to some of those comments and made the requested changes regardless. we can roll some of them back. as for isEnabled name im open to suggestions, i cant think of something more concise ;). Also, i added isValid to indicate the instance has a valid context. as you said, users dont need to know what context is, only if it works or it doesnt

clean all resources in test

make requested changes and add proposed ones

make requested changes and add proposed ones

rolling back diff (hopefully)

arc diff just did somethign strange to that py file. is there a way to revert that?

make requested changes and also add the proposed ones

the context is set to -1 at "cleanup" (line 200~) when the context in the native lib is destroyed. We can set it up to -1 before assigning it at the constructor call. that way we can handle the case where the exception thrown at construction is caught for some dumb reason. Perhaps we should also check if the context is not -1 before every lib call. Theres also fixing cloneContext, which IMO should result in returning a new NativeSecp256k1 class with the cloned native lib context in, but i was hoping to fix that on another ticket

Remove Secp256k1Context and move the loading to NativeSecp256k1

Roll back formating changes

checking if library is enabled, otherwise throw exception

In D5490#132848, @deadalnix wrote:

The getContext used to have the benefit of checking that the context is properly initialized. Now there is no such check, so if cleanup is called, then something else is, you'll get an horrible crash instead. I do not think this is somethign java devs expect.

Fix spacing issue

fixing typo