Add IAragonApp interface

[bingen]

Fixes #587
See: #586 (comment)

[coveralls]

Coverage increased (+0.003%) to 98.505% when pulling 122850f on aragon_app_interface into bf226a2 on next.

[facuspagnuolo]

Should we leverage this change and make these external?

[bingen]

Why? Those functions are being called by other contracts and in Solidity 4 we can’t override from external to public.

[facuspagnuolo]

I don't see a problem in introducing a breaking change if we are going to use semver properly to release a breaking version, if those contracts want to use this new version then they must update their usage

[bingen]

Ok, but what’s the advantage of making them external?

[facuspagnuolo]

Well, it is the expected way to define interfaces in solc... you won't notice a warning here because it is defined as a contract, but otherwise it would complain. However, not a blocker at all, we still need to define what do we want to include in the next breaking version, so we can discuss whether to address this or not

[facuspagnuolo]

LGTM!

[facuspagnuolo]

Well, it is the expected way to define interfaces in solc... you won't notice a warning here because it is defined as a contract, but otherwise it would complain. However, not a blocker at all, we still need to define what do we want to include in the next breaking version, so we can discuss whether to address this or not

[sohkai]

Have a couple more comments... maybe it would be worth discussing the abstractions offline :).

It feels natural that the supportsInterface method lives where the rest of the interface is defined

I wouldn't necessarily say so; ultimately the implementation contract has to decide which interfaces it'd like to publicly advertise. I also wouldn't couple an interface with ERC165, especially as a lot of ERC interfaces only suggest that a contract support ERC165.