Added gaps at the end of the storage of each contract.

[nventuro]

Fixes #13.

Despite IERC721 and friends being contracts and not interfaces (due to a Solidity limitation), I kept this out of those, since they should be used as interfaces.

[nventuro]

The main reason behind 50 is that you'd need 5 in a row in the inheritance tree to go over 256, causing the PUSH1 to change to PUSH2, making the bytecode longer and deployment more expensive. If we bite the bullet and accept this (very small) extra cost, we could increase the gap size to e.g. 2000, future-proofing this implementation.

[spalladino]

We will be removing Initializable and loading it from zos-lib afterwards, right?

[nventuro]

Yes - should we add this there?

[spalladino]

I'd say we are fine with 50 slots per contract :-)

[spalladino]

I'd say the opposite, better keep Initializable as clean as possible

…

On Tue, Oct 9, 2018 at 6:34 PM Nicolás Venturo ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In contracts/Initializable.sol <#29 (comment)> : > @@ -46,4 +46,6 @@ contract Initializable { assembly { cs := extcodesize(address) } return cs == 0; } + + uint256[50] private ______gap; Yes - should we add this there? — You are receiving this because your review was requested. Reply to this email directly, view it on GitHub <#29 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAaOJAB2z0lTjWlC_EEnbfmUhy0WBClKks5ujNAlgaJpZM4XNZRq> .

[nventuro]

Any specific reason to not add the gap? I wouldn't discount the possibility of Initializable becoming more complex and needing more storage.

[coveralls]

Pull Request Test Coverage Report for Build 67

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 99.758%

---

Totals
Change from base Build 66:	0.0%
Covered Lines:	577
Relevant Lines:	577

---

💛 - Coveralls

[spalladino]

I was going to suggest backwards compatibility with the Initializable from 1.x, but that one used a single boolean variable, while the current one uses two. And this exactly proves your point. So, can you send a PR to zeppelinos to add the gap there as well?

[nventuro]

Sure thing! Opened OpenZeppelin/openzeppelin-sdk#215