Training wheels for solidity
This is a Proof Of Concept that has not been formally verified, use at your own risk
- Training wheels for Solidity
- Security silver bullet that automagically makes your contract secure
- A replacement or competitor of Solidity: SafeSol is targeted at novice users who aren't aware of all the pitfalls and subtle security considerations of Solidity, experienced users should not use it.
- Built-in SafeMath: No more writting '.add()' or '.sub()', get automatic overflow/underflow protection without worrying about it!
- Non-strict equality checks on contract balance: If a contract does strict equality checks on its balance it's possible for an attacker to make these checks fail by sending 1 wei to the selfdestruct function. SafeSol solves that by turning all strict checks into non-strict checks.
- Conversion from tx.origin to msg.sender: According to Consensys' documentation, there's been some misuse of tx.origin to identify the transaction sender, which could lead to vulnerabilities. This feature prevents that by replacing tx.origin with msg-sender, which is safe to use for authentication.
- Force oversized types: Using bigger variables doesn't incur a lot of costs and can avoid many problems down the line, therefore SafeSol always uses big types. Linked with var resolution.
npm install safesolOriginally, this project started as a fork of the Solidity compiler, but, due to the constant changes in that codebase and the fact that, due to the way the compiler is structured some features of SafeSol were hard to implement, I moved away from that to a solution based on generating AST from the Solidity source code of a smart contract, operating on that AST to incorporate SafeSol features and then compiling the result. Unfortunately currently there doesn't exist any AST-to-Solidity code generator so right now SafeSol is using Solium/Ethlint, which allows for the creation of rules that operate on the source code using the AST as a reference. Operating on the source code directly is not a good solution, as it has previously introduced many subtle bugs and makes some features quite hard to implement, therefore the path forward for SafeSol is to develop an AST-to-Solidity generator and rewrite SafeSol using that.