Kamil Śliwak

With the warning/error the idea is that such getters are inherently misleading and you should not have them. I.e. when you get the warning, you're better off switching from a...

I pulled the "Disallow suffixes with storage and calldata return values" part into the base PR. I'm still reviewing the rest of the PR.

> - [ ] Include suffix definition as a secondary location in errors. Some hints on this since I just had to do this in the operators PR. To add...

I rebased this together with the base PR. Both are now on top of the operators PR that is currently in review.

I'm pulling the commits for signed exponent check and array size check into the base branch.

Also, looks like the tests with huge exponent massively increase the running time of `soltest_all` (from 5 min to 30+ min). Given that it's very parallelized (has 50 parallel jobs),...

The syntax and type checking for suffix definitions is now done in the base branch. Now definitions need the `suffix` modifier. Some locations in error messages shifted due to this,...

> - [ ] Fix: Suffixes not detected as function calls by control flow graph (no "unreachable code" warning). Here's a hint about how to tackle this point. It refers...

Not sure what happened here, but this was supposed to be closed, not merged into the base branch.

Just ran into it myself. We really need to fix it. ```solidity interface I { function f(uint) external returns (uint); } function f(uint) returns (uint) {} ```