chriseth

We should use a pragma to activate the standard library behaviour on a per-file basis. For this, we need a list of symbols that should be excluded from the global...

Please note that I think it makes sense to re-unify this with tuple type again. For now, it helps to find places where it is used as an inline array,...

Summary of the optimizer problems: There should be an issue about it but I currently cannot find it. The problem is the memory tracking inside DataFlowAnalyzer. For every memory write,...

I have a bad feeling allowing this. Unchecked overflow is a very different thing than unchecked array access. My hope is that the new optimizer will find most of these...

If you know the loop is short and want to optimize sloads radically, what about a fixed-size array?

The reason we do not allow "goto" is more that we would otherwise have to remove some optimizations we are currently able to do.

Is it in any way possible to compile this via yul and run it through an interpreter? Then we would at least not have to re-implement everything.

We are currently going the route of declaring functions at file level and then attaching them to structs. The encapsulation aspect, though, could still be considered.

To compare the two metadata structures, I provide both in their full form and re-use items that are identical in both structures by abbreviating them in angle brackets. current metadata:...

The idea is that multiple versions of evm bytecode can be introduced with limited backwards-compatibility breaks. We can at least be rather sure that no useful existing contract can have...