Daniel

So what's your current plans around this (since we had at least a very brief discussion about this at devconnect)? In general, the ABI should really be specified as a...

Thank you for the PR! This is mainly missing (a lot of) semantics tests. We have to make sure that, under any circumstances, the correct functions are called in the...

> Sure. Let me know what other semantic tests you'd like and I'll work on them. I just tried to :-). But let me try to be clearer :-): The...

Ah, right, I must have missed that one semantics test, sorry about that. But yeah, testing is the main issue with this, especially since, as you say, the effects of...

@matheusaaguiar Just to make this airtight, can you add the following proof scripts as sanity checks: This checks that ``BVSignedCleanupFunction`` actually matches what ``YulUtilFunctions::cleanupFunction`` does, and that both of them...

@bshastry Do we have infrastructure set up to easily fuzz develop pre- and post this change before our next release :-)? Affected is any solidity code with effects that depend...

@bshastry Yes, this will affect legacy *and* IR code generations and it will indeed kick in *before* optimization. So the only thing one could do here would be differential fuzzing...

Yeah, ok... can the fuzzers spawn processes? Then using two separate solc binaries and haviung them produce bytecode and then feeding those two bytecodes to evmhost separately (without actually ever...

For debugging the optimized evm code generation path, it may be good to know that you can visualize the control flow graphs and stack layouts by running our tests: You...

I've, out of curiosity, just built the latest PR version with additionally my last three comments addressed - and that version in fact passes all of our semantics tests, so...