Deprecation warning for SELFDESTRUCT

[ekpyron]

Following https://eips.ethereum.org/EIPS/eip-6049 we should emit deprecation warnings on any use of selfdestruct.

[morganjweaver]

Hello there. I'm a newer Solidity dev with a C++ and compilers background. I've found the Contributing to Solidity 101 doc, the Workflow for PRs doc, and the PR Review Checklist doc, and intend to submit a PR. Please let me know if there is anything not covered in the documentation that I should know. I've also got the gitter solidity-dev channel open. Thank you!

Also: since the callcode deprecation warning is most similar, I wanted to verify that I should include inline assembly in the check + warning, unlike callcode--is this correct?

[ekpyron]

Hi there and thank you very much for wanting to contribute!

Yes, callcode may be a good reference for emitting the deprecation warning on the solidity level - that was done a long time ago and the code base has changed a lot since, but the StaticAnalyzer should still work pretty much the same.

And indeed there should also be a warning about selfdestruct in inline-assembly, resp. Yul - the place to look for that would be libyul/AsmAnalysis.cpp - but we can also start with a PR that only addresses Solidity and take it from there!

Feel free to ask for any help you may need here or in the gitter/matrix channel - unfortunately, it seems like we're in different timezones, so we may have to rely on asynchronous communication.

I hope https://docs.soliditylang.org/en/develop/contributing.html will be a good enough reference to get you started e.g. on running the test suite. https://docs.soliditylang.org/en/develop/contributing.html#writing-and-running-syntax-tests and running isoltest should be the most relevant for this issue.

[ekpyron]

@morganjweaver Any update on this? Do you need any more input and/or help?

[nikola-matic]

@morganjweaver I'll be taking this over since we plan on releasing soon, and we'd rather like for this to go into the 0.8.18 release.

[cameel]

And indeed there should also be a warning about selfdestruct in inline-assembly

Maybe it would be better not to issue a warning for its use in inline assembly? At the moment the opcode is still in the EVM so if you need and want its functionality, using selfdestruct() is still the way to go. There is no alternative construct available now and there's no way to silence this warning.

[nikola-matic]

And indeed there should also be a warning about selfdestruct in inline-assembly

Maybe it would be better not to issue a warning for its use in inline assembly? At the moment the opcode is still in the EVM so if you need and want its functionality, using selfdestruct() is still the way to go. There is no alternative construct available now and there's no way to silence this warning.

The draft PR only issues a warning when used from Solidity, so we're good so far if that's the case.

[cameel]

Yeah, I saw that. That's good for now but it was stated here that we should add it in inline assembly too so I wanted to get a decision on that.

[ekpyron]

Definitely also warn in assembly. The opcode is what's being deprecated.

[trendespresso]

Definitely also warn in assembly. The opcode is what's being deprecated.

As an EVM developer a little out of the know: What should I replace selfdestruct() with? In the case of a payable() inside, how should that look: selfdestruct(payable(msg.sender)) --> ??

Thanks for any info! Appreciate everyone who works on Solidity
Cheers

[trendespresso]

If we want to destroy contracts, how should we do so now that selfdestruct() is deprecated?

[Tshembani04]

I used solidity: "0.8.17", after trying to get anything resourceful from the Ethereum docs and chatGPT