Support point_evaluation_precompile when hardfork is cancun or later

[dajuguan]

Describe the feature

https://github.com/ethereum/EIPs/blob/master/EIPS/eip-4844.md#point-evaluation-precompile

Search terms

point evaluation precompile, 4844, eip4844, eip-4844, kzg proof

[fvictorio]

Thanks @dajuguan for reporting this.

Note for ourselves: a very minimal test is make an eth_call to 0x000000000000000000000000000000000000000a without any input data. Since the precompile checks that the input data is 192 bytes long, it should revert. Right now it doesn't.

We could also have a test it with a proper proof, but that would be like testing REVM itself, which I don't think makes sense.

[dajuguan]

Thanks @dajuguan for reporting this.

Note for ourselves: a very minimal test is make an eth_call to 0x000000000000000000000000000000000000000a without any input data. Since the precompile checks that the input data is 192 bytes long, it should revert. Right now it doesn't.

We could also have a test it with a proper proof, but that would be like testing REVM itself, which I don't think makes sense.

FYI, I've written a simple test case in js and python.

[fvictorio]

Ah, that could be useful, thanks!

[Wodann]

We released EDR v0.3.5 with a fix for this issue.

If you’re using the latest Hardhat release, in order to use EDR v0.3.5, you can remove your node_modules directory and npm/yarn/pnpm lock file and reinstall your dependencies. You can verify that you’re using the correct version of EDR by running the following command:

npm ls @nomicfoundation/edr

It should print 0.3.5.