This change makes use of uin256 to represent balance in state. It touches primarily upon statedb, stateobject and state processing, trying to avoid changes in transaction pools, core types, rpc and tracers.
This PR moves our fuzzers from tests/fuzzers into whatever their respective 'native' package is.
The historical reason why they were placed in an external location, is that when they were based on go-fuzz, they could not be "hidden" via the _test.go prefix. So in order to shove them away from the go-ethereum "production code", they were put aside.
But now we've rewritten them to be based on golang testing, and thus can be brought back. I've left (in tests/) the ones that are not production (bls128381), require non-standard imports (secp requires btcec, bn256 requires gnark/google/cloudflare deps).
This PR also adds a fuzzer for precompiled contracts, because why not.
This PR utilizes a newly rewritten replacement for go-118-fuzz-build, namely gofuzz-shim, which utilises the inputs from the fuzzing engine better.
* core/vm: set basefee to 0 internally on eth_call
* core: nicer 0-basefee, make it work for blob fees too
* internal/ethapi: make tests a bit more complex
* core: fix blob fee checker
* core: make code a bit more readable
* core: fix some test error strings
* core/vm: Get rid of weird comment
* core: dict wrong typo
* Experimental: EVMLogger fixed error code
* First pass of review
* Move error to `core/vm/errors.go`, ensure also we strictly implement `rpc.Error` (without depending on it)
Adding a space beween function opOrigin() and opcCaller() in instruciton.go.
Adding a space beween function opkeccak256() and opAddress() in instruciton.go.
This way, we avoid having a `cost` that is negative which does not make sense for a `uint64`. Having the `old, new` also yields correct value and the delta can then be negative and be holded in a `int64`.
# Conflicts:
# core/vm/evm.go
This way, we avoid having a `cost` that is negative which does not make sense for a `uint64`. Having the `old, new` also yields correct value and the delta can then be negative and be holded in a `int64`.
With this change, the full gas loop of a transaction can be tracked going from initial balance (`gasLimit`) back down to 0.
I validated on my regression test suite that those conditions applied to all transactions I have:
- First `OnGasConsumed` is always going from 0 -> `trx.gasLimit`
- Last `OnGasConsumed` "new value" is always 0
- trx.gasLimit - last balance of last `OnGasConsumed` is equal to `trx.gasUsed`
The addition of `reason` make it possible for a logger to filter out unwanted signal, for example one could want to drop all OpCode related gas change and only cares about "transaction" level. The reason is also a good thing for visibility and for creating powerful debug tools for Ethereum transaction execution.
Also, I would like that we change the `OnGasConsumed` logic from `OnGasConsumed(actual, cost uint64, reason)` to `OnGasConsumed(old, new uint64, reason)`. With the new tracing I've done, we now have negative `cost` that needs to be passed around but the `cost` is `uint64` so it creates overflow. Everything works fine if you record the new/old value `gas-cost` as the final value goes back to a valid range. But it's weird for consumer that would log `cost` that it's `18127127187219...`. An alternative would be to accept a `int64` for the cost which is also valid, not full coverage theorically but I don't see why an operation would cost more than i64 max value. Maybe it make more sense than to have signature `OnGasConsumed(gas uint64, cost int64, reason)`, I'm fine with it also, I leave the decision to you.
