debug_standardTraceBlockToFile returns only inner gas in gasUsed

[MysticRyuujin]

I'm trying to figure out how to use debug_standardTraceBlockToFile - but this output doesn't make any sense / is wrong. So I don't know if I'm doing it wrong, or it's broken 😄

$ curl -H "Content-Type: application/json" -d '{"jsonrpc":"2.0","method":"debug_standardTraceBlockToFile","params":["0x0627c2e829d606bcf19703d5c9f6967211ec2296f76e51f57124e2fc0e4e075c",{"txHash":"0x36ff2af6044075f1d37b090c6c07779bc7b4f05cdc731ba4fca4107f85e5686c"}],"id":1}' http://localhost:8545

{"jsonrpc":"2.0","id":1,"result":["/tmp/block_0x0627c2e8-250-0x36ff2af6-361212364"]}

$ sudo cat /tmp/block_0x0627c2e8-250-0x36ff2af6-361212364 
{"output":"","gasUsed":"0x0","time":436}

https://etherscan.io/tx/0x36ff2af6044075f1d37b090c6c07779bc7b4f05cdc731ba4fca4107f85e5686c

[MysticRyuujin]

Hmm I tried more tx with more stuff going on and they seem to output a lot more, but that one above is just weird. Unfortunately I don't have a Geth archive node to re-try THAT tx with. And I don't know enough to say if its working correctly or not...

[holiman]

Isn't it just a plain value transfer with no exec? If so looks correct to me. 

[MysticRyuujin]

Yes, it is, but gas used isn't 0? Or is this just another quirky Ethereum thing because ETH is special? Haha

[holiman]

Myeah, I think the gasUsed, as returned by the tracer, is the actual gas used by execution, excluding the "intrinsic gas" , i.e, the gas cost for the tx itself (21K) plus the cost for the input data, plus the cost for access list, if any.

That behaviour is indeed a bit wonky, so might deserve it's own ticket. Or at least changing the title.

[MysticRyuujin]

Thanks Martin.

I'm not looking to break compatibility or anything, I just didn't understand the output in this context, so I'll leave that decision to change the behavior up to y'all.

[rjl493456442]

Personally I think we should extend the tracer a bit (as well as the tracing spec https://eips.ethereum.org/EIPS/eip-3155) to include the "intrinsic gas". Although "intrinsic gas" is a bit ugly, but it's also a part of the consensus and can be wrong in different major client implementations. And also for better UX it should also be included.

We will discuss it in the next issue triage(tomorrow).