eth/tracers: replace hardcoded txGas with actual

[s1na]

So a few comments here:

• I needed isHomestead and isIstanbul values and didn't store them in the ctx map because that would expose them to js. So stored them in the Tracer struct instead
• My bindata output seems different. It might be because I'm using v3.1.2? which version do you use?
• Also I still need to rebase master to fix the merge conflicts

[holiman]

How about  setting 'intrinsicGas' as another param on the ctx, instead of adding it to gasUsed. The way you did it makes certain usecases work better, and other usecases regress (if one doesn't care about intrinsic, but just wants the gas used by the execution).Also it's a bit unintuitive that the go-method CaptureEnd gets a gasUsed as input, but that's not what the JsTracer sees. So I'd pefert a second field personally

[s1na]

How about setting 'intrinsicGas' as another param on the ctx, instead of adding it to gasUsed.

Makes sense, I'll do that

So I'd pefert a second field personally

Just to clarify, you're saying we should compute intrinsicGas in evm.go and pass it to CaptureEnd (in addition to the current gasUsed)? I can do that

[holiman]

Just to clarify, you're saying we should compute intrinsicGas in evm.go and pass it to CaptureEnd (in addition to the current gasUsed)? I can do that Actually I didn't mean that, I meant that the js thingy would calculate it, like you did, and set it inside the javascript-object. But the other way might work aswell, but have larger ramifications.

[s1na]

Actually I didn't mean that, I meant that the js thingy would calculate it, like you did, and set it inside the javascript-object. But the other way might work aswell, but have larger ramifications.

Sorry I had misunderstood and thought this is a different point from intrinsicGas in ctx.

Fixed that issue now, but I messed up the PR...I merged master into this branch to fix the conflicts but now it has a massive diff which makes it hard to review, sorry. Any ideas?

[holiman]

Fixed that issue now, but I messed up the PR...I merged master into this branch to fix the conflicts but now it has a massive diff which makes it hard to review, sorry. Any ideas?

Yep, here's what we do

git rebase -i eth/master 

Well, depending on how you have it configured. I recommend having eth or upstream as go-ethereum, and origin be your own fork.
After that,

git push -f

To push it to your branch

[s1na]

Thanks. I chose the easy way and used merge instead of rebase...I've now rebased but this PR still has the massive diff I think because it has a different history from Guillaume's branch. I created a fresh PR ethereum#22038 directly against master. Will close this afterwards.