Reorder evmc_message fields

[chfast]

Resolves #123.
Resolves #122.

[axic]

While this isn't changed in this PR, but I think this comment is invalid, it can be create even for depth 0.

[axic]

So didn't #123 say that move code_hash from message to execute?

[chfast]

So didn't #123 say that move code_hash from message to execute?

Looks like it is not need at the moment, unless WAVM needs it. It was never taken from the database, was computed on the fly so VM is capable of doing the same.

[axic]

Looks like it is not need at the moment, unless WAVM needs it. It was never taken from the database, was computed on the fly so VM is capable of doing the same.

With the extcodehash opcode in Constantinople will that stay like that?

[chfast]

The extcodehash is not useful here because you don't know the address of the code.

[axic]

The extcodehash is not useful here because you don't know the address of the code.

Isn't #123 saying that execute() should have a code_hash parameter? The code is certainly known there.

[chfast]

Yes it says that, but now I prefer to skip it.

[axic]

I don't need it myself so OK to remove it.

[chfast]

I prefer add features when there is use for them. And execute() together with message is a ton of params anyway.