utACK 1e1db2f

Updated adding utilmoneystr, which I forgot will be needed in libconsensus for this line: https://github.com/bitcoin/bitcoin/blob/master/src/main.cpp#L1961
(although if many people agree with me, I would prefer to keep it out and just print satoshis).

ACK c078fd1

I really don't think that we should be moving everything that libconsensus is code-dependent on into the consensus directory. That does not help clarity of the code base. Things like serialization and basic data structures in primitives probably deserve to be their own library, which libconsensus links against. Moving them all into the consensus directory doesn't help IMHO.

I'm all ears for suggestions on other internal packages for the consensus one to depend on besides crypto and libsecp256k1. I think this approach does help the clarity of the codebase by explicitly putting together all the files included (or that are planned to be included) in the same building package (consensus). For example, when I ask "please don't include this header from a consensus file", it would simplify things (not everybody looks at src/Makefile.am).
Also if we plan for libconsensus to become its own repository eventually, I see this as the most straightforward path. Some people on IRC have expressed that maybe there could be some intermediary libraries that both libconsensus and bitcoin core depend on.
So it seems there's mixing feelings about this and in any case this can be done at any time (in fact, it was phase 4 in my newer published plan).
Therefore I'm closing the PR.

Regarding the separation of utilmoneystr in a different commit, it was because I believe it shouldn't be part of libconsensus. At least @sipa seems to agree with me there, so at some point before exposing verifyTx in libconsensus I will create a PR removing the one line dependency and ping you both.