MIME-Version: 1.0
In-Reply-To: <CADm_WcY8Vy+k+5BaBS+jV6D6tmSXrok8rAxoPxxKOzUhyPWgMg@mail.gmail.com>
References: <CADm_WcY8Vy+k+5BaBS+jV6D6tmSXrok8rAxoPxxKOzUhyPWgMg@mail.gmail.com>
Date: Tue, 22 Sep 2015 20:12:41 +0200
Message-ID: <CABm2gDoXa9ERY7iSsouxjypq1PwV_9HuBrtFQ_jrs5pGFst=KQ@mail.gmail.com>
From: =?UTF-8?B?Sm9yZ2UgVGltw7Nu?= <jtimon@jtimon.cc>
To: Jeff Garzik <jgarzik@gmail.com>
Content-Type: text/plain; charset=UTF-8
Cc: Bitcoin development mailing list <bitcoin-dev@lists.linuxfoundation.org>
Subject: Re: [bitcoin-dev] libconsensus and bitcoin development process
Precedence: list

On Tue, Sep 15, 2015 at 6:10 AM, Jeff Garzik via bitcoin-dev
<bitcoin-dev@lists.linuxfoundation.org> wrote:
> [collating a private mail and a github issue comment, moving it to a
> better forum]
>
> On libconsensus
> ---------------
> In general there exists the reasonable goal to move consensus state
> and code to a specific, separate lib.
>
> To someone not closely reviewing the seemingly endless stream of
> libconsensus refactoring PRs, the 10,000 foot view is that there is a
> rather random stream of refactors that proceed in fits and starts
> without apparent plan or end other than a one sentence "isolate
> consensus state and code" summary.
>
> I am hoping that
> * There is some plan
> * We will not see a five year stream of random consensus code movement
> patches causing lots of downstream developer headaches.
>
> I read every code change in every pull request that comes into
> github/bitcoin/bitcoin with three exceptions:
> * consensus code movement changes - too big, too chaotic, too
> frequent, too unfocused, laziness guarantees others will inevitably
> ACK it without me.
> * some non-code changes (docs)
> * ignore 80% of the Qt changes
>
> As with any sort of refactoring, they are easy to prove correct, easy
> to reason, and therefore quick and easy to ACK and merge.
>
> Refactors however have a very real negative impact.
> bitcoin/bitcoin.git is not only the source tree in the universe.
> Software engineers at home, at startups, and at major companies are
> maintaining branches of their own.
>
> It is very very easy to fall into a trap where a project is merging
> lots of cosmetic changes and not seeing the downstream ripple effects.
> Several people complained to me at the conference about all the code
> movement changes breaking their own work, causing them to stay on
> older versions of bitcoin due to the effort required to rebase to each
> new release version - and I share those complaints.
>
> Complex code changes with longer development cycles than simple code
> movement patches keep breaking.  It is very frustrating, and causes
> folks to get trapped between a rock and a hard place:
> - Trying to push non-trivial changes upstream is difficult, for normal
> and reasonable reasons (big important changes need review etc.).
> - Maintaining non-trivial changes out of tree is also painful, for the
> aforementioned reasons.
>
> Reasonable work languishes in constant-rebase hell, and incentivizes
> against keeping up with the latest tree.
>
>
> Aside from the refactor, libconsensus appears to be engineering in the
> dark.  Where is any sort of plan?  I have low standards - a photo of a
> whiteboard or youtube clip will do.

Just because you don't understand the changes proposed it doesn't mean
that they are random.
I may have done a poor job in communicating "my plan for libconsensus"
but I have tried many times and in many ways.
#bitcoin-dev logs show that I have not worked "in the dark" at all, on
the contrary, I've been very tenacious when asking for review and
opinions, to the point that several people (at least @laanwj and
@theuni have complained about their github inboxes being full of
"spam").
This is a relatively recent thread where I describe my plan:
http://lists.linuxfoundation.org/pipermail/bitcoin-dev/2015-July/009568.html
Not my first attempt on this list.

It is very frustrating that everybody seems to agree that separating
libconsensus is a priority to maximize the number of people that can
safely contribute to the project, but at the same time, nobody thinks
that reviewing the necessary refactors to do so is a priority.
I tried creating big PRs for people to see "the big picture" #5946 but
those were too many commits and nobody wanted to read it. Gavin asked
for an API.
So I tried a smaller step: exposing just VerifyHeader in libconsensus
and leave VerifyTx and VerifyBlock for later #5995
Again, this was "too big" and "a moving target". In the meantime I
always had smaller one-little-step PRs that were part of a longer
branch:

** [8/8] MERGED Consensus
- [X] Consensus: Decouple pow from chainparams #5812 [consensuspow]
- [X] MOVEONLY: Move constants and globals to consensus.h #5696
[consensus_policy0]
- [X] Chainparams: Refactor: Decouple IsSuperMajority from Params()
#5968 [params_consensus]
- [X] Remove redundant getter CChainParams::SubsidyHalvingInterval()
#5996 [params_subsidy]
- [X] Separate CValidationState from main #5669 [consensus]
- [X] Consensus: Decouple ContextualCheckBlockHeader from checkpoints
#5975 [consensus_checkpoints]
- [X] Separate Consensus::CheckTxInputs and GetSpendHeight in
CheckInputs #6061 [consensus_inputs]
- [X] Bugfix: Don't check the genesis block header before accepting it
#6299 [5975-quick-fix]
** [5/5] DELETED
*** DELETED Refactor: Create CCoinsViewEfficient interface for
CCoinsViewCache #5747 [coins]
*** DELETED Chainparams: Explicit Consensus::Params arg in consensus
functions #6024 [params_consensus2]
*** DELETED MOVEONLY: Move most of consensus functions (pre-block)
#6051 [consensus_moveonly] (depends on consensus-blocksize-0.12.99)
*** DELETED Consensus: Refactor: Separate CheckFinalTx from
main::IsFinalTx #6063 [consensus_finaltx]
*** DELETED Consensus: Refactor: Turn CBlockIndex::GetMedianTimePast
into independent function #6009 [consensus_mediantime]
*** DELETED Consensus: Adapt declarations of most obviously consensus
functions #6591 [consensus-params-0.12.99]
*** DELETED Consensus: Move blocksize and related parameters to
consensusparams ...without removing consensus/consensus.h [#6526
alternative] #6625 [consensus-blocksize-0.12.99]

After a while I stop rebasing the longer branches and just maintained
a few small consensus-related PRs at a time.

Now I consolidated 3 of them in

*** REVIEW Optimizations: Consensus: In AcceptToMemoryPool,
ConnectBlock, and CreateNewBlock #6445 [consensus-txinputs-0.12.99]

with the hope that it would be merged relatively fast.
After that it will be much simpler to start talking about potential C
APIs for VerifyHeader, VerifyTx and VerifyBlock; as well as separating
the library to a subtree.

I'm more than happy to answer any questions anyone may have about any
of the PRs or commits, until everybody interested is convinced that
there's nothing random in the proposed changes.
I'm also more than happy to get advice on how to better communicate my
plans and structure my PRs.