Re: [bitcoin-dev] BIP draft: HTLC transactions

[Johnson Lau <jl2012@xbt.hk>]

> On July 20, 2016 at 2:17 AM Luke Dashjr via bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org> wrote:
> 
> On Wednesday, July 20, 2016 5:46:54 AM Peter Todd via bitcoin-dev wrote:
> 
> > On Tue, Jul 19, 2016 at 10:35:39PM -0600, Sean Bowe via bitcoin-dev wrote:
> > 
> > > I'm requesting feedback for Hash Time-Locked Contract (HTLC) transactions
> > > in Bitcoin.
> > > 
> > > HTLC transactions allow you to pay for the preimage of a hash. CSV/CLTV
> > > can be used to recover your funds if the other party is not cooperative.
> > > These
> > > 
> > > scripts take the following general form:
> > >  [HASHOP]  OP_EQUAL
> > >  OP_IF
> > > 
> > > 
> > > 
> > > OP_ELSE
> > > 
> > >  [TIMEOUTOP] OP_DROP 
> > > 
> > > OP_ENDIF
> > >  OP_CHECKSIG
> > 
> > Note that because you're hashing the top item on the stack regardless
> > scriptSig's that satisfy HTLC's are malleable: that top stack item can be
> > changed anything in the digest-not-provided case and the script still
> > passes.
> 
> OP_SIZE
> OP_IF
>  [HASHOP] <digest> OP_EQUALVERIFY
>  <seller pubkey>
> OP_ELSE
>  <num> [TIMEOUTOP]
>  <buyer pubkey>
> OP_ENDIF
> OP_CHECKSIG
> 

This is incompatible with my proposal for fixing the OP_IF/NOTIF malleability in segwit ("MINIMALIF"). In this case only the timeout branch may be executed.

To make it compatible, you may use one of the following 2 scripts:

OP_SIZE OP_0NOTEQUAL
OP_IF
 [HASHOP] <digest> OP_EQUALVERIFY
 <seller pubkey>
OP_ELSE
 <num> [TIMEOUTOP] OP_DROP
 <buyer pubkey>
OP_ENDIF
OP_CHECKSIG

or

OP_IF
 [HASHOP] <digest> OP_EQUALVERIFY
 <seller pubkey>
OP_ELSE
 <num> [TIMEOUTOP] OP_DROP
 <buyer pubkey>
OP_ENDIF
OP_CHECKSIG

The overall witness size are the same for these scripts. They are 1 byte larger than Luke's script, in case MINIMALIF is not enforced.

(btw, the OP_DROP after TIMEOUTOP is missing in Luke's script)