1 2016-10-13 17:22:36 <arubi> jl2012, re. bip-mastopcodes, bip114, how does 'scriptcode' in signing midstate look when a check(multi)sigverify is done?  is it the serialized script within the current executing branch?  the entire ser. script concatenated?  hashes of branch/path?   I love script+MAST.  it makes so much sense to do it this way :)
 2 2016-10-13 17:37:32 <arubi> also, I have an urge to try and use some interesting scripts that are now possible with CAT, SUBSTR..  would you say that it makes sense to make some generalized, long, re-occurring parts of script (e.g. make a DER sig from R and S in bip-mastop) across branches into their own branch and treat it like a function that could appear in a path?  I'm really excited for this bip.
 3 2016-10-13 17:44:40 <jl2012> arubi, I'm making a new version, but stopped since I need to complete #8499 first
 4 2016-10-13 17:45:18 <arubi> jl2012, ah.  so it's not v2 or v3?  something completely new?
 5 2016-10-13 17:45:26 <jl2012> v3
 6 2016-10-13 17:45:36 <arubi> oic, and that has no docs yet, correct?
 7 2016-10-13 17:45:52 <jl2012> with doc, just not pushed. wait a second
 8 2016-10-13 17:46:09 <arubi> sure
 9 2016-10-13 17:46:14 <instagibbs> jl2012, don't DoS the BIP numbering please ;)
10 2016-10-13 17:47:16 <jl2012> instagibbs: i'll update 114
11 2016-10-13 18:08:42 <jl2012> arubi: https://github.com/jl2012/bips/blob/bip114v3/bip-0114.mediawiki
12 2016-10-13 18:09:09 <arubi> jl2012, cheers!  I'll build v3 and go over it.  thanks!
13 2016-10-13 18:09:24 <jl2012> code is incompete
14 2016-10-13 18:09:34 <jl2012> the link in the bip is wrong
15 2016-10-13 18:10:25 <jl2012> if you just want to use OP_CAT, try v2. But the script is a little incompatible with v3
16 2016-10-13 18:10:39 <jl2012> also, the checkmultisig in v2 is broken
17 2016-10-13 18:11:18 <arubi> hm.  okay, I'm planning on only using checksigverify for signatures, and I could probably figure out what to change in v3
18 2016-10-13 18:11:47 <arubi> or rather, what to take into account beforehand
19 2016-10-13 18:11:54 <jl2012> you just need an additional empty witness stack item to migrate from v2 to v3
20 2016-10-13 18:12:38 <arubi> also keyScriptCode_K now appears on the stack, which is important
21 2016-10-13 18:12:52 <arubi> if I'm using checksig that is, I guess!
22 2016-10-13 18:12:54 <jl2012> and there was a bug in v2 which is fixed in v3. v3 requires that each subscript (now called keyScriptCode) must be valid standalone
23 2016-10-13 18:13:24 <arubi> ah so no longer subscripts have to returng nothing and not fail?
24 2016-10-13 18:13:48 <jl2012> in v2, all subscripts are merged as 1 script
25 2016-10-13 18:13:57 <jl2012> in v3, they are run one-by-one
26 2016-10-13 18:14:08 <jl2012> the final result must be empty stack
27 2016-10-13 18:14:19 <arubi> I see!  that's what I thought v2 was about, nice.
28 2016-10-13 18:36:35 <arubi> that signature thing that's done in bip-mastop is something I've been interested in for some time now, waxwing, remember us and proslogion were discussing this a while back..  can you sign off a transaction to a scriptpubkey that is only redeemable to a specific set of outputs?  it's tricky, I though I was getting close it in 0895E97E9C4CE7EBE04E15E0835BB0788053FBFDBBB2F3F25F81631687D7B857 (testnet), redeeming a tx that has a self signed sig in
29 2016-10-13 18:36:36 <arubi> it, but it's essentially an anyonecanspend.  you _have_ to use the committed sig, but pubkey is free for all!
30 2016-10-13 18:37:17 <arubi> having CAT, SUBSTR..  I'm trying to figure out if producing such an output that doesn't depend on a prev txid, that can only be redeemed by paying specific amounts to specific outputs.  sounds impossible, and it is with current script, but it remains that if we could figure out a way to sign an actual complete transaction which "only" lacks a prev txid and (maybe!) input index, then it might be possible to create a such a scriptpubkey that commits
31 2016-10-13 18:37:17 <arubi> to 'r', 's' and 'pubkey', and by that also commits to 'd' and 'k'.  anyway, I'm really thankful for this mast bip.  really awesome work
32 2016-10-13 18:49:07 <arubi> sighash single bug came close too, but sadly "one" is the whole sighash, not just the prevtixid, and the bug is fixed in segwit
33 2016-10-13 18:51:31 <arubi> so really, since there is no way to have only the prevtxid+index missing from the sighash midstate, and assuming you are only going to pay to that redeemed scriptpubkey once (security), there is no actual way to do this yet