12020-08-26T02:39:12  <jeremyrubin> roconnor: unclear do you want it to work with bare keys?
 22020-08-26T05:05:21  <aj> roconnor: the path proposed in PR 943 is taking over an unknown pubkey type, yeah. (an alternative approach would be taking over witness version 16, in which case it could work via the key path as well, but that seems unnecessary/premature to me)
 32020-08-26T05:07:21  <aj> roconnor: practically, i think we're just about at the point where taproot is usable on signet (see my signet-tr branch on github), so once that works, i'd like to have signet support anyprevout as well, at which point we should be able to test out eltoo and do non-theoretical attacks against anyprevout to see if it all makes sense
 42020-08-26T05:07:54  <jeremyrubin> aj: can I PR CTV on signet too?
 52020-08-26T05:08:09  <aj> luke-jr: not today, but that's the goal, yeah
 62020-08-26T05:08:14  <aj> luke-jr?
 72020-08-26T05:08:23  <aj> jeremyrubin: not today, but that's the goal, yeah
 82020-08-26T05:09:05  <aj> jeremyrubin: (and you can certainly setup your own signet fork with CTV enabled today, of course)
 92020-08-26T05:09:41  <jeremyrubin> yeah may do this... could be fun to set up an amazon signet AMI so you can click-deploy a new network
102020-08-26T05:10:04  <jeremyrubin> if I'm feeling like it maybe I'll put it together
112020-08-26T05:11:16  <luke-jr> aj: ?
122020-08-26T05:11:48  <aj> luke-jr: misdirected
132020-08-26T05:12:57  <luke-jr> aj: did I remember correctly, that you had argued omitting hash_type was justified by a reduction in CPU time?
142020-08-26T05:13:40  <aj> luke-jr: was going to say: if you've got many inputs to a tx spending via SIGHASH_ALL means fewer hash recalculations than the other options, so justifies saving a byte. being able to spend an extra byte with the explicit sighash_all mode i don't have a real justification for
152020-08-26T05:14:16  <luke-jr> well, can always pad junk data other ways too, so no point stopping that
162020-08-26T05:14:36  <luke-jr> if you have all SIGHASH_SINGLE (for example), though, isn't the sum hashing identical to SIGHASH_ALL for all of them?
172020-08-26T05:15:07  <luke-jr> (ignoring that the current code precalculates SIGHASH_ALL unnecessarily in this case)
182020-08-26T05:19:03  <aj> luke-jr: if you're doing SINGLE, you have to generate sha_single_output for each input as an additional hashing op compared to doing SIGHASH_ALL
192020-08-26T05:19:35  <luke-jr> aj: but that's *instead* of hashing *all* the inputs for SIGHASH_ALL
202020-08-26T05:19:55  <luke-jr> so you have 1+1+1+1 rounds instead of 4 rounds in a single hash
212020-08-26T05:20:51  <aj> luke-jr: right, for 10 ALL vs 10 SINGLE; for 10 ALL vs (1 ALL plus 9 SINGLE) you get the saving
222020-08-26T05:21:40  <luke-jr> for 1 ALL + 9 SINGLE, the additional cost is in the ALL, not the SINGLEs :P
232020-08-26T08:04:37  *** jonatack has quit IRC
242020-08-26T08:08:57  *** midnight has quit IRC
252020-08-26T08:32:43  *** digi_james has quit IRC
262020-08-26T08:36:02  *** fjahr has quit IRC
272020-08-26T08:39:13  *** digi_james has joined ##taproot-bip-review
282020-08-26T08:39:32  *** belcher has joined ##taproot-bip-review
292020-08-26T08:39:33  *** fjahr has joined ##taproot-bip-review
302020-08-26T09:03:46  *** jonatack has joined ##taproot-bip-review
312020-08-26T10:22:05  *** jonatack has quit IRC
322020-08-26T13:36:14  *** jonatack has joined ##taproot-bip-review
332020-08-26T13:38:27  *** midnight has joined ##taproot-bip-review
342020-08-26T14:01:36  *** xoyi- has joined ##taproot-bip-review
352020-08-26T15:02:44  <roconnor> aj: what is PR 943?
362020-08-26T17:30:36  *** belcher_ has joined ##taproot-bip-review
372020-08-26T17:34:07  *** belcher has quit IRC
382020-08-26T18:27:32  *** xoyi- has joined ##taproot-bip-review
392020-08-26T20:01:32  *** xoyi- has quit IRC
402020-08-26T22:03:57  *** ghost43 has quit IRC
412020-08-26T22:04:15  *** ghost43 has joined ##taproot-bip-review