--- Day changed Thu Oct 01 2020
00:45 -!- jonatack [~jon@37.164.132.126] has joined ##miniscript     
01:07 -!- Aleru [sid403553@gateway/web/irccloud.com/x-yujxdzfveoxjwazt] has joined ##miniscript     
01:07 -!- felixweis [sid154231@gateway/web/irccloud.com/x-krefugjevbfmhthz] has joined ##miniscript     
01:52 -!- Netsplit *.net <-> *.split quits: Aleru, felixweis
02:10 -!- Netsplit over, joins: Aleru, felixweis
03:04 -!- jb55 [~jb55@gateway/tor-sasl/jb55] has quit [Ping timeout: 240 seconds]     
03:05 -!- jb55 [~jb55@gateway/tor-sasl/jb55] has joined ##miniscript     
03:15 -!- jonatack [~jon@37.164.132.126] has quit [Ping timeout: 260 seconds]     
05:21 -!- jonatack [~jon@2a01:e0a:53c:a200:bb54:3be5:c3d0:9ce5] has joined ##miniscript     
05:33 < andytoshi> jeremyrubin: i thought we had such a thing
05:34 < andytoshi> https://docs.rs/miniscript/1.0.0/miniscript/policy/semantic/enum.Policy.html unsatisfiable and trivial
05:35 < andytoshi> oh they don't exist on the compiler policy type
05:36 < andytoshi> yeah i think they probably should, it does make composeability hard for those if there is no initial/terminal object
08:46 < jeremyrubin> andytoshi: https://github.com/rust-bitcoin/rust-miniscript/pull/138
08:58 < andytoshi> thanks
08:58 < andytoshi> fyi in the past i have disagreed with sipa about this, and it means that the rust-miniscript concrete policy language differs from sipa's
09:06 -!- jonatack [~jon@2a01:e0a:53c:a200:bb54:3be5:c3d0:9ce5] has quit [Quit: jonatack]     
09:08 < jeremyrubin> interesting
09:09 < jeremyrubin> btw I'm not quite sure how to debug the failing fuzzer
09:12 < andytoshi> yeah for some reason we removed the fuzz unit tests
09:13 < andytoshi> there are insrtuctions in fuzz/README
09:13 < andytoshi> lol hmm apparently i acked this
09:14 < andytoshi> https://github.com/rust-bitcoin/rust-miniscript/pull/95/commits/351576ae2a30bd2ef62a9c6f63c05d133d300805
09:14 < andytoshi> we should probably just revert that commit, it makes the fuzz tests so much more annoying to use
09:15 -!- jonatack [~jon@2a01:e0a:53c:a200:bb54:3be5:c3d0:9ce5] has joined ##miniscript     
09:19 < sanket1729> jeremyrubin: Added a comment for fixing fuzz issue
09:27 -!- roconnor [~roconnor@host-104-157-196-197.dyn.295.ca] has quit [Ping timeout: 260 seconds]     
09:35 < jeremyrubin> sanket1729: to be clear some other patch should change that right
09:35 < jeremyrubin> it's not b/c something I introduced
09:36 < sanket1729> Earlier it was because compiler failed on compiling TRIVIAL policy 
09:36 < sanket1729> Because it marked trivial as unreachable node
09:37 < jeremyrubin> Ah i see.
09:37 < sipa> i thought we already had true/false, but apparently only in miniscript itself, not in the policy?
09:37 < sanket1729> We had that in semantic policy, but not in concrete one
09:38 < sanket1729> I am happy that the fuzzer works :P 
09:40 < jeremyrubin> sanket1729: is there any other bad ramification from adding the from_true?
09:41 < sipa> andytoshi: i'm not too worried about policy languages differences... the whole point of miniscript is that it shouldn't matter what compiler (or source language) people use
09:41 < sipa> though it's useful for cross testing
09:41 < jeremyrubin> sipa: is there a reason you think Policy shouldn't have trivial/unsat?
09:41 < jeremyrubin> curious for the reasoning
09:42 < sipa> no, i don't see why not
09:42 < sipa> i assumed we already had them
09:43 < sipa> the internal optimizer even has it, apparently; there is just no syntax mapping to it
09:43 < sanket1729> jeremyrubin: I think we had initially as a form of sanity check. 
09:44 < sanket1729> I cannot think anything bad happening because we compiled a true policy
09:49 < andytoshi> yeha i thnk it was just a question of usefulness
09:50 < andytoshi> jeremyrubin: can you change the PR though to use lowercase names?
09:50 < jeremyrubin> andytoshi: i did it to be the same as in semantic policy
09:50 < andytoshi> o.o     
09:51 < andytoshi> ok never mind, it's fine then
09:51 < andytoshi> but i wonder why on earth i did that
09:51 < jeremyrubin> :shrug:
09:52 < jeremyrubin> could clean up both as follow up work, but it would be a breaking change :(
09:52 < jeremyrubin> I guess you could add another alias?
09:55 < andytoshi> i think we should just do a breaking change :P
09:55 < andytoshi> we're going to release a new version anyway in the next day or two
10:02 < andytoshi> jeremyrubin: needs rebase
10:25 < andytoshi> darosior: on #131 can you make sure that git rebase -x 'cargo test --features "compiler"' passes?
10:25 < andytoshi> i.e. rebase to fix the compilation errors 
10:26 < sipa> woah
10:26 < sipa> TIL rebase -x
10:26 < andytoshi> haha wow, your life will be changed
10:33 < darosior> andytoshi: rebased
10:33 < andytoshi> thanks
10:35 < darosior> Hmm do you want bisectability for the tests on each commit ?
10:35 < andytoshi> yeah
10:35 < darosior> Somehow i messed up one of the rebase of this PR last week, i can fix it now..
10:36 < andytoshi> though i don't usually check with feature flags on ... i happened to this time because i noticed some unit tests didn't make sense visually and then noticed they were feature-gated on the compiler
10:41 < darosior> andytoshi: Yeah the derive() tests ended up in the first commit. Fixed now :)
10:44 < andytoshi> thanks :) yeah, that's what caught my eye
10:44 < andytoshi> i was trying to figure out where the old implementation of `derive` had gone, since i know it was removed and it didn't appear in any of the diffs
10:44 < andytoshi> and yet it was called from a test..
10:46 < andytoshi> ok gimme a couple minutes to get my local rebase -x finished
10:46 < andytoshi> and i'll merge
10:47 < sanket1729> yep. -x is quite handy. I wonder if we can add that to the CI. 
10:49 < aj> MAKEFLAGS=-j12 git rebase -x 'make && make check' origin/master # might be a fun new alias for me
10:49 < sipa> andytoshi: y u no 'make -j12 check'
10:50 < sipa> eh     
10:50 < sipa> aj: ^
10:50 < aj> ehj     
10:50 < aj> i usually do make -j12 && make -j5 check
10:50 < andytoshi> lol, well, the reason i don't is because some projects (including 2015-era bitcoin, iirc) do bad things with -j
10:50 < sipa> andytoshi: in bitcoin core you need 1G ram per -j
10:51 < aj> but i like to see the build succeed before any tests start, but sure
10:51 < sipa> approximately
10:51 < sipa> i think it only starts checks once the build succeeds anyway
10:51 < sipa> if it doesn't, i've never noticed
10:51 < andytoshi> the problems i remember involved building thigs in the wrong order, not ram limitations-
10:51 < sipa> ah     
10:51 < aj> never tried. it doesn't build the benchmarks until check though :(
10:52 < sipa> andytoshi: that may have been the case long ago
10:52 < andytoshi> though i assume this hasn't been a problem in years, -j is so popular
10:52 < andytoshi> yeah
10:52 < andytoshi> sanket1729: ok lets get 119 in and then we can do a release
10:52 < sipa> yeah, i can do make -j30 fine now, on systems with enough ram
10:53 < andytoshi> or do you think we should do #116? it's marked WIP still and i don't recall where we landed on the design discussion
10:53 < andytoshi> sanket1729: can you rebase #119
10:53 < sanket1729> rebasing 119
10:54 < andytoshi> dope
11:01 -!- Netsplit *.net <-> *.split quits: Aleru, felixweis
11:17 < sanket1729> andytoshi: Rebased and pushed. Awaiting CI
11:17 < sanket1729> can you look at discussion here? https://github.com/rust-bitcoin/rust-miniscript/pull/119#discussion_r487689431
11:24 < andytoshi> sanket1729: i'm confused, how could a general purpose `finalize` method work?
11:24 < andytoshi> in rust-bitcoin
11:25 < sanket1729> no, rust-bitcoin does not a generic finalizer
11:25 < andytoshi> you say " I think there definitely should be a finalize method in rust-bitcoin that works for all scripts. "
11:25 < sanket1729> It only has method to set final_witness and final_scriptsif
11:25 < andytoshi> ah     
11:25 < sanket1729> I mean, scripts that not miniscripts
11:26 < sanket1729> and rust-miniscript has a "smart" finalizer that works for subset(miniscript) which outputs final_scriptsig and final_witness
11:27 < andytoshi> i'm a bit confused as to what the point of this API is in rust-bitcoin ... you have a mehtod which takes a witness/scriptsig and just puts those in the appropriate places in the hsahmap
11:27 < andytoshi> why can't callers (and rust-miniscript) directly put the fields in the hashmap
11:28 < sanket1729> Conceptually finalizing is exactly putting those in places
11:28 < sanket1729> not figuring out what to put
11:29 < sanket1729> I guess what I mean is the role of finalizer belongs to rust-bitcoin?
11:29 < sipa> finalizer is something that's supposed to understand the scriptPubKey being signed for (or otherwise determine if enough information is provided to construct a witness)
11:30 < sipa> it's separate from extractor, as the extractor can be dumb
11:31 < sanket1729> What about things that are not miniscript? There cannot be a finalizer for it?
11:31 < sipa> you need to use a finalizer that understands what the scripts do
11:32 < sipa> if there are miniscript constructions being signed, then that means you need a miniscript-capable finzlier
11:33 < sipa> you'll have at least one party present that does in this case, as the updater and/or signer need to understand it too
11:34 < sipa> but you could have a single transaction that consists of say miniscript inputs, and CrazyWallet inputs (which is something miniscript like, but not compatible with it, and no software exists that supports both miniscript and CrazyWallet)
11:35 < sipa> in that case you'll need to finalize with a miniscript finalizer and a CrazyWallet finalizer, combine, and extract
11:36 < sanket1729> makes sense
11:40 < sanket1729> andytoshi: agreed. Let's the API as it is then finalize(&mut Psbt) -> Result<(), Error>
11:40 < sanket1729> Btw, CI is psasing on #119
11:47 -!- Aleru [sid403553@gateway/web/irccloud.com/x-yujxdzfveoxjwazt] has joined ##miniscript     
11:47 -!- felixweis [sid154231@gateway/web/irccloud.com/x-krefugjevbfmhthz] has joined ##miniscript     
12:02 < andytoshi> 18:31 < sanket1729> What about things that are not miniscript? There cannot be a finalizer for it?
12:03 < andytoshi> yes :) this was one of the original motivations for miniscript
12:03 < andytoshi> arguably -the- original motivation
12:05 < andytoshi> ok i'll do a(nother) full review of 119 now
12:05 < andytoshi> and hopefully we can get it in :)
12:05 < andytoshi> and then do a new release
12:05 < sanket1729> awesome
12:07 < andytoshi> and then we can bump msrv to 1.29, update rust-bitcoin dep, and clean up some rust 1.22-isms
12:19 < andytoshi> o.o what is `run-parts`
12:19 < andytoshi> is that a standard utility?
12:20 < sanket1729> yeah
12:20 < sanket1729> It runs all scripts in a directory
12:20 < andytoshi> nice
12:21 < andytoshi> this `interpreter_check` method is pretty cool :P
12:24 < sanket1729> :)      
12:37 -!- harrigan [~harrigan@ptr-93-89-242-235.ip.airwire.ie] has quit [Ping timeout: 240 seconds]     
12:38 < andytoshi> sanket1729: ok, review done
12:38 < sanket1729> awesome. addressing it
12:38 < andytoshi> awesome
12:39 -!- harrigan [~harrigan@ptr-93-89-242-235.ip.airwire.ie] has joined ##miniscript     
13:16 < sanket1729> fixed 119
13:21 < andytoshi> ok just running the rebase check agian
13:21 < andytoshi> sanket1729: btw should we merge jeremyrubin's pr
13:22 < andytoshi> which uses btreemap in the campiler
13:22 < sanket1729> I am not against it. 
13:22 < sanket1729> I think that maybe helpful in testing. 
13:22 < sanket1729> based on the previous discussion, I think we should not make effort to *gurantee* determinism in rust-miniscript compiler project. 
13:23 < sanket1729> https://github.com/rust-bitcoin/rust-miniscript/pull/109#issuecomment-702375954
13:23 < andytoshi> ok i'll merge it then
13:24 < sanket1729> I think shesek was interested in a simple deterministic compiler rather a optimal one. 
13:26 < sanket1729> There is value some value in having  `compile(compose(p1,p2)) = compose(compile(p1),compile(p2))` that we do not currently have
13:26 < andytoshi> lol now travis is taking forewer on 109
13:27 < sipa> sanket1729: i don't think that can work in general
13:27 < andytoshi> yeah i agree if the compiler were functorial in that sense it'd be great
13:27 < andytoshi> heh it may be impossible
13:27 < sanket1729> we don't have travis?
13:27 < andytoshi> oh lol
13:27 < andytoshi> github ci then
13:28 < sanket1729> I think it's probably because it was opened before we swtich CI
13:29 < andytoshi> no, it was rebased earlier today
13:29 < sipa> or at least not if you want an optimal solution, for any reasonable definition of optimal
13:29 < sipa> if you want composability, use miniscript, not compilation
13:30 < sanket1729> andytoshi: It was rebased today
13:30 < andytoshi> yes it will certianly be impossible if you want to optimize
13:30 < sanket1729> that was another PR by jeremyrubin about True/False policy
13:30 < andytoshi> sanket1729: oh lol
13:30 < andytoshi> jeremyrubin: can you rebase #109
13:34 < sanket1729> I can cherry-pick that if we want to work on release commits now
13:34 < andytoshi> yeah actually, that'd be great
13:34 < andytoshi> can you do a release PR? i am on a call right now
13:34 < andytoshi> i guess we just need a changelog and to cherry-pick 109
13:34 < sanket1729> sure
13:34 < andytoshi> awesome thanks :)
13:35 < sanket1729> andytoshi: Should we also update to 0.25?
13:35 < sanket1729> rust-bitcoin
13:35 < andytoshi> sanket1729: not yet
13:35 < andytoshi> we need to do a release with MSRV at 1.22
13:35 < andytoshi> then a followup one that bumps MSRV to 1.29 and bitcoin to 0.25
13:35 < sanket1729> cool
13:36 < sanket1729> andytoshi: What was that graph git log command that you shared last time :P ?
13:37 < andytoshi> git log HEAD --not <last tag> --oneline --graph
13:42 < andytoshi> also if you do --min-parents=2
13:42 < andytoshi> you'll get only merge commits so you can see the PRs
13:42 < sipa> --merges is the same as --min-parents=2
13:42 < andytoshi> oh nice
13:43 < sipa> confusingly --no-merges is the same as --max-parents=1
13:43 < sanket1729> andytoshi: I think we should also update the psbt interface for satisfier
13:44 < sanket1729> Change it from psbt::Input to new struct (psbt, index)
13:44 < andytoshi> moreso than we did in the PRs we merged?
13:44 < andytoshi> oh right .. i forgot about that
13:44 < sanket1729> It is indeed as of now
13:44 < andytoshi> we currently just don't support sequence numbers?
13:44 < sanket1729> I thought I would do that along with hash preimages
13:44 < andytoshi> ah i see
13:45 < sanket1729> I will add that too.
13:45 < andytoshi> i guess we can't since we won't have hash preimagise til bitcoin 0.25?
13:45 < sanket1729> It's not merged yet :) 
13:45 < sanket1729> after 0.25.1?
13:45 < andytoshi> heh yeah
14:02 < sanket1729> andytoshi: #141 and #142 are up
14:02 < sanket1729> working on release PR now.
14:06 < andytoshi> dope thanks
14:26 < sanket1729> up. 
15:03 < andytoshi> sanket1729: need to update the psbt pr
15:03 < andytoshi> with height/time checks
15:03 < sanket1729> ah     
15:03 < sanket1729> oops
15:05 < andytoshi> also lol when did we do 1.0?
15:05 < andytoshi> seems premature given that we know we're gonna break shit when taproot comes out
15:12 < sanket1729> We did it around July. I guess we were feeling confident about the stability. 
15:13 < sanket1729> pushed a new fix. Using Older/After structs from satisfy.rs
15:14 < andytoshi> nice
15:17 < andytoshi> ok act 142 .. waiting on ci
15:26 < andytoshi> published 2.0 :)
15:26 < sanket1729> awesome
15:26 < sanket1729> I just found a typo in changelog :|
15:26 < andytoshi> i'm tempted to pull a rust-ism and bump the MSRV in 2.1..
15:26 < andytoshi> oh lol "miscellaneous"
15:27 < sanket1729> you found another one :P
15:27 < andytoshi> hah what was the first one?
15:31 < andytoshi> lol i was all excited to use https://github.com/rust-bitcoin/rust-bitcoin/pull/387 but actually we haven't released it yet..
15:31 < andytoshi> i need to do a new rust-bitcoin (minor) release befoer we can use those functions
15:33 < sanket1729> pushed a pr
15:37 < sanket1729> I don't see why not do 3.0.0?
15:38 < sanket1729> Let's change the rust-ism
16:09 < jeremyrubin> hey sorry was AFK
16:11 < jeremyrubin> thanks for the rebase sanket1729!
16:50 < andytoshi> sanket1729: because it looks bad to have enormous version numbers and to be doing several releases in a year
16:51 < andytoshi> for supposedly 1.0 software
16:55 -!- Netsplit *.net <-> *.split quits: Aleru, felixweis
17:05 -!- Netsplit over, joins: Aleru, felixweis
19:23 -!- jb55 [~jb55@gateway/tor-sasl/jb55] has quit [Remote host closed the connection]     
19:24 -!- jb55 [~jb55@gateway/tor-sasl/jb55] has joined ##miniscript     
19:33 -!- shesek [~shesek@164.90.217.137] has joined ##miniscript     
19:33 -!- shesek [~shesek@164.90.217.137] has quit [Changing host]     
19:33 -!- shesek [~shesek@unaffiliated/shesek] has joined ##miniscript     
20:21 -!- Netsplit *.net <-> *.split quits: Aleru, felixweis
20:32 < sanket1729> andytoshi: heh, sure we can keep it at 2.1 . 
20:49 -!- Aleru [sid403553@gateway/web/irccloud.com/x-yujxdzfveoxjwazt] has joined ##miniscript     
20:49 -!- felixweis [sid154231@gateway/web/irccloud.com/x-krefugjevbfmhthz] has joined ##miniscript     
21:23 -!- hugohn [sid304114@gateway/web/irccloud.com/x-wnvicsfksfayqswb] has quit [Ping timeout: 256 seconds]     
21:24 -!- hugohn [sid304114@gateway/web/irccloud.com/x-wigsyrnczlzuvegj] has joined ##miniscript     
21:29 -!- ghost43 [~daer@gateway/tor-sasl/daer] has quit [Remote host closed the connection]     
21:29 -!- ghost43 [~daer@gateway/tor-sasl/daer] has joined ##miniscript     
22:00 -!- ghost43 [~daer@gateway/tor-sasl/daer] has quit [Ping timeout: 240 seconds]     
22:01 -!- ghost43 [~daer@gateway/tor-sasl/daer] has joined ##miniscript