< fanquake> Need someone to tiebreak the non-determinism in the aarch64 and riscv64 rc1 gitian builds
< achow101> oh noes non-determinism :(
< gmaxwell> fanquake: have you looked to see what is different in the build products?
< * sipa> feels like tiebreaking will not help nondeterminism
< fanquake> gmaxwell not yet, just going to wait for my own builds to finish
< achow101> my gitian build results are available at https://github.com/achow101/bitcoin/releases/tag/v0.18.0rc1 if you need them
< fanquake> achow101 Looks like my builds match with yours
< fanquake> Maybe it's a reoccurrence of something like https://github.com/fanquake/core-review/blob/master/determinism.md#ld64-threading heh
< bitcoin-git> [bitcoin] gwillen opened pull request #15534: In lint-format-strings, open files sequentially (fix for OS X failure) (master...feature-fix-lint-osx) https://github.com/bitcoin/bitcoin/pull/15534
< zhangzf> hello, Who have testnet bitcoin? Can you give me some, I want to test code. My address is mgSRBDvv9h2bRgZHGWLo4xQU9PLvssK6Yi
< aj> zhangzf: google "testnet faucet"
< zhangzf> aj: Thank you, I will try it.
< zhangzf> aj: "bitcoinfaucet.uo1.net is currently unable to handle this request"
< zhangzf> aj: The web page is right?
< achow101> zhangzf: there's lots of different testnet faucets. if one is down, try another
< zhangzf> achow101: Ok. I see I have recevied 3 BTC. I guess someone sended to me.
<@gwillen> sipa: do you have thoughts on how to handle the dependency issue I have run into
<@gwillen> I am trying moving tx_verify into common
<@gwillen> but it's going to be awhile before I find out whether it worked, because building in a linux VM is very slow
<@gwillen> I don't really understand why it doesn't break under clang/OS X
< sipa> gwillen: i think osx linker probably discards unused function symbols
< sipa> so if you're not using any of the functions that have unresolved dependencies, it doesn't error
< sipa> which is pretty fragile, but shows that there must exist a solution :)
<@gwillen> ohhhh, heh, that makes sense
< bitcoin-git> [bitcoin] Deano1167 opened pull request #15535: Dander (0.14...master) https://github.com/bitcoin/bitcoin/pull/15535
< bitcoin-git> [bitcoin] fanquake closed pull request #15535: Dander (0.14...master) https://github.com/bitcoin/bitcoin/pull/15535
< rafalcpp> regarding the attacks on RPC (replay message, or replace body of message) is there a plan how to solve it?
< luke-jr> rafalcpp: just don't let anyone get your password..
< luke-jr> or access your RPC
< bitcoin-git> [bitcoin] laanwj pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/14023c966c51...4b57b7f0f0c3
< bitcoin-git> bitcoin/master fad76e7 MarcoFalke: doc: Remove pr release notes file in wrong dir
< bitcoin-git> bitcoin/master 4b57b7f Wladimir J. van der Laan: Merge #15527: doc: Remove pr release notes file in wrong dir
< bitcoin-git> [bitcoin] laanwj pushed 1 commit to 0.18: https://github.com/bitcoin/bitcoin/compare/6a178e52618d...71ac4ebe4893
< bitcoin-git> bitcoin/0.18 71ac4eb MarcoFalke: doc: Remove pr release notes file in wrong dir
< bitcoin-git> [bitcoin] laanwj merged pull request #15527: doc: Remove pr release notes file in wrong dir (master...1903-docRelPr) https://github.com/bitcoin/bitcoin/pull/15527
< jonasschnelli> hmmm.... looks like the linux build is not deterministic: https://bitcointools.jonasschnelli.ch/data/builds/1007/bitcoin-linux-0.18-build.assert
< jonasschnelli> diverges from wumpus
< jonasschnelli> +signatures
< jonasschnelli> Only arm64 and riscv though
< wumpus> achow101's was divergent too, maybe it matches his ?
< jonasschnelli> Oh. We merged divergent signaturs... :/
< jonasschnelli> fanquakes is not identical to yours wumpus but to mine
< fanquake> Yea, mine match achows
< wumpus> gitian sigs should always be merged, *especially* if they don't match, it's evidence
< jonasschnelli> fair point
< wumpus> okay so only mine are different? this is interesting
< wumpus> if it's that then if I build a few more times I should get different outputs?
< wumpus> trying
< fanquake> it'd be interesting if we'd somehow reintroduced that. More interesting given it's constrained to riscv and aarch64
< wumpus> I'm going to try if it's a race condition by trying to build a few times, if not I'd like to compare executables
< fanquake> wumpus achow's build results are up here: https://github.com/achow101/bitcoin/releases/tag/v0.18.0rc1. I can upload mine as well if needed
< wumpus> thanks! if they're the same it's not necessary, if you get *yet another* output, well :)
< bitcoin-git> [bitcoin] laanwj pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/4b57b7f0f0c3...3800ca606896
< bitcoin-git> bitcoin/master 3eac2d5 Alistair Mann: docs: add "sections" info to example bitcoin.conf
< bitcoin-git> bitcoin/master 3800ca6 Wladimir J. van der Laan: Merge #15513: docs: add "sections" info to example bitcoin.conf
< bitcoin-git> [bitcoin] laanwj merged pull request #15513: docs: add "sections" info to example bitcoin.conf (master...rebase-fixup-15387) https://github.com/bitcoin/bitcoin/pull/15513
< * wumpus> confused by #15521, I think what ever shouldbe the solution to this, it's not doing a re-checkout of .
< gribble> An error has occurred and has been logged. Please contact this bot's administrator for more information.
< rafalcpp> is there any reason for RPC to use http Basic Auth header, instead of adding the password as request parameter (so, in body)?
< luke-jr> rafalcpp: the latter would be ugly
< jonasschnelli> rafalcpp: both are kinda ugly...
< jonasschnelli> But the authentication does not protect from attackers on system level and exposing RPC to the internet is not recommended
< rafalcpp> the idea would be {"method":"getblockchaininfo","params":[],"id":1,"hmac-tag":"34d3f35460166cc84cc1b51a61abe9c5a9994558b777125969aed8d9d0c29588","hmac-nonce":"eaaa1578252fc8de8fe115e5802e999a67353bce", "hmac-clock":"1551790542"}
< luke-jr> rafalcpp: that violates the JSON-RPC protocol I'm pretty sure..
< jonasschnelli> I once proposed an ecc based authentication with no replay protection: https://github.com/bitcoin/bitcoin/pull/6604
< jonasschnelli> rafalcpp: if you need better authentication, consider adding a reverse proxy (Apache / Nginx), add TLS and cert based user authentication (on top of Cores HTTP BASIC AUTH)
< rafalcpp> jonasschnelli: yeah Im proposing the reverse, no DH/pubkey but full replay protection. of course, we could combine both methods and then pin to pubkey in server and in client
< jonasschnelli> rafalcpp: I think there is no reason to implement that in core...
< jonasschnelli> we don't want that people expose it to the internet,...
< rafalcpp> I will include motivation in BIP
< jonasschnelli> we would suddenly become maintainer of a http server and it's pitfalls
< rafalcpp> (it is about localhost, not about internet)
< jonasschnelli> Just use existing authentication layer and reverse proxy to it
< luke-jr> rafalcpp: this is not a BIPable topic
< jonasschnelli> weak agree with luke-jr. The RPC layer is very core specific
< rafalcpp> * in the RFC :)
< wumpus> tried rebuilding; got the same hash for risc-v and aarch64, now trying with wiped cache
< bitcoin-git> [bitcoin] MarcoFalke pushed 7 commits to master: https://github.com/bitcoin/bitcoin/compare/3800ca606896...a74d588f2154
< bitcoin-git> bitcoin/master fa6bf21 MarcoFalke: scripted-diff: test: Use py3.5 bytes::hex() method
< bitcoin-git> bitcoin/master fab5a1e MarcoFalke: build: Require python 3.5
< bitcoin-git> bitcoin/master fa0e65b MarcoFalke: scripted-diff: test: Remove brackets after assert
< bitcoin-git> [bitcoin] MarcoFalke merged pull request #14954: build: Require python 3.5 (master...Mf1811-buildPython3) https://github.com/bitcoin/bitcoin/pull/14954
< bitcoin-git> [bitcoin] MarcoFalke pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/a74d588f2154...d8a62db8bf68
< bitcoin-git> bitcoin/master 4d4e4c6 Russell Yanofsky: Suggested interfaces::Chain cleanups from #15288
< bitcoin-git> bitcoin/master d8a62db MarcoFalke: Merge #15531: Suggested interfaces::Chain cleanups from #15288
< bitcoin-git> [bitcoin] MarcoFalke merged pull request #15531: Suggested interfaces::Chain cleanups from #15288 (master...pr/wclean2) https://github.com/bitcoin/bitcoin/pull/15531
< jonasschnelli> Looks like ChaCha20Poly1305 is pretty fast on arm64(aarch64)
< jonasschnelli> CHACHA20_1MB, 5, 340, 13.3704, 0.00772624, 0.00806671, 0.00775259
< jonasschnelli> CHACHA20_256BYTES, 5, 250000, 2.41802, 1.92206e-06, 1.95887e-06, 1.92964e-06
< jonasschnelli> CHACHA20_64BYTES, 5, 500000, 1.28993, 5.07247e-07, 5.26875e-07, 5.14802e-07
< jonasschnelli> HASH_1MB, 5, 340, 22.5945, 0.0128931, 0.0135617, 0.0132931
< jonasschnelli> HASH_256BYTES, 5, 250000, 6.84513, 5.43544e-06, 5.54961e-06, 5.45136e-06
< jonasschnelli> HASH_64BYTES, 5, 500000, 7.47419, 2.94502e-06, 3.04259e-06, 2.99531e-06
< jonasschnelli> POLY1305_1MB, 5, 340, 10.2145, 0.00591661, 0.00609435, 0.00601109
< jonasschnelli> POLY1305_256BYTES, 5, 250000, 1.88446, 1.48188e-06, 1.52889e-06, 1.52119e-06
< jonasschnelli> POLY1305_64BYTES, 5, 500000, 1.08101, 4.25452e-07, 4.43012e-07, 4.29685e-07
< jonasschnelli> ^ sipa gmaxwell
< jonasschnelli> results are from a RK3328 Quad-Core ARM Cortex A53 64-Bit
< jonasschnelli> (no SHA2 hardware acceleration)
< jonasschnelli> Which shows a tendency that something like BIP151 may speed up processing performance on ARM... especially small packets
< jonasschnelli> ~72MB/s
< bitcoin-git> [bitcoin] MarcoFalke pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/d8a62db8bf68...4952a953585e
< bitcoin-git> bitcoin/master 21be609 Glenn Willen: In lint-format-strings, open files sequentially
< bitcoin-git> bitcoin/master 4952a95 MarcoFalke: Merge #15534: [test] lint-format-strings: open files sequentially (fix for...
< bitcoin-git> [bitcoin] MarcoFalke merged pull request #15534: [test] lint-format-strings: open files sequentially (fix for OS X) (master...feature-fix-lint-osx) https://github.com/bitcoin/bitcoin/pull/15534
< didoerpl> hey
< didoerpl> when will zksnarks be implemented for privacy
< didoerpl> why still no privacy
< bitcoin-git> [bitcoin] instagibbs opened pull request #15538: wallet_bumpfee.py: Make sure coin selection produces change (master...bumpfee_change_test) https://github.com/bitcoin/bitcoin/pull/15538
< provoostenator> Gitian woes: bin/gbuild:21:in `system!': failed to run on-target setarch x86_64 bash -x < var/build-script > var/build.log 2>&1 (RuntimeError)
< provoostenator> Last line in build.log: checking for BSD- or MS-compatible name lister (nm)... nm
< provoostenator> Ah, out of disk space :-)
< wumpus> okay, after cleaning the cache I still get the same riscv/aarch64 output
< MarcoFalke> wumpus: Mind to upload the binaries?
< MarcoFalke> or anyone else who has matching hashes
< jonasschnelli> gmaxwell: about the message histogram: I logged about 5h, ~50% of the messages are invs while ~90% of the traffic is due to the 5% block messages
< jonasschnelli> The median message size is 103 bytes (only the payload)
< jonasschnelli> (and thats sending)
< jonasschnelli> receiving: 34% invs, 39% tx, median message size: 31 bytes
< jonasschnelli> 88% of inbound traffic is inv/tx (73% of all messages), median inv size: 109, median tx size: 325
< gmaxwell> the 5% block messages, also then means these numbers aren't representative for all peers (they're from your node helping IBDing peers, and most nodes on the network don't participate in that).
< bitcoin-git> [bitcoin] MarcoFalke closed pull request #8471: Key origin metadata, with HD wallet support (master...keyorigin_hd) https://github.com/bitcoin/bitcoin/pull/8471
< mmgen> gmaxwell: is this a permanent ban from #bitcoin, greg? Just asking. To be honest, I'm astounded. As a long-time participant here, I wasn't expecting this
< mmgen> gmaxwell: trying to fathom the motive
< gmaxwell> You are only in there continually promoting your "key generator tool" that sets off a multiple red flags, it would be difficult for it to look more like malware if you were trying.
< gmaxwell> It's not welcome in there.
< mmgen> gmaxwell: it's a complete wallet and cold storage system, not a key-generating tool
< mmgen> a 5-year old project
< mmgen> and I don't see the "continual promotion". I mention it in response to peoples' queries, that's all
<@gwillen> mmgen: just as a devil's advocate here, a popular thing for people to do is take over projects with a lot of history and inject malware
<@gwillen> so that's not totally dispositive as to intent
<@gwillen> (note that I only have an @ here for spam-fighting reasons, I am not in charge here and defer to gmaxwell on all matters of relevance)
<@gwillen> (or to Appropriate Authorities)
<@gwillen> also, in general #bitcoin has a lot of people who are a danger to themselves in terms of running whatever software is in front of them
<@gwillen> and risking their coins in the process
<@gwillen> so there has to be a pretty high bar in there in general
<@gwillen> (this is why e.g. most links are looked on with extreme suspicion)
< mmgen> gwillen: ok
<@gwillen> FWIW I think your tool looks cool, although I am skeptical that your alternative to BIP32 is an improvement but I'd be interested to hear about the motivation behind it (but not in this channel, perhaps #bitcoin-dev would accept such a conversation)
<@gwillen> (I'm not sure where the best place is)
< mmgen> gwillen: thanks for the interest
< gmaxwell> My logs show mmgen first joined #bitcoin on feb 23rd and in that time he his linked his tool 35 times.
< mmgen> s/tool/wallet/
< mmgen> and I've been here much longer than that
< mmgen> for well over a year
< gmaxwell> In any case, this discussion is offtopic here.
< mmgen> well if i'm banned on #bitcoin, how can I get unbanned?
< mmgen> that's the only reason I'm discussing it here
< mmgen> my nick was registered around July 2017, by the way
< mmgen> gwillen: my concern with bip32 is that is uses ecc, which could be a problem after the advent of quantum computing
< mmgen> gwillen: that's why my key-derivation scheme is hash-based
< gmaxwell> mmgen: hardend bip32 is a hash derrived private key, you are spreading disinformation claiming that it has any different security properties.
< mmgen> gmaxwell: but it uses ec as well
< sipa> mmgen: please take this elsewhere
< sipa> (and others)
< booyah> mmgen: #bitcoin-bans is about discussing bans
< mmgen> sipa: well, I'd be happy to take it to #bitcoin, but gmaxwell has banned me there, for promoting "snake oil"
< sipa> mmgen: sorry to hear that, but not my problem. it's completely off topic here
< mmgen> booyah: thanks for the pointer