< ossifrage>
I think I just managed to get the client to take the fee from the wrong side of the transaction
< achow101>
I want to change addmultisigaddress's output from an address to a json object with an address and redeemscript field, so what's the best way to do that?
< jonasschnelli>
achow101: just do it and make sure it's mentioned in the release notes?
< jonasschnelli>
API changes should be avoided, yes. But also we should not have every new RPC call static made in stone
< luke-jr>
IMO just do it for that specific case, mainly because it's not useful for non-testing
< achow101>
ok
< bitcoin-git>
[bitcoin] achow101 opened pull request #11415: [RPC] Disallow using addresses in createmultisig (master...createmultisig-no-addr) https://github.com/bitcoin/bitcoin/pull/11415
< esotericnonsense>
looking back on the meeting; I've done some work previously on serverside ZMQ tracking of topics, it's kind of a hack on the protocol but you can detect when a client disconnects and queue up messages for them for some period, if you combine that with sequence numbers and overlap them you can mitigate missing messages to some extent, but it's still a hack on the ZMQ semantics for pubsub
< esotericnonsense>
(and makes the notifications stateful)
< bitcoin-git>
bitcoin/master bb174e1 fanquake: [docs] Remove partial gitian instructions from descriptors dir
< bitcoin-git>
bitcoin/master bbc901d Wladimir J. van der Laan: Merge #11414: [docs] Remove partial gitian build instructions from descriptors dir....
< bitcoin-git>
[bitcoin] laanwj closed pull request #11414: [docs] Remove partial gitian build instructions from descriptors dir. (master...gitian-cleanups) https://github.com/bitcoin/bitcoin/pull/11414
< esotericnonsense>
wumpus: on the Termux thing; I did get it to work, I had to hack a few of the #defines in leveldb, compile libevent, and make that #include change from the issue.\
< esotericnonsense>
I'm heading off to work now but I can write it up later if you're interested, it's running on a few different phones I have here.
< bitcoin-git>
[bitcoin] maaku opened pull request #11418: Add error string for CLEANSTACK script violation (master...cleanstack-error-code) https://github.com/bitcoin/bitcoin/pull/11418
< bitcoin-git>
[bitcoin] OmeGak opened pull request #11419: Utils: Fix launchctl not being able to stop bitcoind (master...fix-macos-init) https://github.com/bitcoin/bitcoin/pull/11419
< ThomasV>
bip173 merged :)
< mryandao>
hi, i just ran `ldd` on bitcoin-cli and noticed that it requires libboost, is it really necessary?
< wumpus>
esotericnonsense: cool! can you post the patch somewhere please? we should probably handle the leveldb changes through upstream but it'd be nice to try
< bitcoin-git>
bitcoin/master ff4cd60 MarcoFalke: Merge #11319: [qa] Fix error introduced into p2p-segwit.py, and prevent future similar errors...
< bitcoin-git>
[bitcoin] MarcoFalke closed pull request #11319: [qa] Fix error introduced into p2p-segwit.py, and prevent future similar errors (master...2017-09-fix-p2p-segwit) https://github.com/bitcoin/bitcoin/pull/11319
< esotericnonsense>
wumpus: sure. i'll recompile on master and post a diff (I just downloaded the tarball and hacked about so it's annoying, also i've make cleaned so might as well wait it out anyway).
< esotericnonsense>
i'll try test_bitcoin as well. pretty sure the issue is just fs::temp_directory_path() giving an inaccessible directory.
< esotericnonsense>
unrelated, i would appreciate input on https://github.com/bitcoin/bitcoin/pull/11359 if someone could be so kind. it's a tiny diff, just curious on whether anyone has a better approach.
< BlueMatt>
jonasschnelli: care to close 8757 (superceded)
< BlueMatt>
cfields: wanna just close 9000? It needed rebase a long time ago and you even said you dont particularly care if it goes upstream
< esotericnonsense>
yeah, that's it. boost::filesystem::temp_directory_path() gives /data/local/tmp which is denied access (if it even exists, i don't have root on this device).
< esotericnonsense>
it's pretty much the bare minimum hack to get it to compile and for test_bitcoin / bitcoind to run, hardcoding #define OS_ANDROID is my lazy way of not wanting to face autotools stuff :P
< esotericnonsense>
i'm guessing it just sees Linux in uname -a or whatever and that's that
< esotericnonsense>
i'm using ./configure --disable-wallet, and also termux-chroot, i didn't note the dependencies as I went along, unfortunately, but everything else should be in termux repos
< bitcoin-git>
[bitcoin] TheBlueMatt opened pull request #11422: Verify DBWrapper iterators are taking snapshots (master...2017-09-leveldb-check-snapshots) https://github.com/bitcoin/bitcoin/pull/11422
< bitcoin-git>
[bitcoin] jl2012 opened pull request #11423: [Policy] Make OP_CODESEPARATOR and FindAndDelete in non-segwit scripts non-std (master...const_scriptcode) https://github.com/bitcoin/bitcoin/pull/11423
< bitcoin-git>
bitcoin/master cee28fb Mark Friedenbach: Add error string for CLEANSTACK script violation, preventing an "unknown error" if the CLEANSTACK error condition is set.