< bitcoin-git>
[bitcoin] kallewoof opened pull request #9387: RAII of libevent stuff using set of wrappers (master...raii-libevents) https://github.com/bitcoin/bitcoin/pull/9387
< jtimon>
I would really welcome help from experts on ./rpc-tests/segwit.py or ./rpc-tests/p2p-compactblocks.py I really can't understand why changing the genesis block of regtests should be an issue or how can I correct it or find my error
< jtimon>
sorry, I'm talking about #8994 again, which changes all rpc/py tests from "regtest" to "custom" (the tests should be identical except for the genesis block, those 2 tests resist)
< jtimon>
in each of them, one test is failing, let me go back to what line was failing on each
< gmaxwell>
\O/ hurray about the RC.
< instagibbs>
jtimon, please give more info thanks
< jtimon>
oh, sorry, I forgot
< jtimon>
in segwit.py the first error is in L428, in L844
< jtimon>
in segwit.py the first error is in L428, in qa/rpc-tests/p2p-compactblocks.py in L844
< jtimon>
in all the rest everything seems to be working fine, my fear is omse of them aren't calling setup_chain()
< jtimon>
in segwit.py the error is "missing inputs", my guess is that the tests are somehow using regtest genesis block instead of the custom chain one
< MarcoFalke_web>
Does the new chain work with the caching in the python tests?
< MarcoFalke_web>
looking at your pull, it should
< jtimon>
the custom chain should be completely identical for regtest and custom by default expect for the genesis block
< jtimon>
ideally, that commit would only touch test_framework.py and be squashed somewhere
< jtimon>
I'm pretty convinced it has something to do with the mininodes, but there still some chain-dependent code I'm looking fore
< cfields>
gitian builders: detached sigs for 0.13.2rc1 pushed.
< timothy>
hi, I have a slightly question. I'm trying to integrate bitcoin with tor under docker
< timothy>
it works, but unlucky -proxy and -torcontrol doesn't support hostnames, but only IPs and this doesn't work under docker-compose
< gmaxwell>
the tor support in Bitcoin Core really intends the tor daemon to be on localhost.
< gmaxwell>
if it's on another host right now it will end up banning the tor host.
< timothy>
well, it works also if you have the tor daemon on another host, but I need to specify the ip :P
< gmaxwell>
It doesn't work for long.
< timothy>
ok, wrong idea. thank you :)
< sipa>
-proxy certainly accepts IP addresses
< gmaxwell>
hopefully we'll fix that in an upcoming version.
< gmaxwell>
sipa: yes, outbound would be fine.
< jtimon>
right, for simple proxies an env variable for docker may be enough
< gmaxwell>
Proxy won't take a hostname?
< btcdrak>
I dont mean to be a stick in the mud, but why are merge commits not signed in the gitian.sigs repository?
< MarcoFalke_web>
What would it help?
< sipa>
gmaxwell: seems it doesn't
< jtimon>
no, but environment variable would give the ip, no? I'm not really sure what he is trying to do so probably I shouldn't talk
< gmaxwell>
lol, gitian sigs are already sigs.
< MarcoFalke_web>
btcdrak: Your linux has a mismatch
< btcdrak>
MarcoFalke_web: just the linux one?
< MarcoFalke_web>
jup
< btcdrak>
weird.
< gmaxwell>
A request for 0.13rc2 testers: Please run ./qa/pull-tester/rpc-tests.py ... we have this fantastic test suite and I get the impression that a lot of people only use it via travis. If there are bad interactions with your system and the software these tests are probably more likely to reveal them then just your local testing alone.
< achow101>
I'm gitian building at the same time, so that might be interfering with something too
< MarcoFalke>
Could be also OOM
< achow101>
could be
< morcos>
sipa: i've spent too long confused by move semantics already today, but i think there is an easily fixable potential bug... can you take a quick look
< morcos>
in wallet.cpp under the logic for -walletrejectlongchains
< morcos>
hasn't txNew been moved from?
< sipa>
what line?
< morcos>
search for walletrejectlongchains, i'm not on master
< morcos>
right under that
< morcos>
sorry second occurence, line 2566
< sipa>
yup
< sipa>
i wonder how that code works at all
< morcos>
thats what i mean by confused
< morcos>
i made the same error myself on some other wallet code i was working on earlier
< sipa>
txNew in practice will just have empty txin and txout
< morcos>
same bug in 0.13.2
< morcos>
if you don't mind taking it from here.. i have a couple things i'm rushing to finish before dinner
< morcos>
but out of curiousity, why is that such an easy mistake to make? shouldn't there be a compiler warning or something
< sipa>
there can't be
< sipa>
well, it could be a suggestion
< sipa>
but in general, use after move is well defined