< esotericnonsense>
tell me to STFU and go elsewhere as necessary
< esotericnonsense>
has anyone (other than the ABCore project) made a mobile wallet based on bitcoin core, e.g. remote?
< esotericnonsense>
it's kind of obvious, but I'm realising that if I expose getnewaddress, sendtoaddress, and maybe a QR code in my interface, you have an android wallet that's backed by bitcoin core
< esotericnonsense>
it seems trivial enough that someone must have already done it and realised that no-one (other than themselves personally) finds this useful?
< luke-jr>
esotericnonsense: you shouldn't expose RPC over the internet
< esotericnonsense>
luke-jr: right, don't expose RPC over the internet, expose the subset that you need through a proxy that has sanity checking
< esotericnonsense>
with proper authentication (e.g. session token over TLS)
< luke-jr>
esotericnonsense: it's something I've thought about for a long time
< esotericnonsense>
i don't think it's ever going to be viable for like, thousands or tens of thousands of dollars, because the attack surface is too high
< esotericnonsense>
or rather, I think it'd be possible to make secure, but I'm just too paranoid about every layer
< esotericnonsense>
i'm kind of at the 90/10 stage where I have all the infrastructure in place, but packaging it sanely to make it something that anyone else can use is quite difficult
< gwillen>
esotericnonsense: as a practical matter I think very few people would ever use an android app that requires setting up one's own backend
< esotericnonsense>
luke-jr: thanks for that link, i may have to edit it as I go if I decide to do this
< gwillen>
very few people can be bothered, even technical people, and I say this as someone who runs my own mailserver
< esotericnonsense>
gwillen: it'd be a webapp not android specifically, but yes, I see that being very unlikely too
< esotericnonsense>
the frustrating thing is that actually if namecoin or similar took off, I think it could be packaged into a simple 'run this file and leave your computer on, then go to www.whatever.com on your phone'
< gwillen>
you don't really need namecoin for service discovery, e.g. you could use IPFS
< esotericnonsense>
so I'm thinking it can be made one-click on the desktop side, but end users don't have their own domain
< gwillen>
"scan this QR code using your phone"
< gwillen>
sorry, this is getting off-topic for this channel I fear
< esotericnonsense>
yes, indeed, heh, sorry :)
< sipa>
esotericnonsense: if namecoin took off, it'd be useless as a name registration mechanism :)
< esotericnonsense>
anyway, i'll tinker with it myself and see where I get
< luke-jr>
esotericnonsense: note that link is from 2011; it probably needs serious rethinking
< esotericnonsense>
luke-jr: I think that ultimately someone would have to make a demo of it in order to actually understand what the issues are; some of the stuff in there on a quick skim seems like premature optimization
< esotericnonsense>
> Cooperative mining (see below) :D
< * esotericnonsense>
feels young again
< meshcollider>
wumpus: I think #14689 is RTM already
< gwillen>
meshcollider: I'm not sure what it's currently supposed to do, but my understanding was that it turned out that updating a comment still bumps the date on the issue
< gwillen>
so the change didn't really help in the way it was hoped it would
< meshcollider>
but still, it makes it difficult to read a PR thread if there are multiple Drahtbot posts with conflicts instead of just one up-to-date one
< meshcollider>
i think its a bug
< kabaum>
I noticed that a bitcoind logs to stdout and debug.log by default instead of just debug.log as was the case at least in 0.16.0. What was the rationale behind this change?
< sipa>
kabaum: it does so when running in the foreground
< sipa>
if you run with -daemon it doesn't
< sipa>
the rationale is that if you're running things on the foreground in a terminal, you probably want to see something
< kabaum>
Ok, I was just curious on why the default behavior changed from 0.16.0.
< meshcollider>
it was mentioned in the 0.17 release notes btw
< kabaum>
Oh, I glanced over it, but somehow it escaped me. Thanks.
< meshcollider>
kabaum: and the original PR is 12689 if you're interested
< luke-jr>
any ideas why I can't reply to MeshCollider? :/
< meshcollider>
wow that's very weird
< meshcollider>
luke-jr: looks like you can reply on the files tab
< meshcollider>
might be because I was already replying to conversation threads further up
< kallewoof>
sipa: I constantly start bitcoind with '&' to do something real quick, then ^D^C to stop it. The new change made this a bit less convenient, unfortunately, but I'll live. (regarding default printtoconsole)
< kallewoof>
sipa: i want to try some command on a new instance real quick so i do e.g. mkdir x; ./bitcoind -datadir=x -regtest &; [do stuff]; ^D^C and the instance is shutdown. I could just do './bitcoin-cli -datadir=x stop' but ^D^C is .. shorter.
< kallewoof>
wait, not ^D^C, fg;^C
< kallewoof>
still shorter.
< sipa>
killall bitcoind
< sipa>
which also works with -daemon
< kallewoof>
sipa: that works most of the time but i do have several servers with an actual bitcoind mainnet full node running.
< provoostenator>
sipa: I'm tempted to take your "killall bitcoind" line out of context :-P
< luke-jr>
meshcollider: yeah, replying from the files tab worked
< esotericnonsense>
kallewoof: do it the other way around?
< esotericnonsense>
e.g. make bitcoind the foreground instance
< bitcoin-git>
[bitcoin] ken2812221 opened pull request #14698: build: Add bitcoin-tx.exe into Windows installer (master...win-bitcoin-tx) https://github.com/bitcoin/bitcoin/pull/14698
< bitcoin-git>
[bitcoin] MarcoFalke opened pull request #14700: qa: Avoid race in p2p_invalid_block by waiting for the block request (master...Mf1811-qaPassOnCentOs) https://github.com/bitcoin/bitcoin/pull/14700