00:18:14  * warbrettquit (Quit: Connection closed for inactivity)
00:21:23  * pfrazejoined
00:27:20  * pfrazequit (Remote host closed the connection)
00:43:15  * groundwaterjoined
00:48:25  * robertkowalskiquit (Remote host closed the connection)
00:48:25  * pfrazejoined
00:48:35  * robertkowalskijoined
00:50:24  * contrahaxjoined
01:05:30  <thealphanerd>ey substack I'm trying to get mkdirp into the node smoketesting suite
01:05:50  <thealphanerd>if you have time to checkout https://github.com/substack/node-mkdirp/pull/100
01:05:55  <thealphanerd>that fix should have it ready to go
01:06:18  <thealphanerd>old version of mock-fs doesn't work on v5 >
01:07:40  <feross>substack: the tls cert on htmlb.in is expired
01:08:29  <substack>feross: yes I know
01:08:44  <substack>feross: I'm going to have a new version that uses service workers instead of hash domains
01:08:44  <feross>i don't mind clicking through the warning, but i can't get curl to post to it
01:08:50  <substack>curl -k
01:09:05  <substack>I'm going to use service workers so I can get a free ssl cert from letsencrypt
01:09:30  <substack>and! I'm going to distribute hashes to webtorrent with https://github.com/substack/swarmlog
01:09:33  <feross>substack: hm, why does using a service worker affect whether you can get a free cert?
01:09:58  <substack>feross: wildcard domains cost money
01:10:33  <substack>but a service worker can give the same benefits
01:10:44  <substack>for hashed URLs that can't be updated
01:10:45  <feross>wow, even with lets encrypt? i thought the whole point was to make it free?
01:10:55  * groundwaterquit (Quit: My MacBook has gone to sleep. ZZZzzz…)
01:10:56  <feross>swarmlog looks excellent!
01:11:08  <substack>https://github.com/letsencrypt/letsencrypt/issues/66
01:12:01  <substack>for the next hyperboot, I want to publish webapp metadata to a swarmlog
01:12:18  <substack>clients will replicate the swarmlog with each other and the docs in the warmlog will have magnet links for the html payloads
01:12:55  <substack>for a related feature, is there any way in webtorrent to re-use an existing swarm?
01:13:46  <substack>because I'm working on a live streaming experiment where I already have a webrtc swarm for the metadata but I'll need something to distribute the payloads
01:14:07  <substack>the clients will already be connected to each other, so it would save some latency to use existing connections in webtorrent if possible
01:15:25  <feross>substack: the bittorrent protocol doesn't support this - it just opens multiple connections to the same peer
01:15:43  <feross>the handshake contains which info hash the connection is to be used for right up front
01:15:51  <feross>but maybe we can hack something for webtorrent
01:16:00  <feross>since webrtc allows multiplexed datachannels over the same peer connection
01:16:27  <feross>maybe if we notice that the peer id that we're about to connect to is the same as one we already have, we just reuse that peer connection and open a new data channel on it
01:16:31  <feross>to save resources
01:16:35  <feross>and latency
01:16:44  <feross>and lighten the load on the signalhubs
01:17:32  <feross>substack: does that make sense? the torrent protocol would remain unchanged, i.e. one bittorrent-protocol stream per swarm
01:18:26  <feross>but they'd be multiplexed over the same peer connection, and we'd need to add a reference count or something so the peer connection doesn't get destroyed if there's still a swarm using it
01:21:58  * groundwaterjoined
01:27:21  <substack>yes that makes sense
01:27:52  <substack>I think the optimization with a live streaming swarm is that the nodes you're already connected to for the update metadata will also have the archived chunks
01:36:52  * groundwaterquit (Quit: My MacBook has gone to sleep. ZZZzzz…)
01:47:12  * pfrazequit (Remote host closed the connection)
01:54:25  * pfrazejoined
01:59:29  * pfrazequit (Remote host closed the connection)
02:10:52  * pfrazejoined
02:17:58  * pfrazequit (Remote host closed the connection)
04:11:14  <feross>substack: right
05:04:24  * contrahaxquit (Quit: Sleeping)
05:51:19  * groundwaterjoined
07:45:08  <substack>now seeing how far I can get using swarmlog+webtorrent for the new hyperboot
07:46:08  <substack>eventually a custom peering scheme will have lower latency but this will be fine to start with
07:51:28  * rusintezjoined
08:31:02  * fotoveritequit (Quit: fotoverite)
08:38:45  * fotoveritejoined
08:39:16  * groundwaterquit (Quit: Textual IRC Client: www.textualapp.com)
08:53:02  * fotoveritequit (Quit: fotoverite)
09:03:17  * rusintezquit (Remote host closed the connection)
11:22:18  * jhermsjoined
13:27:31  * pfrazejoined
13:45:11  * jhermsquit (Ping timeout: 276 seconds)
14:00:01  * Guest62409quit
14:00:23  * Guest62409joined
14:00:28  * fotoveritejoined
15:36:06  <mafintosh>super close to having https://github.com/mafintosh/utp-native be stable as well
15:36:14  <mafintosh>i know c++ now
15:41:36  <mafintosh>i wanna use a dns server as a hole punching server to get good p2p connectivity
15:41:50  <mafintosh>+ the bt dht should help as well
16:55:14  * posejoined
17:04:27  <yangwao>use dns server as a hole punching? well, thats pretty golden idea :)
17:08:15  * posequit (Remote host closed the connection)
17:08:52  * posejoined
17:12:58  * posequit (Ping timeout: 240 seconds)
17:17:09  * pfrazequit (Remote host closed the connection)
17:20:01  * posejoined
17:30:08  <ralphtheninja>mafintosh: "i know c++ now" <-- that made me think of matrix :D
17:30:24  <mafintosh>that was my intention :D
17:30:28  <ralphtheninja>substack: mind publishing a new version of stream-adventure?
17:30:30  <ralphtheninja>haha
17:32:06  * posequit
17:35:56  <mafintosh>daviddias: thanks for adding a bunch of tests to, https://github.com/mafintosh/utp-native
17:36:38  <daviddias>no problem :)
17:37:27  <daviddias>since you are using bindings to libutp, would you say it is worth it to do interop tests as well?
17:38:06  <mafintosh>daviddias: how would that work?
17:38:53  <daviddias>firing a process with ucat and connecting to it through utp-native and check if nothing goes crazy
17:39:25  <mafintosh>daviddias: ah
17:39:57  <mafintosh>daviddias: if i can get ucat compiled on all platforms yea. i'm migrating the liputp makefile to gyp
17:40:03  <mafintosh>daviddias: so it'll work on windows
17:40:37  <daviddias>awesome!
17:57:13  * pfrazejoined
18:01:36  * pfrazequit (Ping timeout: 245 seconds)
18:42:50  * pfrazejoined
21:45:32  <jjjohnny>retweet with happened before this https://twitter.com/johnnyscript/status/696086058337849344
23:00:15  * fotoveritequit (Quit: fotoverite)