<jbenet>
hey daviddias as you go through all this-- may be worth it if we work on expanding https://github.com/ipfs/specs/ -- we can do this together as you tackle each subcomponent?
zabirauf has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<daviddias>
to organize ideas, but right now it is more a mind map than proper documentation
<jbenet>
sgtm
<jbenet>
lgtm
spikebik1 has joined #ipfs
chriscool has joined #ipfs
<daviddias>
also, one result is that I'm writing expected api specs (have to write down more there :))
chriscool has quit [Client Quit]
chriscool has joined #ipfs
chriscool has quit [Client Quit]
chriscool has joined #ipfs
<ipfsbot>
[go-ipfs] jbenet pushed 2 new commits to master: http://git.io/vqZW5
<ipfsbot>
go-ipfs/master 7ede3f4 Sean Lang: Add link to init system examples - closes #1430...
<ipfsbot>
go-ipfs/master 1694b79 Juan Batiz-Benet: Merge pull request #1447 from slang800/patch-1...
<ipfsbot>
[go-ipfs] gatesvp opened pull request #1448: Add ignore and hidden file support to add (master...master) http://git.io/vqZlT
spikebik1 has quit [Remote host closed the connection]
spikebike has joined #ipfs
gatesvp has joined #ipfs
chriscool has quit [Read error: Connection reset by peer]
chriscool has joined #ipfs
<gatesvp>
@jbenet, there's the `.ipfsignore` and `hidden` feature with all of the changes manually "smushed" together and `sharness` tests added.
zabirauf has joined #ipfs
<gatesvp>
it's now one change ahead of ipfs/master, so it should be mergable (instead of my previous multi-week mess)
<jbenet>
hey whyrusleeping are you going to finish https://github.com/ipfs/go-ipfs/pull/1420 and batching before tomorrow? because my iprs package uses an extracted out merkledag and thus extracted out blocks (it has to because iprs requires local dagstore that does not depend on network). if yes, i can rebase my work on top, else if mine is done first it will be
<jbenet>
harder to rebase your work on top. (maybe i can keep it dangling or something)
<jbenet>
gatesvp: awesome!
<gatesvp>
as always feedback accepted and welcome... though I would also be happy to get this add and make future fixes as bugs b/c I think it's finally "there"
<jbenet>
gatesvp sounds good
<jbenet>
will only comment on major things, thanks so much for pushing this
<gatesvp>
currently programming from a different country and timezone... the internet is amazing :)
kbala has quit [Quit: Connection closed for inactivity]
rollick has joined #ipfs
<cryptix>
gmorning
<jbenet>
cryptix: morning
<cryptix>
hey jbenet how is live treating you?
<jbenet>
cryptix: good :)
<jbenet>
cryptix: yourself?
<cryptix>
heat stole my sleep - otherwise good :)
<cryptix>
like all nothern germans we also like to bitch about good wether ^^
<cryptix>
i*
<fd0>
cryptix: all Germans complain about the weather all the time, be it good or bad :P
<cryptix>
fd0: true :)
<cryptix>
beeing from hamburg i still wonder how many people dare to complain about good weather, thou :))
<cryptix>
rawtaz: re zfs: i use zfs send and recv but not as in 'zfs send set@snapA | ssh root@remote recv remote/set' - the pipe into ssh is replaced with multiple steps and there is no direct connection between the pool and the remote backup site
<crest>
fd0: depends on what you call good weather
<cryptix>
crest: imho >15degC plus a couple of days without rain
<crest>
cryptix: that i like and nobody will bitch about it here (bremen)
<crest>
but 37C and 70% humidity suck
<crest>
but they provide a good excuse to check the server room AC in person
<cryptix>
yea but we wont have to tolerate that for long
<cryptix>
hahaha
<fd0>
I see, many fellow Germans in here
domanic has quit [Ping timeout: 250 seconds]
zabirauf has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
simonv3 has quit [Quit: Connection closed for inactivity]
infinity0 has quit [Remote host closed the connection]
Leer10 has quit [Ping timeout: 252 seconds]
infinity0 has joined #ipfs
anshukla has joined #ipfs
anshukla has quit [Ping timeout: 246 seconds]
domanic has joined #ipfs
chrisr_ has quit [Ping timeout: 252 seconds]
hellertime has joined #ipfs
<pguth2>
:)
<pguth2>
Not just our government loves american products and services.
_whitelogger_ has quit [Ping timeout: 252 seconds]
_whitelogger_ has joined #ipfs
domanic has quit [Ping timeout: 244 seconds]
carver has quit [Ping timeout: 256 seconds]
<fd0>
pguth2: I'm using textsecure
<fd0>
and I know moxie and the protocol is good
<fd0>
it's somewhat annoying that the messaging server is hosted in the usa, but yeah.
<jbenet>
fd0 what does the server need to do?
domanic has joined #ipfs
<doublec>
jbenet: it's used to store and forward messages. They also use google and apple's push servers I believe.
<fd0>
jbenet: messages are sent via HTTP POST (tls with certificate pinning), the server notifies the recipients via google cloud messaging or apple push, and the messages are retrieved via HTTP (+tls)
<fd0>
their cryptographic protocol is really good
<fd0>
perfect forward secrecy per message =)
<doublec>
I believe they also store a bunch of precomputed keys from clients to make the intial connection to a new user that may not be online work better
<fd0>
possible
<doublec>
The main issue I have with TextSecure (and it's minor) is that there's no multi-device support and I have to give out my phone number.
<doublec>
And because of the google and apple push server requirement, can't run it behind tor.
anshukla has quit [Remote host closed the connection]
carver has quit [Ping timeout: 265 seconds]
therealplato has joined #ipfs
<whyrusleeping>
git is being weird...
<pjz>
?
<whyrusleeping>
i just reset my branch to the remotes version of it
<whyrusleeping>
so i have a merge commit on top now, that i didnt have locally before
<whyrusleeping>
and now i'm doing a 'git rebase -i master'
<whyrusleeping>
and the merge commit doesnt show up
<whyrusleeping>
and after the rebase, its gone
<pjz>
weird
<whyrusleeping>
yeah...
anshukla has joined #ipfs
anshukla has quit [Ping timeout: 255 seconds]
<ipfsbot>
[go-ipfs] wking created tk/file-ls-help-fixup (+1 new commit): http://git.io/vqC1Z
<ipfsbot>
go-ipfs/tk/file-ls-help-fixup b304eef W. Trevor King: core/commands/unixfs/ls: Drop multi-column example from --help...
<ipfsbot>
[go-ipfs] wking opened pull request #1449: core/commands/unixfs/ls: Drop multi-column example from --help (master...tk/file-ls-help-fixup) http://git.io/vqC18
<jbenet>
yep +1, everyone take a moment to write up update so we can paste it in
<mafintosh>
jbenet: there was a logging where webrtc would become unresponsive after writing some log messages
<mafintosh>
jbenet: which meant that long running processes always broke down after a while
<jbenet>
mafintosh: ah
<mafintosh>
jbenet: they fixed it in master. not sure if its out yet
<whyrusleeping>
alright, ere goes:
<whyrusleeping>
This last week I worked on getting my PR's for version 0.4.0 all fixed up and ready. I also implemented multiplex in go, and its faster that spdystream. which is kinda odd, but whatever.
<whyrusleeping>
Things remaining from last week are my GC PR which should be ready (pending a final review)
<whyrusleeping>
and merging the multistream muxer into dev0.4.0, I just fixed a bug in a test on that one.
<whyrusleeping>
i also made sure that go-multistream and node-multistream were compatible
<jbenet>
<whyrusleeping>
so now we can move forward with connecting node and go :)
<whyrusleeping>
yeap! i hope to get that in soon as well
<whyrusleeping>
it really does good things for add eprf
<whyrusleeping>
perf*
<jbenet>
(can discuss this later, but will want to merge your PRs before we use the record pkg)
<jbenet>
ok anything else whyrusleeping?
<whyrusleeping>
not that i can think of, although the coffee hasnt quite kicked yet...
<jbenet>
ok, next, krl ?
<daviddias>
I can go next :)
<daviddias>
here it goes:
<jbenet>
(not sure who's here, was going in order, but-- ) sure!
<daviddias>
- spdystream (go) and spdy-transport (node) now interop (mux all the streams!! :))
<daviddias>
- wrote a bunch of tests and docs for spdy-transport
<daviddias>
- contributed to multicast-dns(node) to get it working with the go impl (mdns)
<daviddias>
- started working in a more complete DHT spec which eventually led me to start writing things for both the Routing Layer and Network layer entirely
<daviddias>
- In the node-ipfs land
<daviddias>
- We have mDNS Discovery working, peers are found and added to the DHT
<daviddias>
- DHT does DHT things, meaning k-bucketing peers added, splits, and math to find 'closest peers'
<daviddias>
Also, learned about the hole punching hacks from @jbenet , pretty excited about that
<daviddias>
thank you :)
<jbenet>
daviddias: anything else?
<jbenet>
hahaha
<jbenet>
ok next? lgierth?
<krl>
o2
<lgierth>
yup
<lgierth>
c&p incomin
<krl>
jbenet: forgot about time shift
<jbenet>
(krl go after lgierth?)
<lgierth>
landed the :5001/debug/metrics/prometheus endpoint, with a couple more fixes to the ipfs_p2p_peers_total metric
<lgierth>
prometheus is now up at pluto.i.ipfs.io/prometheus and collecting metrics from the gateway ipfs daemons. it'll disappear behind auth in a bit.
<lgierth>
i'm giving promdash a try before going with grafana (PR incoming). the latter's prometheus integration doesn't seem to be very stable or up-to-date.
<lgierth>
published gateway ansible scripts at github.com/ipfs/infrastructure
<lgierth>
integrated ssh authorized_keys into ansible scripts
<lgierth>
c&p eof
<lgierth>
also i'll have the camp and battlemesh invoices / payment details for you @jbenet in a bit
vijayee_ has joined #ipfs
<jbenet>
<whyrusleeping>
i do :D
<jbenet>
lgierth
<lgierth>
:]
<jbenet>
that all lgtm. anything else?
<lgierth>
that's about it
<lgierth>
neeext
<krl>
ok
<whyrusleeping>
ALL THE FEAT/MSS-MUX TESTS PASS!!! :D :D :D
<lgierth>
:]
<krl>
I made ipfs-apps that can recursively load each other
<krl>
however i realize that the way i chose to implement it, is probably no the right way of doing it
<krl>
the way i implemented asset loading, does not really allow the app to be shipped as one mdag object
<krl>
i.e, ipmlicit links
<jbenet>
ah -- yeah will want that (we can discuss the details post sync up so this part goes faster for others)
<krl>
made a ipfs-app-tool that can be used to 'run' the app, which opens it in a browser with a small bootloader
<krl>
halfway blocked on unixfs support for js ipfs
<krl>
so that we can set up and import a unixfs directory that is the app
<krl>
which seems like a more elegant solution, than having hashes in js
<jbenet>
ah, that's nice-- can you drop the link of ipfs-app-tool and any apps into the https://github.com/ipfs/pm/issues/20 issue? (that way we can CR)
<krl>
so, in essence, false start that got me seeing the problem from a better angle, and most of things i can reuse
<jbenet>
ok, sounds good
<krl>
yep
<daviddias>
whyrusleeping: wohooo! ^5!
<jbenet>
kk, anything else?
<krl>
nop
<jbenet>
ok, cool, next up, wking or kbala around?
<jbenet>
ok i'll go
<jbenet>
- reviewed various PRs with @whyrusleeping and either merged or updated. Many are still inactive, so we can either ping them again or otherwise address them this week
<jbenet>
- made iprs package, which i'll push up later today. still needs tests, but looking good.
<jbenet>
- checked out cjdns code for embedding
<jbenet>
- didn't get to mss in go-ipfs, i'll try to get to that today
<whyrusleeping>
i can still manually use it though
<jbenet>
yeah
<whyrusleeping>
was just surprising, thats all
<jbenet>
that's not ideal
<jbenet>
for relay, we should think about how to restrict or bw-cap the "enabled-by-default" relay traffic, so that NAT punching can go through any relays easy, but larger traffic doesn't
<ogd>
jbenet: i think it shows you yourself as well as other participants
<ogd>
jbenet: it will arrange all participants in a big circle
<jbenet>
ogd: but i'm the only one in
<ogd>
jbenet: did someone else join your server?
<jbenet>
ogd: i'll restart it
<ogd>
jbenet: i can try and join. its possible the webrtc stack broke since that was written (webrtc api itself has changed a lot since then)
<jbenet>
hmmm i should screencat show you this
<jbenet>
i think it's a rendering problem
<ogd>
lol
<ipfsbot>
[go-ipfs] whyrusleeping force-pushed feat/mss-mux from fefbdfe to 2474247: http://git.io/vtN3H
<ipfsbot>
go-ipfs/feat/mss-mux d7ccde4 Jeromy: WIP: using multistream muxer...
<ipfsbot>
go-ipfs/feat/mss-mux 84cbc11 Juan Batiz-Benet: multistream mux: fix ID service stream...
<ipfsbot>
go-ipfs/feat/mss-mux 8674278 Jeromy: make the relay service use msmux...
domanic has joined #ipfs
dread-alexandria has joined #ipfs
<kbala>
whyrusleeping: hey for netsim how come you use a buffer and write every 4 ms in transport instead of just writing the messages as you receive them? is it a performance thing?
<whyrusleeping>
kbala: it mimics the tcp ring buffer for coalescing outgoing writes
anshukla has joined #ipfs
anshukla has quit [Ping timeout: 256 seconds]
pfraze has joined #ipfs
<jbenet>
ogd thanks that was awesome
<jbenet>
ogd +9001 engelbart would be proud
<ogd>
haha
<ogd>
sad it didnt work from my living room server and we had to resort to the cloud though
<jbenet>
its ok, we're getting there
<jbenet>
soon you may be able to ask ipfs dht nodes to help you with webrtc signaling
<jbenet>
p2p all the things
* whyrusleeping
looks forward to that
<whyrusleeping>
i also really want ipfs pipe to arbitrary peers
<dread-alexandria>
hey @jbenet, i’m going to be in SF Thursday & Friday, can we meet up for coffee sometime?
heliumcraft has quit [Quit: Be back later ...]
chriscool has quit [Ping timeout: 255 seconds]
therealplato1 has quit [Ping timeout: 256 seconds]
therealplato has joined #ipfs
<jbenet>
hey dread-alexandria: i'm actually in Seattle until I go off to Berlin
<jbenet>
but can do a +hangout
<whyrusleeping>
mafintosh: why doesnt playback work on linux yet :(
<dread-alexandria>
oh shoot, ok, then lets plan to do that when im back at my desk next week, when do you leave for Berlin?
<mafintosh>
whyrusleeping: my friend had it running on ubuntu i think with mac ui - which part isnt working?
<whyrusleeping>
the part on your readme that says its not supported
<whyrusleeping>
i'll try running it and see what happens
<mafintosh>
whyrusleeping: ah yea we don't have linux builds thats true