chattered has quit [Remote host closed the connection]
everywhen has joined #ipfs
corvinux has quit [Ping timeout: 258 seconds]
everywhen has quit [Ping timeout: 240 seconds]
corvinux has joined #ipfs
tilgovi has quit [Ping timeout: 250 seconds]
nunofmn has joined #ipfs
herzmeister has quit [Quit: Leaving]
herzmeister has joined #ipfs
nunofmn has quit [Ping timeout: 260 seconds]
_whitelogger has joined #ipfs
SuperPhly has joined #ipfs
<whyrusleeping>
ShawkHawk[m]1: what OS?
<ShawkHawk[m]1>
mac 10.9
<whyrusleeping>
Ah, yeah.
<whyrusleeping>
i don't think docker volume mounting stuff works on OSX
dominic_ has quit [Ping timeout: 250 seconds]
<ShawkHawk[m]1>
i'm not sure that's true, i've successfully used the -v flag with other docker images
<whyrusleeping>
yeah?
<whyrusleeping>
huh
<ShawkHawk[m]1>
yeah i know, it's weird
SuperPhly has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
Falconix has quit [Ping timeout: 260 seconds]
Falconix has joined #ipfs
tilgovi has joined #ipfs
Mateon1 has quit [Remote host closed the connection]
Mateon1 has joined #ipfs
FrozenFox is now known as ThawedFox
everywhen has joined #ipfs
fleeky_ has quit [Remote host closed the connection]
Oatmeal has quit [Ping timeout: 245 seconds]
cemerick has joined #ipfs
O47m341 has quit [Ping timeout: 258 seconds]
fleeky has joined #ipfs
anonymuse has joined #ipfs
herzmeister has quit [Quit: Leaving]
herzmeister has joined #ipfs
horrified has quit [Quit: quit]
lkcl_ has quit [Ping timeout: 250 seconds]
lkcl_ has joined #ipfs
horrified has joined #ipfs
captain_morgan has quit [Remote host closed the connection]
captain_morgan has joined #ipfs
c5e3 has quit [Quit: bye]
c5e3 has joined #ipfs
c5e3 has joined #ipfs
c5e3 has quit [Changing host]
slothbag has quit [Quit: Leaving.]
infinity0 has quit [Ping timeout: 245 seconds]
cemerick has quit [Ping timeout: 246 seconds]
infinity0 has joined #ipfs
tilgovi has quit [Ping timeout: 256 seconds]
mguentner2 has quit [Quit: WeeChat 1.6]
kulelu88 has quit [Quit: Leaving]
cemerick has joined #ipfs
mguentner has joined #ipfs
<deltab>
ShawkHawk[m]1: fsync is a system call to say "make sure all the changes have been sent to the disk before continuing"
<deltab>
so I don't think it's anything to do with permissions; more likely it's something in the docker filesystem stuff that's not able to pass the fsync message on
anonymuse has quit [Remote host closed the connection]
<daviddias>
How can I be more proactive in testing orbit with js-ipfs>
<daviddias>
?
<jbenet>
daviddias thinks it's the gateway not understanding dag-cbor?
<haad>
daviddias: you mean how to test it with Orbit?
<daviddias>
it would be sweet to be able to do `npm run dev:browser:jsipfs` & `npm run dev:browser:goipfs` & `npm run dev:electron:jsipfs` & `npm run dev:electron:goipfs`
<daviddias>
yeah, where do I plug things in
<haad>
daviddias: in the client, so orbit/client/ is where you can npm link the latest js-ipfs in (afair you need to delete the node_modules/ in js-ipfs after linking, otherwise it'll pull some wrong dependencies in)
<haad>
daviddias: 'npm run build' is what builds the client to dist/ and once built, you can open index.html in dist/
<daviddias>
Oh, understood now :)
<haad>
I'm close to take out the electron app code from that repo, so after that there will be only one place to look at :)
<haad>
(I know, it's a bit confusing atm)
<daviddias>
Is this the 'state of the build', or something that is for testing the current iteration? Will there be a more streamline version of just firing up the several options? (I really like my npm scripts idea, but I might be biased :D)
<whyrusleeping>
jbenet: the gateways don't understand ipld well
<whyrusleeping>
they should be basically running 'tree'
<haad>
daviddias: it's not quite structured that way as you propose in your npm commands. the UI will differentiate between "am I in electron or in the browser" and based on that will use js-ipfs or get the js-ipfs-api instance from electron. so atm it's not really possible to run "npm run dev:electron:jsipfs" because that is the same as "npm run browser", and "npm run dev:browser:goipfs" won't run because the app will think it's in the browser and use js-ipfs.
<jbenet>
whyrusleeping right. im not sure what i woudl expect it to show me-- we probably need to think a bit about the gateway pages listing
<haad>
daviddias: let me think about it for more than 1min and I'm sure we can make it more streamlined and granular for you. then again, the most streamlined thing exists atm: 'make' :)
<jbenet>
to consider the case of listing things that are trees but not directly dirs, may have values
<jbenet>
this would porbably be a lot easier to think about if the gateway had an `/ipld/` route that showed the raw data
<daviddias>
oh, I would be fine if these are make commands too
<daviddias>
I just want to be able to fire a bunch of sessions with the different combinations
<daviddias>
and test them simultaneously
<daviddias>
and make tweaks as I go
<haad>
daviddias: are you testing interop? or just orbit-to-orbit?
<daviddias>
I just have floodsub interop
<daviddias>
and libp2p interop
<daviddias>
orbit interop I do not have
Kingsquee has joined #ipfs
rendar has joined #ipfs
<haad>
so you want to see the interop working in orbit between go-ipfs app and web app? what address do you need to set in the web app for it to connect to go-ipfs?
<daviddias>
the websockets one
<daviddias>
you don't need to set any special add in js-ipfs
<daviddias>
just go-ipfs
<haad>
daviddias: so you need to be able to configure those addresses in the ui, right? I need to add that back in to the web version then, took it out as all it uses is the signal server address
<daviddias>
that would be interesting, but even simpler than that
<daviddias>
you just need to do swarm.connect (from the browser one) to the websockets multiaddr of the go-ipfs one (in electron)
<daviddias>
or, you can just spawn two electron ones, one with go-ipfs and another with js-ipfs and it should be good to go as well
<haad>
daviddias: hmmm. so, there's no "connect to peer" functionality in the UI atm but everything is automatic. and for "two electrons", like said, the electron app only runs go-ipfs+js-ipfs-api and browser only js-ipfs. I could change it so that you can choose whether to "use go-ipfs daemon" or "use nodejs ipfs" but that's a bit more work.
<haad>
daviddias: do you need to know the "websocet s multiaddr of the go-ipfs one" atm and (manually) connect them? or can the ws address be discovered and automatically dialed to by libp2p?
<haad>
I'm just trying to think what the best approach here is so that we don't do unnecessary UI work but can still cover the test cases
<jbenet>
whyrusleeping is ipld broken out into a package yet or what?
<jbenet>
go-ipld *
<haad>
daviddias: regardless, if you want to try to test it quickly in Orbit, I would recommend to put the swarm.connect call ion line 71 in client/src/components/stores/IpfsDaemonStore.js
<haad>
then npm run build and refresh the index.html
<daviddias>
re: manually connect websockets -> unfortunately yes, there is no cross runtime discovery at the moment for go and js (only for browser and nodejs)
<jbenet>
haad: send us an orbit build when you're done today, we need to iterate fast for the demos.
<haad>
jbenet: will definitely do
<haad>
daviddias: I don't think I understand everything from that ^ so that I can test the interop today (I mean: it doesn't say how to test it)
<haad>
daviddias: if you can leave me instructions before you head to bed, I can try to test it today
<daviddias>
Ok, will write at the PR so it gets saved
<haad>
daviddias: re. https, it's not working out of the bo you're right, but I've gotten it to work in Chrome by clicking "load unsafe scripts" on the first load.
<haad>
daviddias: cool, thanks
<jbenet>
whyrusleeping: could you roll utp under an experimental flag in the next day? along with your multiplex?
<jbenet>
daviddias: does utp work in node js-ipfs?
<jbenet>
and whyrusleeping daviddias: websockets go-ipfs <--> js-ipfs works, right?
<daviddias>
right
jchevalay has quit [Quit: Page closed]
<haad>
daviddias: btw. last week when I tried webrtc-star (sig-star) locally, I just got an error message "no protocol for <some number>", was running it locally. anything specific to running it that I need to know?
<daviddias>
jbenet: re: uTP, 'it works™' as in: It works till you issue a `swarm disconnect`, once you do that, once of the sides will blow up the stream and the other side will make spdy feel weird causing the process crash. Actually, this might not be a problem anymore since we handle those havoc situations way better
<daviddias>
haad: I haven't seen 'that error' before, can you replicate
<haad>
daviddias: I'll see if it reproduces today and open an issue if that happens
jchevalay has joined #ipfs
<daviddias>
thank you :)
SuperPhly has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
SuperPhly has joined #ipfs
SuperPhly has quit [Client Quit]
ylp has joined #ipfs
corvinux has joined #ipfs
lkcl_ has quit [Ping timeout: 256 seconds]
lkcl_ has joined #ipfs
pocketss has quit [Quit: ChatZilla 0.9.93 [Firefox 50.0.2/20161130094234]]
chungy has quit [Ping timeout: 260 seconds]
<A124>
Hey there. Missed last week, gotta read up.
<A124>
Serious question: How could/should I proceed minimizing the memory footprint for the ipfs daemon?
corvinux has quit [Ping timeout: 250 seconds]
brodo has quit [Remote host closed the connection]
chungy has joined #ipfs
lkcl_ has quit [Ping timeout: 248 seconds]
lkcl has quit [Ping timeout: 250 seconds]
corvinux has joined #ipfs
jokoon has joined #ipfs
jokoon has quit [Read error: Connection reset by peer]
s_kunk has quit [Ping timeout: 240 seconds]
ecloud has quit [Ping timeout: 265 seconds]
bastianilso_____ has joined #ipfs
<Kubuxu>
go-ds-metrics needs to be propagated to go-ipfs it would save ~10-20MiB
lkcl has joined #ipfs
ecloud has joined #ipfs
gavinmcdermott has joined #ipfs
<A124>
Yeah, I am more talking about 50-75% reduction.
<A124>
Looking into profiling memory in go. And other things. Found some notices about leveldb eating some.
palkeo has quit [Quit: Konversation terminated!]
s_kunk has joined #ipfs
chriscool has joined #ipfs
ygrek_ has quit [Ping timeout: 264 seconds]
lkcl_ has joined #ipfs
PseudoNoob has joined #ipfs
Kingsquee has quit [Quit: Konversation terminated!]
wallacoloo___ has quit [Quit: wallacoloo___]
pringles has joined #ipfs
<pringles>
s
pringles is now known as pringless
<pringless>
Hey I read that IPFS is shifting to curve ec25519, but cannot find details on the KDF used for the actual cipher. Any idea guys?
<Kubuxu>
pringless: we are in a weird state, currently we are rocking custom crypto protocol based of one of the TLS 1.2 modes
<Kubuxu>
we hope that TLS1.3 impls are done soon so we can move to that
<pringless>
ec25519 for key exchange and then cant the secret key be used to generate encryption based on aes192? Nodejs has native support for it
<Kubuxu>
we will be moving to ed25519 as main node key (currently it is RSA)
<Kubuxu>
yes it could be, but we don't want to relay on self-crafter crypto protocol
<Kubuxu>
we would prefer to use something already vetted
<pringless>
aes been there for ages. isnt is peer reviewed adequately ?
<Kubuxu>
it isn't about AES only, it is about how do you do noncing, seq tracking, authentication and so on
<Kubuxu>
and AES especially has many pitfalls you are not expecting that can seriously downgrade the security
espadrine has joined #ipfs
<pringless>
oh so you are using TLS?
<Kubuxu>
we would love to move to TLS 1.3 when implementations are out
<pringless>
If key exchange is so secure
<pringless>
then the KDF should not matter as much correct?
<Kubuxu>
KDF is only small part of the equation
<pringless>
you heard of Signal from open whisper systems?
<Kubuxu>
yes, read to paper too
<pringless>
in there, the key exchange between two users chatting
<pringless>
follow ecdh followed by a KDF. IN that case, can kdf be aes192 (cause afaik it is) or TLS is actually needed as noncing, seq tracking is not really used
<Kubuxu>
it is different, they do ecdh to exchange key "pair" and then KDF to derive per message key
<Kubuxu>
it isn't necessary in encrypted wire communication
<pringless>
ohk so for VoiP types aes192 is fine
<Kubuxu>
you would use ecdh to derive "session" key
<Kubuxu>
and then use it in with AES, Salsa20 ChaCha20 to encrypt the comms
<pringless>
isn't it open to brute force by goventment spy agencies
<Kubuxu>
brute force with those algs will never be an option
<Kubuxu>
ok, not never
<Kubuxu>
never say never ;p
<pringless>
cause Signal uses this thing called ZRTP which claims to have complete detection of mitm attack during the key pair exchange
<pringless>
:D
<pringless>
but I read NSA made some voice simulator to crack that as well ...
<Kubuxu>
:/
<pringless>
cause you get two words and the parties have to say one each, unless you recognize the voice isn't it a flop
<pringless>
like If i do a voip with you, I have no clue how you sound, you dont know if I am a guy or girl so
<pringless>
mitm to this zrtp used by them can still fool us both
<pringless>
is this safe? Says independent peer reviews is left Kubuxu :
<Kubuxu>
if there are not any critical mistakes in the exchange protocol and they authenticated through the channel that they think they are the same person they are
<Kubuxu>
it is
<pringless>
i mean ecdh is vulnerable to mitm entirely
<Kubuxu>
yes, so what you do is exchange the keys, encrypt the channel and through it you ask the other party to confirm its identity using its long term key
brodo has joined #ipfs
<Kubuxu>
it is the closes what you can get to "safe" of djb's crypto in JS I think
<Kubuxu>
it is based of TweetNaCl which was written by djb and people we comparing and testing it throughoutly
<pringless>
so the npm pack is usable
<pringless>
cause idk why indepedent reviews is left
<Kubuxu>
yeah, this is the package I recommend personally
<pringless>
even proving via long term key is vulnerable
<pringless>
cause that step is done after the key exchange, which is where mitm happens
treaki has joined #ipfs
<Kubuxu>
because JS is tricky and very little of cryptographers will have the domain knowledge to vet it
<pringless>
one thing I thought is the encryption is done via a server (routing), so users must sign the messages using dsa, if it fails then server rejects it and notifies users of the attack
<pringless>
this way, both sides will know if the public key via ecdh is signed correctly or not
<Kubuxu>
yup
<victorbjelkholm>
good morning world!
<Kubuxu>
\o
<pringless>
so client creates ecdsa pair, sends public key to server. Then during key exchange signs using the private key. Server checks with its public key
<pringless>
cause I noticed even Silent Circle which makes enterprise encrpted com tools, uses zrtp without any dsa in it?
<Kubuxu>
no idea
<pringless>
I have been going through source code if chat apps and havn;t found a real mitm proof protocol anywhere
<pringless>
Bleep, Telegram, silent circle, signal
<Polychrome[m]>
What about RetroShare?
<pringless>
ipfs encryption seems to be stronger than these
<pringless>
havn't seen retroshare in detail
PseudoNoob has quit [Quit: Leaving]
<pringless>
Kubuxu: Hey sorry If I am harassing you :D but nacl.box.open(box, nonce, theirPublicKey, mySecretKey) suggests that the protocol does key exchange and encryption? So no need to use a kdf?
Soft has quit [Read error: Connection reset by peer]
Soft has joined #ipfs
ylp has quit [Quit: Leaving.]
pfrazee has joined #ipfs
<lgierth>
flyingzumwalt: we do the "what did you work on" round right after the roll call and the call for additional agenda items, yeah?
<lgierth>
PSA: sprint calls starting in 45 minutes: https://github.com/ipfs/pm/issues/285 -- add your agenda items if you want to talk about something
PseudoNoob has joined #ipfs
muvlon has quit [Quit: Leaving]
<richardlitt>
sprint-helper: next
<sprint-helper>
There are no more events today.
<richardlitt>
Well, that wasn't supposed to happen.
<richardlitt>
The All Hands call is in 30 minutes, ya'll.
<lgierth>
richardlitt: i'll be moderating, we do the "what did you work on" round after the roll call yeah?
<richardlitt>
Sounds good to me. I would ask the team and gauge whether they want it again (most likely very yes)
ebel has quit [Ping timeout: 260 seconds]
ebel has joined #ipfs
<daviddias>
haad: checking the notes you left
<haad>
daviddias: there's a bunch of them :)
<daviddias>
Re: restart nodes - yeah, that is because floodsub is not broadcasting in an interval the subscriptions
<daviddias>
Can you add a test to interface-IPFS-core so that we cover that ?
<daviddias>
Ahah true
<haad>
my day is over after the calls, so you'll have to pick that up
<Kubuxu>
haad: +1
<haad>
I'm not sure why it's related to broadcasting, shouldn't the restarted node broadcast to the other peers that it subscribed again?
<daviddias>
With regards to bitswap and Cid, I believe that what dignifiedquire was the trying to say is that Cid bitswap is on the way and has a ton of changes, changing the old bitswap might not be efficient use of time and it is better just to ship the new one
<Kubuxu>
5mins till the stream
kulelu88 has joined #ipfs
<Kubuxu>
daviddias: I think whyrusleeping integrated it nicely in go-ipfs
ulrichard has joined #ipfs
<haad>
daviddias: do you think we can do that in 2 days? :)
sprint-helper has quit [Remote host closed the connection]
<daviddias>
Maybe not, there is so much stuff today
sprint-helper has joined #ipfs
<daviddias>
When this happens, please ( pretty please) add tests
<daviddias>
It means the tests of pubsub are comprehensive
<daviddias>
I believe we can make it nicer
<daviddias>
lgierth: haven't looked into port binding again, today! :)
<richardlitt>
sprint-helper: next
<sprint-helper>
The next event is "All Hands Call", in 3 minutes.
<richardlitt>
:D!
<richardlitt>
Woohoo!
<haad>
daviddias: the down side of all that bitswap is that orbit browser version is not working without the fixes as it can't find the blocks
<daviddias>
It is not working at all?
<Kubuxu>
richardlitt: would it be possible to have the call link 5mins early
<haad>
daviddias: nope :/
<richardlitt>
Working on it now.
<Kubuxu>
so people can start coming into early
<daviddias>
Wait, how did it work past Friday ?
<Kubuxu>
so we don't start late
<Kubuxu>
(as usual :/)
<haad>
daviddias: I believe the screenshot I posted earlier than reporting the bug contained some old code so it worked.
jchevalay has quit [Quit: Page closed]
<haad>
daviddias: dunno what was changed between Friday and this morning in regards to bitswap but it's not working atm :/
<haad>
daviddias: GOOD NEWS is that floodsub works and it's really fast :)
<haad>
daviddias: so it's not all doomed
sametsisartenep has quit [Ping timeout: 250 seconds]
<richardlitt>
Kubuxu: Yes. I am trying to automate this so that you don't have to ask again - just fixed the sprint helper's 'next' command, took a bit longer than I expected, and I didn't have time to prepare the link a few minutes before this week. Sorry about that.
<richardlitt>
sprint-helper: next
<sprint-helper>
The next event is "All Hands Call", 5 minutes ago.
<richardlitt>
:P.
<haad>
daviddias: "have you refresh installed today" <-- oh so many times! :D
<dignifiedquire>
I write a lot of javascript, so yeahhh mostly just what Chrome gives me
avatar_ has joined #ipfs
<richardlitt>
d3
<avatar_>
Hello everyone, I have guided here to ask my question related to ipfs. You could see my question from here: http://stackoverflow.com/q/41105951/2402577 it is basically "Is it possible to mount ~/.ipfs folder to an external device?"
<A124>
The question alone is wrong. You don't mount folders.
<A124>
What filesystem is on that device?
SuperPhly has joined #ipfs
<avatar_>
Sorry for the wrong explanation. mount devices* || filesystem: linux - ext3
<A124>
And the cifs fits where?
cubemonkey has joined #ipfs
<A124>
If you are using samba, I believe there are some server/client or both switches to enable more granular access to the filesystem on the other side.
arpu has joined #ipfs
bsm117532 has quit [Read error: Connection reset by peer]
<avatar_>
It is WD My book External Hard Drive, honestly i dont know which file system it uses.
bsm117532 has joined #ipfs
everywhen has joined #ipfs
bsm117532 has quit [Ping timeout: 258 seconds]
ashark has quit [Ping timeout: 265 seconds]
everywhen has quit [Ping timeout: 258 seconds]
<avatar_>
I was not using samba, but I will have a look as you suggested. -cifs => recommended to mount WD My book External Hard Drive.
<avatar_>
the question arise that I'm sure it's not designed to have that file system shared between nodes (simultaneously?)
<richardlitt>
sprint-helper: next
<sprint-helper>
The next event is "All Hands Call", an hour ago.
<richardlitt>
Fixing.
SuperPhly has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<daviddias>
haad: it didn't change to ipld-dag-cbor and 'it never will', the `ipfs object api` will continue to use dag-pb. The `ipfs dag api` is the one that enables you to pick.
<A124>
avatar_ Well I have hard time understanding the whole situation, where is samba, where is external device, etc. Then it is probably NTFS. Not sure I am able to help here, but to m ake things clearer, you should state, where is the drive connected. Can't talk about the design, but this seems to be cifs related problem, being that fsync is called with argument that cifs does not accept.
<avatar_>
@A124, My server is linux ubuntu, as external device probably it is NTFS as you mentioned. The documentation is lack related to WD My book External Hard Drive (it seems like a block box) , and it does not allow me to install filesystem that I want to.
<A124>
avatar_ To the contrary, the disk should look like regular device.
<keks>
avatar_: formatting allows changing the files system
<A124>
But you would have to backup the data, then re-format with ext4 or other you like.
bastianilso_____ has joined #ipfs
<keks>
formatting means initializing a new file system. You "make a new one" and thus the command is called mkfs.ext4
<keks>
avatar_: there should be a manpage for that, they really help undestanding what you are doing
bsm117532 has joined #ipfs
<avatar_>
@A124, @keks "My Book Live Home Network Drive" just allows me to format to the factory setting, I can't format it like i format an external drive. But thanks for the guide maybe I should use an external drive instead of a network drive.
ashark has joined #ipfs
<keks>
avatar_: ah, i see
<keks>
call `mount` and post the line that regards the network drive
sprint-helper has quit [Remote host closed the connection]
<daviddias>
<3
<avatar_>
@keks || what do you mean by post the line? sorry i did not get it.
<A124>
$ mount | grep '.ipfs'
<keks>
can you just mount the network drive
<keks>
somewhere
<keks>
and show the line that corresponds to that mount
sprint-helper has joined #ipfs
<A124>
keks: If I may have a question of my own, how can one get minimal build of go-ipfs? Is there anything to exclude except fuse? And what is the best way to achieve that, just GOFLAGS? I really need to squash the footprint and problems. Fresh node starts up with about 256M which is max we can afford. But something leaks, gotta drill that down or rewrite what needs to be, replace leveldb or what else.
<richardlitt>
sprint-helper: next
<sprint-helper>
The next event is "Sprint: Apps on IPFS", in 6 minutes.
<sprint-helper>
Right now, "Sprint: go-ipfs discussion" is happening.
sprint-helper has quit [Remote host closed the connection]
sprint-helper has joined #ipfs
<richardlitt>
sprint-helper: now
<sprint-helper>
The current event is "Sprint: go-ipfs discussion", which started 26 minutes ago and ends in 4 minutes.
<richardlitt>
All fixed.
<richardlitt>
:)
<dignifiedquire>
👍
kobajagi has joined #ipfs
kobajagi has quit [Read error: Connection reset by peer]
<flyingzumwalt>
jbenet agreed. that makes the readme much easier to read
<flyingzumwalt>
richardlitt could you add something like that to standard-readme? it almost helps more than a TOC
<whyrusleeping>
I think it would look good if i wasnt using a custom stylesheet for github
brodo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<pjz>
oh, looks like I missed the apps on ipfs call... oh well. http://pinbits.io/ is still going! Actually working now! Free first-come first-serve beta of 40GB to eat
<dignifiedquire>
jbenet: phaser is super fun to work with :)
<lgierth>
pjz: hah great to hear :)
<pjz>
I'm doing the business stuff now so I can go fully live soon
<pjz>
and I need to write some dashboarding stuff, I think
<pjz>
also, I'm totally open to comments, suggestions, etc.
<whyrusleeping>
btw, adding multihash codes longer than a single byte is gonna break lots of things
<whyrusleeping>
not sure is people are yet aware of this...
<whyrusleeping>
jbenet: o/
<A124>
pjz What about mine suggestions I mentioned last time (credits, resource management)?
gavinmcdermott has joined #ipfs
Encrypt has quit [Quit: Quit]
<A124>
Also you might want to show gigabyte/month also, as a figure.
everywhen has joined #ipfs
<pjz>
A124: It's on the todo list, for sure. It's going to require user accounts and etc, and to get *that* I'm going to have to write some infrastructure
<pjz>
GB/mo is a terrible measure... how many days in a month?
<pjz>
...but fair enough
<A124>
Oh and the pin image needs rework of the edges.
<A124>
720 hours a month.
<Kubuxu>
whyrusleeping: how many?
<A124>
Standard figure in bussiness work.
<pjz>
so 30 days in a month? okay
<Kubuxu>
3 or 5?
<dignifiedquire>
whyrusleeping: yeah nobody supports varints currently in multihash
herzmeister has quit [Remote host closed the connection]
<Kubuxu>
<3
herzmeister has joined #ipfs
everywhen has quit [Ping timeout: 240 seconds]
<jbenet>
whyrusleeping: do you mean because old tools will panic?
<jbenet>
whyrusleeping: yeah you're right-- we should get on updating all the tools and rolling them out. the good news is they wont fail silently.
<pjz>
is there a good multihash-validation spec somewhere?
<richardlitt>
flyingzumwalt: You're free to suggest on standard-readme
<whyrusleeping>
but at least that one should be caught by the 'valid code' check
<richardlitt>
I don't like it though. It is github-centric, and I don't think it helps very much. I find it distracting, and making the pictures probably has a very high overhead
<richardlitt>
sprint-helper: next
<sprint-helper>
The next event is "Sprint: js-ipfs", in a few seconds.
<sprint-helper>
Right now, "Sprint: Apps on IPFS" is happening.
<daviddias>
sprint-helper: next
<sprint-helper>
The next event is "Sprint: libp2p", in 30 minutes.
<sprint-helper>
Right now, "Sprint: js-ipfs" is happening.
<daviddias>
sprint-helper: now
<sprint-helper>
The current event is "Sprint: js-ipfs", which started a few seconds ago and ends in 30 minutes.
<jbenet>
haad: here's a fun test for Orbit: expose a way for the apps to add state to the channel, then grab one of these games and drop it and the channel can play together http://phaser.io/examples
<jbenet>
flyingzumwalt: thank you-- it was a bit of a clarifying question because issues were moved to in-progress that shouldn't have been according to that o/ (and david's) definition
<flyingzumwalt>
ah.
kaotisk has quit [Quit: Leaving]
herzmeister has quit [Quit: Leaving]
herzmeister has joined #ipfs
bastianilso_____ has quit [Quit: bastianilso_____]
<kumavis[m]1>
daviddias: so i've learned the ipld/resolver/data side of ipfs, now curious about the libp2p/networking side
<kumavis[m]1>
any reccomendations on how to get started understanding how it fits together? or should i just dig through the js-ipfs and see how its done there
<victorbjelkholm>
flyingzumwalt: by the wording from the PM doc, "actively working" could be today or could be this week. As I understand anyways
<kumavis[m]1>
any reccomendations on how to get started understanding how it fits together? or should i just dig through the js-ipfs daemon and see how its done there
lkcl has quit [Ping timeout: 240 seconds]
JayCarpenter has joined #ipfs
<dignifiedquire>
kumavis[m]1: you could start looking at what happens in go-online.js
<kumavis[m]1>
dignifiedquire: during a talk (i forget who / when) i got the impression that libp2p can help you dial to a node you can directly connect with (e.g. a peer over a transport you dont have)
<kumavis[m]1>
is that a real thing or just an idea?
<dignifiedquire>
you would need sth like relay for that
<dignifiedquire>
which there are plans for, but it's not there yet
<dignifiedquire>
what can happen (if you have e.g. a dht) is that B tells A about C and then A connects directly to C
<dignifiedquire>
gotta run now, but daviddias will be awake a little longer ;)
anonymus_ has quit [Read error: Connection reset by peer]
bastianilso_____ has joined #ipfs
anonymuse has joined #ipfs
<kumavis[m]1>
thanks/
cemerick has joined #ipfs
JayCarpenter has quit [Quit: Page closed]
kenshyx has quit [Quit: Leaving]
<flyingzumwalt>
victorbjelkholm yeah. We don't have consensus about what "actively working" actually means, or what it should mean. A scrum team would emphasize a wider time frame (ie. day or week). A Kanban team would emphasize a tighter time frame (ie. hour by hour).
dominic__ has quit [Quit: Konversation terminated!]
<whyrusleeping>
take a look at the ipld object here
<whyrusleeping>
valid links in that object are /code /cats/moustache /catArray/2
<whyrusleeping>
etc
<lgierth>
can a cat be named "moustache/withaslash"?
<lgierth>
cause that'd be the only way to come up with a conflict
<lgierth>
(cross-object conflict, that is)
<ansuz>
if I had a cat
<ansuz>
that's what I'd name it
s0c has quit []
<ansuz>
ls
<lgierth>
:):)
<ansuz>
erm
<ansuz>
hi
<whyrusleeping>
A slash in the name?
<ansuz>
there's a unicode slash
<ansuz>
one of cjd's evil plans
<whyrusleeping>
lgierth: the issue is that if i say 'resolve /ipfs/QmFooBar/ecks/why/zee' am i resolving /ecks/why/zee within /ipfs/QmFooBar ? or am i resolving /why/zee in /ipfs/QmFooBar/ecks ?
maxlath has quit [Quit: maxlath]
<lgierth>
yeah, can be both. and ideally mutually exclusive. that's why i'm asking if "moustache/withaslash" is allowed. because then you could had a link at "/cats/moustache" whose target has something at "/withaslash", and you could have "/cats/moustache/withaslash", and there's your conflict
<lgierth>
or you could just say inline links take precedence, or: that resolution happens within one object only
<lgierth>
should i bring it up with nicola?
nunofmn has quit [Ping timeout: 245 seconds]
rendar has quit [Quit: std::lower_bound + std::less_equal *works* with a vector without duplicates!]
<haad>
jbenet: orbit web build. not quite working correctly yet (you want to open min 3 tabs to make sure at least some of them get messages through).
Encrypt has quit [Quit: Quit]
<haad>
daviddias: ^ that build is my last commit from the pubsub PR. if I add anything after that, it doesn't work anymore (nothing gets through bitswap), and with that it kinda work like it always did (some blocks don't get through). it's using js-floodsub 0.4.1 though.
<haad>
I'm off for today. leave me a message :)
<lgierth>
whyrusleeping: damn, i know what i forgot for 0.4.5 -- symlinks on the gateway
<whyrusleeping>
mmmm
<lgierth>
that'd be reaaaally sweet for the ccc archive
<whyrusleeping>
Yeah... would have been
<lgierth>
oh i'm still gonna post that prominently on the day before
<lgierth>
same for putting the schedule on ipfs like we did adhoc last year
se3000 has joined #ipfs
cemerick has quit [Ping timeout: 268 seconds]
dmr has quit [Quit: Leaving]
Guest46549 has quit [Ping timeout: 265 seconds]
kulelu88 has left #ipfs ["Leaving"]
espadrine has joined #ipfs
ckwaldon1 has joined #ipfs
ckwaldon has quit [Ping timeout: 250 seconds]
ckwaldon1 is now known as ckwaldon
ckwaldon1 has joined #ipfs
ckwaldon has quit [Ping timeout: 250 seconds]
ckwaldon1 is now known as ckwaldon
lkcl_ has joined #ipfs
sprint-helper has quit [Remote host closed the connection]