<Confiks>
lgierth: I was hoping IPFS would solve that problem for me.
amstocker has quit [Ping timeout: 240 seconds]
<lgierth>
Confiks: what you seem to be looking for is a complete, coherent picture of the network
<lgierth>
and that's not what a dht gets you :)
<lgierth>
what it does get you, is a framework for asking around until you determine that you know enough in order to carry on
<Confiks>
Okay, that's a clear answer :)
amstocker has joined #ipfs
f[x] has joined #ipfs
atgnag__ has quit [Ping timeout: 250 seconds]
<lgierth>
M-davidar: oh good that i checked before killing processes -- you're downloading gutenberg eh? wanna download that on biham instead?
Matoro has quit [Ping timeout: 256 seconds]
corvinux has joined #ipfs
chriscool has quit [Quit: Leaving.]
chriscool has joined #ipfs
atgnag__ has joined #ipfs
<Confiks>
lgierth: Might you know if there is any guarantuee with corenet.Listen and .Dial that the peer actually has possesion of the private key associated with its peer.ID
<Confiks>
I'm looking through basic_host.go and I see that it's calling IDService.IdentifyConn, but I don't really see any signature checking anywhere.
pierrebeaucamp has joined #ipfs
Tera2342 has joined #ipfs
tinybike has quit [Ping timeout: 250 seconds]
<lgierth>
Confiks: their peerid is derived from the public key the send us during the handshake
<lgierth>
have a look at p2p/crypto/secio
<lgierth>
oh mh i see what you mean
atgnag__ has quit [Ping timeout: 272 seconds]
Matoro has joined #ipfs
<Confiks>
Alright. I guess I need to use SessionGenerator's NewSession, but I'm not too sure what to plug into its second argument, an insecure ReadWriteCloser.
<Confiks>
But I've found that seccat command, and that looks promising.
patcon has quit [Ping timeout: 250 seconds]
edubai__ has joined #ipfs
pierrebeaucamp has quit [Ping timeout: 256 seconds]
<lgierth>
Confiks: oh word, you wanna check that the peerid they claim via idservice, is the one they used to shake hands in the beginning
<lgierth>
right?
<edubai__>
how can i offer my storage space in exchange for ether ?
<lgierth>
edubai__: check out filecoin.io -- we haven't started building it yet though
<edubai__>
it's just a concept ?
<lgierth>
yeah
<edubai__>
i see, thanks lgierth
<lgierth>
edubai__: one sec i forgot something
<Confiks>
lgierth: Yes, that any data received through corenet.Listen is authenticated ('to' its peerId).
<edubai__>
?
<lgierth>
someone built a bulletin board where you could post in exchange for eth and it'd store it in ipfs
<lgierth>
that's the only thing with ipfs+ethereum that exists
zdm has quit [Ping timeout: 256 seconds]
<edubai__>
do you have a link ?
<lgierth>
yeah i'm looking through history but can't find it
drathir has quit [Ping timeout: 260 seconds]
amstocker has quit [Ping timeout: 240 seconds]
zugz has joined #ipfs
zugzwanged has quit [Ping timeout: 250 seconds]
drathir has joined #ipfs
Whispery is now known as TheWhisper
tinybike has joined #ipfs
rendar has quit [Quit: std::lower_bound + std::less_equal *works* with a vector without duplicates!]
Matoro has quit [Ping timeout: 256 seconds]
atgnag__ has joined #ipfs
NightRa has quit [Quit: Connection closed for inactivity]
Matoro has joined #ipfs
jhulten_ has quit [Ping timeout: 250 seconds]
corvinux has quit [Quit: IRC for Sailfish 0.9]
robcat has quit [Ping timeout: 252 seconds]
<M-davidar>
lgierth: yeah, I only started the download yesterday, so I can probably just restart it on the new node
<M-davidar>
I don't remember exactly what I have on Sirius, but probably best not to trash it
Tv` has joined #ipfs
anticore has quit [Remote host closed the connection]
amstocker has joined #ipfs
Qwertie- has joined #ipfs
Qwertie has quit [Ping timeout: 240 seconds]
f[x] has quit [Ping timeout: 272 seconds]
shyamsk has quit [Ping timeout: 250 seconds]
patcon has joined #ipfs
randomguy has quit [Remote host closed the connection]
amstocker has quit [Ping timeout: 240 seconds]
amstocker has joined #ipfs
shyamsk has joined #ipfs
Tera2342 has quit [Ping timeout: 272 seconds]
jhulten_ has joined #ipfs
<ipfsbot>
[go-ipfs] lgierth created makefile-ldflags (+1 new commit): http://git.io/v0lsE
<ipfsbot>
go-ipfs/makefile-ldflags c38efda Lars Gierth: makefile: add -ldflags to install and nofuse tasks...
astrocyte has quit [Quit: Leaving]
jhulten_ has quit [Ping timeout: 240 seconds]
fiatjaf has quit [Remote host closed the connection]
<ipfsbot>
[go-ipfs] lgierth opened pull request #2066: makefile: add -ldflags to install and nofuse tasks (master...makefile-ldflags) http://git.io/v0lG3
reit has joined #ipfs
r04r is now known as zz_r04r
ianopolous has joined #ipfs
echo_oddly_ has quit [Remote host closed the connection]
f[x] has joined #ipfs
echo_oddly has joined #ipfs
jamie_k_ has joined #ipfs
<lgierth>
M-davidar: cool thanks -- sirius is copying slowly but happily :)
<lgierth>
M-davidar: let me know when it's ok to start copying castor too
<lgierth>
the others have stuff there too so i'll just scp -r /home
reit has quit [Quit: Leaving]
<M-davidar>
lgierth, Yeah, I've stopped Gutenberg, so go ahead with castor
pokeball99 has quit [Quit: Connection closed for inactivity]
pfraze has quit [Read error: Connection reset by peer]
pfraze has joined #ipfs
ehd has quit [Read error: Connection reset by peer]
anderspree has quit [Read error: Connection reset by peer]
<pinbot>
now pinning /ipfs/QmbL23v6s8fysWrtXj1agBTsdr6xCZjsTNSLWhFBtAyrAt
<pinbot>
[host 1] failed to grab refs for /ipfs/QmbL23v6s8fysWrtXj1agBTsdr6xCZjsTNSLWhFBtAyrAt: unknown ipfs-shell error encoding: text/html - "<html>\r\n<head><title>502 Bad Gateway</title></head>\r\n<body bgcolor=\"white\">\r\n<center><h1>502 Bad Gateway</h1></center>\r\n<hr><center>nginx/1.9.3</center>\r\n</body>\r\n</html>\r\n"
NeoTeo has joined #ipfs
<pinbot>
[host 4] failed to grab refs for /ipfs/QmbL23v6s8fysWrtXj1agBTsdr6xCZjsTNSLWhFBtAyrAt: Post http://[fcd8:a4e5:3af7:557e:72e5:f9d1:a599:e329]:5001/api/v0/refs?arg=/ipfs/QmbL23v6s8fysWrtXj1agBTsdr6xCZjsTNSLWhFBtAyrAt&encoding=json&stream-channels=true&r=true&: dial tcp [fcd8:a4e5:3af7:557e:72e5:f9d1:a599:e329]:5001: getsockopt: connection timed out
<multivac>
daviddias: 2015-12-14 - 05:18:46 <richardlitt> tell daviddias to get ready for the endeavor tomorrow, too! https://github.com/ipfs/pm/issues/74
<daviddias>
:)
ansuz has joined #ipfs
<M-davidar>
jbenet (IRC): cool. I had some thoughts about a "programmable wiki" a while ago, but never managed to work out the details
<M-davidar>
Will be interesting to see where they go with this
<ansuz>
I just started working at xwiki
<ansuz>
(a programmable wiki)
<ansuz>
I'm improving the realtime collaborative functionality (think etherpad)
<ansuz>
M-davidar ^
<ansuz>
eventually I'd like to see about incorporating ipfs into more of their technology, but we'll have to see, it's only my second week
fazo has joined #ipfs
fazo has quit [Changing host]
fazo has joined #ipfs
<M-davidar>
ansuz: cool :). What do you think about the talk jbenet linked?
Tera2342 has joined #ipfs
<ansuz>
uhh, I just lost all my scrollback because my VPS rebooted
<ansuz>
relink plz?
<ansuz>
xwiki-core is 10 years of java code, so it's unlikely to be trivial to incorporate new storage
<ianopolous2>
Quick question guys: Is it possible to pin a block? Or does that not really make sense because a block is just random data, and not in ipld or historical format?
<wao>
The_8472: achin thanks! :)
<wao>
btw, are you guys commin to 32c3?
<davidar>
It's entirely possible, but not sure if it makes sense
<achin>
it would make sense to pin a block if that block has important data :)
<ianopolous2>
atm I get http 500 responses trying to pin something i just added as a block
reit has joined #ipfs
<davidar>
ianopolous2 (IRC): does it work using the CLI?
<ianopolous2>
and through the cmd i get Error: pin: incorrectly formatted merkledag node: Unmarshal failed.
<davidar>
Hrm, maybe it isn't possible
<ianopolous2>
yeah, I think that's the case
<davidar>
Could have sworn I've done it before though
<ianopolous2>
as it's not in a canonical format
<davidar>
You're not trying to pin recursively?
<ianopolous2>
nope, my command was "ipfs pin add QmPcwB8HS7Pgm3FwgaZnegvMiG2Fpbjrs6JdCAveMq24Tw"
wking has joined #ipfs
<davidar>
Hmm, I'm not sure why it's trying to unmarshal it, it doesn't seem like pin should care what's in the block :/
<davidar>
ianopolous2 (IRC): you're able to retrieve the block from another node, right?
<achin>
there must be something wrong with that block
<ianopolous2>
I can retrieve the block fine from the same node, haven't tried another node
<ianopolous2>
the block was just random data
<achin>
how was it created?
<ianopolous2>
created with block.put
<davidar>
achin (IRC): blocks don't necessarily have to be unixfs formatted
<ianopolous2>
exactly, it doesn't have a format
<achin>
but there is a merkeldag format -- PBNode
wolf has joined #ipfs
<ianopolous2>
I would have to wrap it with a properly formatted object (or one that links to it) to pin it I believe
wolf is now known as Guest7890
<davidar>
s/unixfs/protobuf/
<ianopolous2>
yep
<davidar>
IIRC stuffing random bytes into blocks is supported, maybe not though, it's been a while since I checked
<davidar>
If it works with some commands but not others, I'd call that a bug
<davidar>
Either it should work everywhere or nowhere
<achin>
i see, hmm.
<ianopolous2>
my understanding is that block are mostly random bytes according to whatever chunking of the files exists, and proper objects link to those raw blocks?
<davidar>
Normally, no, they're usually wrapped in protobuf everywhere
<davidar>
(even the chunk contents)
<ianopolous2>
if that's meant to be the case then yeah, that probably should be enforced
<davidar>
But then when multicodec happens, there will be formats other than protobuf too
<ianopolous2>
whyrusleeping, jbenet: is that the case? :-)
<jbenet>
ianopolous2: AFAIU, and the model i push for, shouldnt be able to store anything under blocks, though i think whyrusleeping wanted to keep that and it may work fine to block put/get
<jbenet>
ianopolous2: we will see about linking to raw data, and how to signal it.
flyingkiwi_ is now known as flyingkiwi
<jbenet>
not yet there.
<jbenet>
ok going to sleep a bit before sprint. o/
<jbenet>
see you all in a few hrs.
<The_8472>
how does one specify ipv6 link-local bind addresses? they need fe80::...%eth0 disambiguation
Encrypt has quit [Quit: Quitte]
pierrebeaucamp has quit [Ping timeout: 272 seconds]
<davidar>
ianopolous2 (IRC): might be worth submitting an issue to clarify the expected behaviour, since it's currently a bit inconsistent
<davidar>
There's not much point being able to add arbitrary blocks if you can't pin them
<davidar>
jbenet: +1 to being able to link to raw data though
neurosis12 has quit [Remote host closed the connection]
pierrebeaucamp has joined #ipfs
jedahan has joined #ipfs
Kane` has joined #ipfs
<dignifiedquire>
richardlitt: what’s up?
<multivac>
dignifiedquire: 2015-12-14 - 05:17:41 <richardlitt> tell dignifiedquire to get ready for the endeavor tomorrow for apps on ipfs: https://github.com/ipfs/pm/issues/74
<richardlitt>
Was working on the apiary docs, but decided it's probably best to just keep conversation in the PR
cemerick has joined #ipfs
hellertime has quit [Quit: Leaving.]
<richardlitt>
dignifiedquire: Although, I guess I can say I am confused by Postman/.
hellertime has joined #ipfs
<dignifiedquire>
richardlitt: :D
<dignifiedquire>
will take a look at your pr later
<richardlitt>
sometimes setting key to arg works, sometimes it doesn't. The problem for me is the lack of feedback as to why things work occasionally.
<richardlitt>
I must be missing some information.
The_8472 has quit [Ping timeout: 264 seconds]
The_8472 has joined #ipfs
ulrichard has quit [Read error: Connection reset by peer]
<dignifiedquire>
richardlitt: not sure that is an issue with postman, but rather with the go server
<dignifiedquire>
there is a lot guess/trial and error work in this..
<richardlitt>
dignifiedquire: could be. Hard for me to diagnose.
<richardlitt>
In any event, if I can't get postman to work, I use the terminal, and do my best
<dignifiedquire>
if you can’t figure out how the req should look like, use the js-ipfs-api in the browser and use chrome to inspect the conections
gperrin has left #ipfs ["Leaving"]
voxelot has joined #ipfs
voxelot has joined #ipfs
<richardlitt>
dignifiedquire: cool. Will try that if I'm lost.
Guest53124 has joined #ipfs
<daviddias>
dignifiedquire: it is really nice to read Rust code from a JavaScripter :D
<daviddias>
thank you for making it so easy for me to understand
<dignifiedquire>
daviddias: well I try my best :D
<daviddias>
yesterday I ended up spending my evening working on get Arch Linux + i3 to work on my Mac
<daviddias>
so I didn't touch the Rust things again until now
pfraze has joined #ipfs
svetter has joined #ipfs
svetter has quit [Client Quit]
svetter has joined #ipfs
hellertime has quit [Quit: Leaving.]
r04r is now known as zz_r04r
hellertime has joined #ipfs
Guest53124 has quit [Ping timeout: 252 seconds]
<dignifiedquire>
daviddias: shame on you, rust is way cooler than arch linux :P
<ansuz>
arch linux + i3 + mac
<ansuz>
hipster trifecta
<ansuz>
you win
<dignifiedquire>
I’m afraid I have no idea what i3 is..
<ansuz>
tiling display manager
<ansuz>
like tmux for everything
<dignifiedquire>
ah right
<dignifiedquire>
I like my os x :)
<ansuz>
I like my making fun of people who like things
* ansuz
grumpycats
<dignifiedquire>
to everyone their own
<ansuz>
NO
<ansuz>
etc
<ansuz>
heh
<daviddias>
ahaha
<dignifiedquire>
well on the day your arch linux runs photoshop we can talk again :P
<daviddias>
dignifiedquire: rust is cool, buy that too :)
<dignifiedquire>
daviddias: the ultimate hipster, you really need to grow a beard
<daviddias>
Arch for me also means that I can use gnu radio stuff with the radio badge I got at CCC without having to go to 1000x sheneningams of running kali on vmware
<ansuz>
lol photoshop
<ansuz>
y u no giump
<ansuz>
gimp even
<dignifiedquire>
gimp…I don’t even
<ansuz>
heh
<dignifiedquire>
Paint > gimp
<ansuz>
I was gonna say
<ansuz>
mspaint
<ansuz>
4lyfe
<daviddias>
maybe, if gimp was based on electron
<daviddias>
so that WebDesigners could easially hack on it
<dignifiedquire>
oh yeah good idea, then it would be slow and have a horrible ui
<ansuz>
cause it doesn't take enough memory
<ansuz>
:D
* ansuz
buzzkills harder
<daviddias>
there would be a ton of contributions and we would have a photoshop open source alternative that people would use
<daviddias>
dignifiedquire: well, Electron has what, months?
<dignifiedquire>
months what?
<ansuz>
total age
<daviddias>
since it was released
fazo has joined #ipfs
<dignifiedquire>
not sure you have to count in atom, which is older
fazo has quit [Changing host]
fazo has joined #ipfs
<ansuz>
daviddias: are you in EU again?
<dignifiedquire>
a year maybe
<ansuz>
I'm going to have to visit portugal
<dignifiedquire>
not saying it’s impossible to have fast things, but atom is pretty slow
<ansuz>
after all I've heard
<dignifiedquire>
and photoshop and all it’s goodness is a big beast
<dignifiedquire>
the current doc I’m working on takes 2gb uncompressed in memory
<ansuz>
:|
<ansuz>
I have 2GB ram on this computer
<daviddias>
ansuz: sweet! when are you coming?
<dignifiedquire>
I don’t want to write anything in javascript to handle that
<dignifiedquire>
ansuz: I have 16gb :D
<dignifiedquire>
for a good reason
<ansuz>
daviddias: I'm in paris, but probably broke for a while until I settle in
<ansuz>
but maybe in the summer, idk
<ansuz>
I'm going to Romania in July for company stuff
<daviddias>
Summer is nice here, you definitely should come in the Summer
<daviddias>
And Portugal is cheap :)
<daviddias>
well, most of Europe is, only when you pay taxes, it is the nightmare, in another hand, must be a paradise for tourists
wiedi has quit [Quit: ^Z]
* whyrusleeping
is awake
<multivac>
whyrusleeping: 2015-12-14 - 05:18:25 <richardlitt> tell whyrusleeping : reminder to get ready for your endeavor tomorrow: https://github.com/ipfs/pm/issues/74
<daviddias>
I'm not sure if I'll be around in the summer, but i should be, at least part of it :)
<richardlitt>
also tell him good morning, get some coffee
<daviddias>
hah
<whyrusleeping>
hrm... theres still a half full glass of whiskey on my desk...
<richardlitt>
irish coffee
<whyrusleeping>
the temptation is strong, but an empty stomache would not be happy with irish coffee
<richardlitt>
Alas, no.
dignifiedquire has quit [Quit: dignifiedquire]
dignifiedquire_ is now known as dignifiedquire
<fazo>
I remember using Atom on my macbook and witnessing huge loading times. It has improved a lot, my laptop now has half the horsepower and loads atom in less than 1 second
nekomune has quit [Ping timeout: 260 seconds]
<fazo>
that's of course if your files aren't bigger than a few kilobytes lol
The_8472` has joined #ipfs
The_8472 has quit [Read error: Connection reset by peer]
<richardlitt>
You all ready for the sync? Please add your tasks to the Sprint
<richardlitt>
Also, remember we're starting a new thing today: please also add in an asterix for the one thing you did this week that you think is best, for a newsletter thing
f[x] has joined #ipfs
<richardlitt>
Well, this is going to be a fun sync
<lgierth>
hi richardlitt o/
The_8472` has quit [Read error: Connection reset by peer]
<daviddias>
- [x] Wrote a Roadmap https://github.com/ipfs/pm/pull/71 which lead me to update https://github.com/ipfs/js-ipfs. I would prefer to merge both, but it is definitely nice to understand what is the list of all the things that need to be done and the subset of those that are our core priority
<daviddias>
- stats - I was unable to get any stats as I am always filling up disks 'disk full'. I'm using now straight registry-mirror on Pollux, so that I can pause and resume anywhere and not be surprised by a massive overhead from ipfs add -r
<daviddias>
- [x] write issue with learnings (post mortem analysis) made with registry-mirror
<daviddias>
- [x] transfer out repos from aws and stop that machine
<jbenet>
dignifiedquire: thanks for all the distributions work -- sorry this is trickier than expected re CR
<dignifiedquire>
jbenet: yeah wasn’t that happy when I read your comments :( especially after whyrusleeping told me he was thinking of moving all the shell scripts to js
<jbenet>
what shell scripts, the distribution ones?
<dignifiedquire>
I think we need a proper strategy and look at who will actually run this code/build releases
<whyrusleeping>
i think it might be easier to just use jq and shell scripts
<dignifiedquire>
yeah the ones for go
<whyrusleeping>
its easier to read than a javascript cli app
<dignifiedquire>
jq?
jbrooker has quit [Ping timeout: 272 seconds]
ianopolous3 has joined #ipfs
<jbenet>
let's please stop trying to change everything for now, and just ship the site update pls.
patcon has quit [Ping timeout: 240 seconds]
<richardlitt>
Cool
<richardlitt>
jbenet: want to go?
<pjz>
whyrusleeping: to clarify, the IPFS API is async, but the python-ipfs-api is *not*
<richardlitt>
Do we have any extra hangouts today? jbenet?
<jbenet>
thanks for running the sprint richardlitt
<richardlitt>
np
<jbenet>
dignifiedquire: yeah
<daviddias>
lgierth: I'm not being able to repo gc on Pollux (probably the node is doing some operation and it still doesn't know to freeze other operations to gc)
<daviddias>
might be, it has been stale for a while though
grahamperrin has joined #ipfs
<lgierth>
daviddias: i mean it's constanyly reading a few MB/s
<lgierth>
docker ps
<jbenet>
dignifiedquire: what's the question? what seems unsafe to you?
pfraze has joined #ipfs
<lgierth>
ups
ianopolous3 is now known as ianopolous
psych has joined #ipfs
<daviddias>
ok, will wait a bit more then
<dignifiedquire>
jbenet: I’m mostly concerned with the fact that you brute force through all the bytes to detect the codes to find boundaries, rather than doing something like:
<dignifiedquire>
- read first 2 bytes
<dignifiedquire>
- bytes -> code
<dignifiedquire>
- code -> size
<dignifiedquire>
- skip size bytes
<dignifiedquire>
- loop
<daviddias>
thank you for checking
<dignifiedquire>
which would work for all except for ipfs which has variable size :(
<dignifiedquire>
though that could be solved, just read the next x bytes for the header, and than skip that length
<dignifiedquire>
or is there something that I’m missing here?
<jbenet>
(i agree it's not the prettiest code ever, it reminds me of some bowels of golang stdlib code
<ipfsbot>
[go-ipfs] whyrusleeping created vendor/update-flatfs (+1 new commit): http://git.io/v0Rxc
<ipfsbot>
go-ipfs/vendor/update-flatfs 5c6c5a9 Jeromy: update to new flatfs code...
<dignifiedquire>
jbenet: okay, right didn’t fully understand the go I guess :/ but then I’m still not sure how you handle the unknown size of an ipfs address in that case
<harlan_>
ah sorry richardlitt missed your message... well for the record, my sprint update:
<nicolagreco>
where are the hangout links today?
<dignifiedquire>
nicolagreco: will be anounced here shortly before the hangouts start
<dignifiedquire>
jbenet: one other thing, https://github.com/jbenet/multiaddr/issues/8 would really appreciate to change the spec to just say 2bytes (also why the use of the numbers > 128 for some protocols, there are only a handful of protocols in there atm)?
<dignifiedquire>
also using varint makes the byte representation incompat between different languages as far as I understand
mildred has joined #ipfs
<dignifiedquire>
though getting language interop working would mean the code transformation needs to be part of the spec as well
amingoia has quit [Quit: Connection closed for inactivity]
<richardlitt>
jbenet: yeah. Il do that
patcon has joined #ipfs
<jbenet>
dignifiedquire: "2 bytes" no, the goal is to not have artificial limits that inevitably will be passed. we keep doing this. even ip6 gave in to the hardware people and will require another flag day eventually to switch out to something else. ip should've been a varint from the beginning, and many pushed for it back then (and continue to push for it) but were
<jbenet>
overrulled by hardware issues then. today varints are not such a big deal. (hello UTF-8!)
<jbenet>
dignifiedquire: "why the use of the numbers > 128 for some protocols" i was trying to follow IANA numbers, but this is fraught.
<jbenet>
dignifiedquire: "also using varint makes the byte representation incompat between different languages as far as I understand" no, why would it? it's just a byte representation
<jbenet>
dignifiedquire: "the code transformation needs to be part of the spec as well" ? if what you mean is "how to read the varint", yes indeed.
<daviddias>
ok, restarted repo gc too lgierth
mvr_ has joined #ipfs
warner has quit [Quit: ERC (IRC client for Emacs 24.5.1)]
wiedi has joined #ipfs
f[x] has joined #ipfs
Tv` has joined #ipfs
patcon has quit [Ping timeout: 250 seconds]
keorn has quit [Ping timeout: 250 seconds]
<achin>
hi all, thanks for starring all these great things! i'll aim to produce a draft roundup post by my end-of-day, and post it on pm/issues/73 for review
<jbenet>
achin yay. just do a PR
<achin>
a PR to the PM repo? ok sounds good
<jbenet>
achin: oh sorry, to the blog
<jbenet>
achin: ok yeah just post a roundup in the issue first, and then we can push it to a blog post
<richardlitt>
How long should `ipfs file ls <file>` really take?
<richardlitt>
It seems to be hanging for a long time
<dignifiedquire>
jbenet: thanks for the details, two things 1. I’m not a 100% which definition of varint you are referring to 2. I’m still not understanding how you parse out ipfs addresses in the go code
<dignifiedquire_>
jbenet: np, need to pass some of it to more people though so you can do this without me ;)
mildred has joined #ipfs
<ion>
jbenet: Re: the discussion, perhaps run both versions on the public gateways and see if a frontend HTTP server could query both and do a transparent fallback.
patcon has joined #ipfs
baselab has joined #ipfs
cemerick has quit [Ping timeout: 250 seconds]
grahamperrin has left #ipfs ["Leaving"]
oguz has joined #ipfs
<richardlitt>
dignifiedquire_: write a doc on how to doit
oguz is now known as ogzy
ogzy has quit [Changing host]
ogzy has joined #ipfs
<ogzy>
assume that i dont want to run ipfs daemon at my node but somehow i want to learn the ip_addresses/ipns/peer_id address, so the ip adress part is not known, any way to learn it somehow?
jbrooker has joined #ipfs
<achin>
you can run "ipfs id" even without a daemon running
<ogzy>
achin, how? in any case shouldn't i need an ipfs command?
<achin>
else look in $HOME/.ipfs/config to get your peerID
eater has quit [Read error: Connection reset by peer]
eater has joined #ipfs
<ogzy>
achin, i am at the ip part, since the content is distributed, how will i learn the nodes ip addresses?
<achin>
a node will use whatever IP addresses are on the machine
<achin>
(i'm not sure i fully understand what you're asking, sorry)
<ipfsbot>
[webui] greenkeeperio-bot opened pull request #123: Update d3 to version 3.5.11
<ogzy>
achin, ipns/peer_id can be learned from dns record, assume that i am reaching a content via some_ip/ipns/peer_id then the node is down, how will i know the other nodes?
<ogzy>
achin, and assume i dont have gateway node
<achin>
are you wondering about your own node? or someone elses node?
chriscool has joined #ipfs
<fazo>
ogzy: do you mean what happens if you're using a gateway to access IPFS or IPNS and that gateway is down?
<ogzy>
fazo, assume i am not using any gatway, i am running ipfs without bootstrap node information
<ogzy>
achin, my node is a regular client on internet
Peer3Peer has joined #ipfs
<fazo>
if you don't have any bootstrap node in the list or you can't connect to any of them, the only way to find new nodes is to manually connect to their IP or via local peer discovery in a LAN
<fazo>
but if you connect to a node, your node will also try to connect to that node's nodes (of course it tries to make as many connections as possible) I think. If you want a private network, I'm afraid there's no way to easily do that but it's planned :)
<ogzy>
fazo, how will i connect them if i dont know their ips, ok here is the clear question, i am a regular user, i questioned a domain got the ipns information and then my browser somehow redirected an ip/ipns/peer_id, but that node somehow down after a while, the content is still somewhere
<ogzy>
fazo, i dont want to be dependent on bootstrap nodes indeed
keorn has joined #ipfs
<fazo>
ogzy: it's impossible to have IPFS work without bootstrap node, unless you use cjdns or a similar protocol
<fazo>
I mean you could connect manually to nodes, or just use lan discovery
<ogzy>
himm
<dignifiedquire_>
whyrusleeping: ping
<fazo>
when you visit /ipns/domain.foo it gets translated to /ipns/<ipns_address> and then it gets resolved to /ipfs/<ipfs_address> and then that file or folder is downloaded from the best source available
<daviddias>
whyrusleeping: you up for the libp2p hangout?
<dignifiedquire_>
whyrusleeping: libp2p^^
<ianopolous>
quick question: does object.patch.set-data work in the http api? If so how is the data transferred? As far as I can see it's not in any of the api implementations.
<ogzy>
fazo, yes i got it, but the design is dependent on bootstrap nodes, thats is not decentralized :)
<fazo>
ogzy: it actually is. I'll find some clarifications about what decentralized and distributed mean for you :) I agree, it would be better without bootstrap node, and there are ways to achieve it.
<fazo>
ogzy: but to work without boostrap nodes it needs a service discovery system, and the common network stack doesn't have any that work over the internet. that's why you need cjdns.
<The_8472>
hrm, is it actually possible to configure separate bind addresses for DHT and swarm connections?
<fazo>
in other words: at the moment the internet isn't able to do that.
<ansuz>
fazo: are you on cjdns?
<ansuz>
on hyperboria, rather
<fazo>
no, but I plan to be
<fazo>
but I have limited time right now and I'm dedicating that to developing ipfs-boards
<fazo>
if anyone has got a better name for the project, please advise :)
<The_8472>
<ogzy> fazo, i dont want to be dependent on bootstrap nodes indeed <- i've seen people make a bigger deal than it is out of that with bittorrent too. bootstrap nodes are not special or important. they are very much replaceable. any node can be used to join a DHT. conceptually you could have smartphones signallying each other as people walk past each other, wifi hotspots talking to each other etc. etc.
<The_8472>
you can download a bootstrap list from anywhere
<The_8472>
it's not a SPOF
<fazo>
yeah, The_8472 you're completely in the right
<The_8472>
and once you are bootstrapped you can keep a cache of recently-seen, stable nodes so you won't ever need the bootstrap list again
<fazo>
but one day we'll have a network stack that allows distributed systems to work without bootstrapping :)
keorn has quit [Ping timeout: 240 seconds]
<fazo>
I see it as a temporary fix
<The_8472>
fazo, that either takes a gossiping mesh network or wan multicast
<fazo>
ogzy: I can't find that cool image that other guys use to show the different topologies between centralized, decentralized and distributed. It really shows well the difference, and I can assure IPFS is completely distributed.
<fazo>
The_8472: I think cjdns has service discovery built in to solve the bootstrapping problem.
<fazo>
you advertise a service so that people can find you
<lgierth>
fazo: it doesn't
jbrooker has quit [Remote host closed the connection]
<lgierth>
you can get a list of known nodes and probe them
<fazo>
oh. I thought it did. I guess I was wrong, sorry to spread misinformation
reit has joined #ipfs
<lgierth>
no worries ;)
<fazo>
lgierth: well that looks like enough to get some connections going :)
<lgierth>
best way to get the right answer is to post a wrong answer :P
<fazo>
lgierth: I'm gonna steal that lol
<lgierth>
fazo: yeah! i have that working btw, it's just missig tests
<ianopolous>
whyrusleeping, jbenet: Does object.patch.set-data work in the http api? If so, how is the data transferred? As far as I can see it's not in any of the api implementations.
<fazo>
lgierth: I should really get cjdns going :(
ipfs-gitter-bot has joined #ipfs
ipfs-gitter-bot has quit [Read error: Connection reset by peer]
<fazo>
anyone have any ideas on a name for a distributed social network over IPFS?
<ansuz>
cjdns nodes discover each other over common interfaces like ethernet or wlan, but that's not really a 'service'
awrelll has joined #ipfs
Matoro has quit [Ping timeout: 240 seconds]
<awrelll>
hello
ipfs-gitter-bot has joined #ipfs
<awrelll>
it's up now
<ianopolous>
fazo: Peergos? :-)
<ipfs-gitter-bot>
(harlantwood) Nice
<ipfs-gitter-bot>
(harlantwood) What was the fix?
<lgierth>
harlantwood: ping
<ipfs-gitter-bot>
(awrelll) i did nothing
<ipfs-gitter-bot>
(awrelll) just checked to see if it’s up
<ipfs-gitter-bot>
(awrelll) i saw those errors
<harlantwood>
hey lgierth
<ipfs-gitter-bot>
(awrelll) and that’s all
<lgierth>
harlantwood: ah ok that's the irc user
<fazo>
ianopolous: wow, nice project you have!
<fazo>
good thing I checked to see if the name was already taken ;-)
<ianopolous>
fazo: thanks. It's still in alpha stage, but progressing fast.
<ipfs-gitter-bot>
(harlantwood) @awrelll strange! well thanks for looking ;)
<ianopolous>
we only switched in IPFS 2 weeks ago to replace out DHT and NAT traversal and storage
<fazo>
ianopolous: cool :) mine has a quite different vision, but overall the projects are similar
<ianopolous>
fazo: looks cool! Who's responsible for storing the data?
<fazo>
the user :) if you try the prototype, it's written right on the landing page
<fazo>
however, I plan on implementing a "health indicator" for ALL content, that shows how many are seeding it
<fazo>
and then set up quite some servers that will seed as much as possible.
arpu_ has quit [Quit: Ex-Chat]
<fazo>
the beauty of it is that anyone can "donate" by setting up a machine that seeds. And you can choose what to seed, to show your support to people and organizations.
arpu has joined #ipfs
<The_8472>
fazo, how do things linked together? i.e. how does person A learn that B responded to them or something like that?
<The_8472>
*how are
<fazo>
check the README, all of that is explained
<fazo>
notifications won't be implemented until IPFS has pub/sub implemented, and real time features such as chatting probably won't be either until "ipfs pipe" is implemented
<fazo>
but I can't think of some feature that I know I can't do, except one: verify that a profile is managed by a human
<ogzy>
ansuz, so to work cjdns i need cjdns running?
<fazo>
there is no google analytics, no dependency, only IPFS: I want this to work in any browser out of the box with no external programs, but also in any operating system, even without internet access and without a web browser.
<The_8472>
fazo, hrrm... so that only works if people already know each other?
<The_8472>
i don't see anything in the readme that explains how "hello, i'm new here, can I join?" would work
<fazo>
The_8472: absolutely not. You can still discover content, but you have to judge yourself :)
<fazo>
it's a little too early for that, but the cool thing is that by the new year I will have exactly that
dignifiedquire has quit [Quit: dignifiedquire]
dignifiedquire_ is now known as dignifiedquire
<fazo>
The_8472: if you want to help, open an issue detailing what information is missing in the getting started section and what can be improved, I'd be very glad and you'd help me speed up development
<ianopolous>
sounds great, fazo!
<fazo>
thanks, ianopolous! :)
<The_8472>
fazo, i'm wondering whether i'm having some misunderstanding of what ipfs can do or whether your documentation falls short
awrelll has quit [Ping timeout: 252 seconds]
<fazo>
The_8472: I know, it's quite a problem: understanding how the app works requires understanding IPFS... and that isn't very easy
rombou has quit [Ping timeout: 240 seconds]
keorn has joined #ipfs
<The_8472>
i can see how two users who already know each other's pubkeys can talk to each other. they can just poll their respective IPNS records to see if the other has appended some data to a data structure. they can then even update references "this guy responded to me"
<The_8472>
but how would user C ever be able to join a conversation?
wao is now known as xiaonormie
<jedahan>
maybe people who care to advertise who they are? sharing public keys through multicast or something?
<jedahan>
@The_8472 I plan on getting some sort of public node up on a mesh network node, and wondering the same thing
<jedahan>
just like, heartbeating out to 255.255.255.255
<jedahan>
or whatever the ipv6 equivalent is, as it were
<The_8472>
jedahan, not sure which conversation you're responding to
<fazo>
The_8472: he'd need A or B's public key hashes (node IDs). Of course his instance of the app could randomly discover them.
<jedahan>
The_8472 regarding how does C find out about A or B
<The_8472>
nono, my question isn't how C finds A or B
<ilyaigpetrov>
The_8472: a chat room may be spinning on ipns address with all participants listed
<The_8472>
my question is how A/B know that C wants to talk to them
<The_8472>
don't you need the *private* key to publish something to an ipns key? the pubkey only is useful for reads, no?
<fazo>
The_8472: A and B would need to discover C (either by manually entering his IPNS address or discovering it)
<fazo>
yeah, the private key is necessary for publishing.
<ilyaigpetrov>
The_8472: you're right. I guess ipfs is more about files, not about building web services like ip
<The_8472>
fazo, so it's impossible for C to join a conversation unilaterally? to knock on the door?
<fazo>
The_8472: for now, yes, but with the implementation of pipes, it will be possible.
voxelot has quit [Ping timeout: 256 seconds]
<The_8472>
i see
<fazo>
That's why I said I'm waiting for pipes to work on notifications :)
<lgierth>
ogzy: cjdns running to run cjdns?
<lgierth>
ogzy: you can peer with nodes on the internet just fine -- ansuz was just saying that cjdns nodes on the LAN discover each other
<lgierth>
similar to ipfs bootstrap vs. discovery
<The_8472>
is there any info on pipes?
<ilyaigpetrov>
fazo: wow, pipes. Is it built upon ipns?
<jedahan>
Has anyone done any integration between scuttlebutt and ipfs?
harlan_ has quit [Quit: Connection closed for inactivity]
<The_8472>
hrrm, i'm not used to reading Go, but it looks like a direct peer-peer connection... transient, not persistent data?
<fazo>
The_8472: yes, it would work only if the receiver is online
<ilyaigpetrov>
they had streams specs somewhere, will it will be integrated with pipes?
<The_8472>
i see. i wondered whether it was some offline mechanism, publishing a small pointer "here, please look at this blob once you come online". apparently not
<fazo>
The_8472: yeah but if the receiver is offline, there's no way to send them a message
<The_8472>
that's what i just said :P
<fazo>
my idea was: in a public chatroom, when someone arrives, other users pipe to him the messages
<The_8472>
an online mechanism is probably fine for rapid communication. but for slower-moving things offline notification might be necessary
<ogzy>
lgierth, i didn't understand the usage of cjdns in my case, i watch the video and reading stuff about it, so the regular user should have cjdns running on its machine or not?
<fazo>
this is way faster than having the new guy look around random profiles to find some messages
<fazo>
The_8472: yeah my current approach is avoiding implementing rapid communication in the app
<fazo>
because it wouldn't work well with the current state of things
Matoro has quit [Ping timeout: 256 seconds]
<jbenet>
whyrusleeping doing go-ipfs ?
<lgierth>
ogzy: yeah they should, you lose a couple of guarantees if you run it e.g. only on the LAN router
<dignifiedquire>
daviddias: :cry: error: `octets` does not live long enough
<daviddias>
??
<dignifiedquire>
daviddias: trying to parse ip addresses with rust
<daviddias>
but what does "not live long enough" mean exactly?
<dignifiedquire>
:D
<dignifiedquire>
in this case it means I’m returning a slice (reference to the part of a byte array), but the original goes out of scope when the function ends
<dignifiedquire>
so I’m returning a reference to nothing
<ogzy>
lgierth, so what is cjdns is solving for me?, i am trying to find a solution for an agentless situation, for a regular user, does cjdns have browser plugins for ex, that may help also?
<dignifiedquire>
whyrusleeping: turn off your video, that should improve things
<lgierth>
ogzy: no browser plugins at this point, also need root permissions to start up -- not sure what the initial question was, just picked up
<daviddias>
dignifiedquire: woa, I see
<daviddias>
you got to return it as a value, or pass some address for some pre allocated memory then
<dignifiedquire>
yeah, but that’s what I didn’t want to do :(
<dignifiedquire>
byte arrays are hard :( (only compiled time known length..)
<ogzy>
lgierth, the problem is to reach ipfs content i need bootstrap nodes, , dont want to use them, assume i am accesing an adres that i learned from dns some_ip/ipns/peer_id, what will happen if the some_ip is down, the content is at some other nodes, how will i know the other ip?
water1_resistant has quit [Quit: Leaving]
psych has quit [Remote host closed the connection]
Dashnakist has joined #ipfs
cemerick has joined #ipfs
<ianopolous>
whyrusleeping, jbenet: I figured it out: The data in object.patch.set-data is an extra argument in the url. This is quite limiting if I want to put arbitrary binary data there (esp if I want 512kb). Thoughts about making it the request body, or a multipart to keep the same as everything else?
<ogzy>
lgierth, how about cjdns is running at the ipfs nodes, can i learn the other node information when i reach one of them?
voxelot has joined #ipfs
voxelot has quit [Changing host]
voxelot has joined #ipfs
jedahan has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<lgierth>
ogzy: sorry i'm in a hangout so a bit slow -- what kind of information do you mean?
<ogzy>
lgierth, the other ipfs nodes, their ips will be enough
Nitori_ is now known as Nitori
<lgierth>
ogzy: tools/peerStats for peerings, tools/dumptable for routing table used for answering dht lookups by other nodes, tools/sessionstats for the nodes you're communicating with
ulrichard has quit [Read error: Connection reset by peer]
tilgovi has joined #ipfs
arpu has quit [Quit: Ex-Chat]
arpu has joined #ipfs
chrisdickinson_ is now known as chrisdickinson
Encrypt has joined #ipfs
jedahan has joined #ipfs
cemerick has quit [Ping timeout: 256 seconds]
wiedi has quit [Quit: ^Z]
<jbenet>
ianopolous: yes that makes sense, can you open an issue on go-ipfs ? (it should become stdin in the command, which will allow it to be passed in req body
<patagonicus>
I should really change my VM setup. Having one machine connect to the other by connecting to a port on the host that is forwarded to the second VM is putting a heavy load on one of my CPUs without providing any benefits. :D
<dignifiedquire>
daviddias: tests passing and compiler happy!!!
<daviddias>
wooo o/ ! :D
cemerick has joined #ipfs
f[x] has quit [Ping timeout: 256 seconds]
compleatang has joined #ipfs
compleatang has quit [Remote host closed the connection]
<daviddias>
thanks for reviewing it, jbenet, will add that :)
<harlantwood>
improvements welcome
<jbenet>
daviddias: it's really good, narration helps a ton!
<daviddias>
:D :D
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #2069: update to new flatfs code (dev0.4.0...vendor/update-flatfs) http://git.io/v0E86
<achin>
whyrusleeping: im testing 0.4.0 -- do you want github issues opened for problems? or just a note in IRC?
<ipfsbot>
[go-ipfs] chriscool force-pushed fix-t-fatal-goroutine from 1bc2ebd to 57c4188: http://git.io/v0G6P
<ipfsbot>
go-ipfs/fix-t-fatal-goroutine f905197 Christian Couder: exchange/bitswap/bitswap_test: fix t.Fatal in a goroutine...
<ipfsbot>
go-ipfs/fix-t-fatal-goroutine f144822 Christian Couder: fuse/readonly/ipfs_test: fix t.Fatal in a goroutine...
<ipfsbot>
go-ipfs/fix-t-fatal-goroutine f41801c Christian Couder: merkledag/merkledag_test: fix t.Fatal in a goroutine...
chriscool has quit [Quit: Leaving.]
chriscool has joined #ipfs
<ipfsbot>
[ipfs] RichardLitt opened pull request #137: Updated ci-status to project-repos (master...feature/update-name) http://git.io/v0EKP
harlantwood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
Dashnakist has quit [Ping timeout: 240 seconds]
The_8472 has quit [Ping timeout: 264 seconds]
amstocker has quit [Ping timeout: 240 seconds]
keorn has quit [Remote host closed the connection]
The_8472 has joined #ipfs
rendar has quit [Ping timeout: 240 seconds]
go1111111 has quit [Ping timeout: 240 seconds]
f[x] has joined #ipfs
wiedi has joined #ipfs
<achin>
whyrusleeping: memory usage so far looks much much better \o/
hellertime has quit [Ping timeout: 240 seconds]
rendar has joined #ipfs
chriscool has quit [Read error: Connection reset by peer]
chriscool has joined #ipfs
chriscool1 has joined #ipfs
harlan_ has joined #ipfs
chriscool has quit [Read error: No route to host]
<The_8472>
wait a minute... is the DHT protocol udp or tcp based?
<achin>
tcp in 0.3.8. optional UDP support in 0.4.0
<The_8472>
oh god, no wonder it's frying my nat
<The_8472>
it's kademlia, yes?
<The_8472>
kademlia with TCP is utter madness
<whyrusleeping>
achin: awesome!!
<achin>
its kademila plus coral, i think
<whyrusleeping>
no coral yet, but thats on the todo liust
NeoTeo has quit [Quit: ZZZzzz…]
<achin>
plus maybe some special interplanetary magic sprinkled in
<The_8472>
kademlia is design for udp
<The_8472>
*designed
<whyrusleeping>
eh, kademlia is a routing protocol.
<The_8472>
yes, but for rapid iterative lookups
<achin>
whyrusleeping: inital data suggests that ipfs-add is faster, too, but i don't have good numbers at the moment. i think you did some work here, so this is probably expected
<The_8472>
chord does forwarding iirc
<whyrusleeping>
achin: yep, add should be a good deal faster
<achin>
(i'm still running ipfs-add with no daemon running)
cemerick has quit [Ping timeout: 272 seconds]
<ipfsbot>
[go-ipfs] chriscool force-pushed ipfs-volume from 781506c to ae53f1b: http://git.io/v0cE1
<ipfsbot>
go-ipfs/ipfs-volume 45959fa Christian Kniep: change Dockerfile order to create VOLUME under user ipfs...
<ipfsbot>
go-ipfs/ipfs-volume 307db9b Christian Couder: Travis-CI: add docker service...
<ipfsbot>
go-ipfs/ipfs-volume d89b3ad Christian Couder: CircleCI: add docker service...
harlantwood has joined #ipfs
<ipfsbot>
[js-ipfs-api] greenkeeperio-bot opened pull request #166: Update babel-runtime to version 6.3.19
harlantwood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<whyrusleeping>
lgierth: do the 0.4.0 bootstrap nodes have a utp swarm address?
<lgierth>
whyrusleeping: not yet
<lgierth>
i'll get to it right after the stuff for mappum
strepitous has joined #ipfs
jedahan has joined #ipfs
keorn has joined #ipfs
rombou has joined #ipfs
The_8472 has quit [Ping timeout: 240 seconds]
The_8472` has joined #ipfs
<whyrusleeping>
<3
rombou has quit [Ping timeout: 240 seconds]
mildred has quit [Ping timeout: 272 seconds]
* The_8472`
just got disconnected from irc due to router 'sploding ~~
<achin>
when i ipfs-add, there is a running filesize counter in the bottom left of my screen. is this counting input data processed? or output data (blocks) written?
rombou has joined #ipfs
go1111111 has joined #ipfs
baselab has quit [Quit: Leaving]
bengl has quit [Ping timeout: 260 seconds]
jedahan has quit [Read error: Connection reset by peer]
keorn has quit [Remote host closed the connection]
ilyaigpetrov has quit [Quit: Connection closed for inactivity]
Guest7890 has quit [Quit: Leaving]
keorn has joined #ipfs
pfraze has quit [Remote host closed the connection]
amstocker has joined #ipfs
pfraze has joined #ipfs
bengl has joined #ipfs
pfraze has quit [Ping timeout: 256 seconds]
borgtu has quit [Remote host closed the connection]
tinybike has quit [Ping timeout: 256 seconds]
jedahan has joined #ipfs
Encrypt has quit [Quit: Sleeping time!]
devbug has joined #ipfs
<ion>
achin: Try copying a gigabyte from /dev/zero and adding that.