<pinbot>
now pinning /ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN
<pinbot>
[host 0] failed to grab refs for /ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN: Post http://[fc98:424c:b433:d7e2:7ee3:9541:73ff:2cdb]:5001/api/v0/refs?arg=/ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN&encoding=json&stream-channels=true&r=true&: trailer header without chunked transfer encoding
<pinbot>
[host 2] failed to grab refs for /ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN: Post http://[fcfe:eab4:e49c:940f:8b29:35a4:8ea8:b01a]:5001/api/v0/refs?arg=/ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN&encoding=json&stream-channels=true&r=true&: trailer header without chunked transfer encoding
<pinbot>
[host 1] failed to grab refs for /ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN: Post http://[fce3:c53b:c3c5:2f54:8bb0:b6d9:898e:f140]:5001/api/v0/refs?arg=/ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN&encoding=json&stream-channels=true&r=true&: trailer header without chunked transfer encoding
<pinbot>
[host 4] failed to grab refs for /ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN: Post http://[fc3d:9a4e:3c96:2fd2:1afa:18fe:8dd2:b602]:5001/api/v0/refs?arg=/ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN&stream-channels=true&r=true&encoding=json&: trailer header without chunked transfer encoding
<pinbot>
[host 5] failed to grab refs for /ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN: Post http://[fcd8:a4e5:3af7:557e:72e5:f9d1:a599:e329]:5001/api/v0/refs?arg=/ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN&encoding=json&stream-channels=true&r=true&: trailer header without chunked transfer encoding
<pinbot>
[host 7] failed to grab refs for /ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN: Post http://[fcdf:a296:afe3:7118:4135:cc0b:ff92:4585]:5001/api/v0/refs?arg=/ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN&encoding=json&stream-channels=true&r=true&: trailer header without chunked transfer encoding
<pinbot>
[host 3] failed to grab refs for /ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN: Post http://[fc4e:5427:3cd0:cc4c:4770:25bb:a682:d06c]:5001/api/v0/refs?arg=/ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN&encoding=json&stream-channels=true&r=true&: trailer header without chunked transfer encoding
<pinbot>
[host 6] failed to grab refs for /ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN: Post http://[fc29:9fda:3b73:c1d2:9302:31e3:964c:144c]:5001/api/v0/refs?arg=/ipfs/QmZULkCELmmk5XNfCgTnCyFgAVxBRBXyDHGGMVoLFLiXEN&encoding=json&stream-channels=true&r=true&: trailer header without chunked transfer encoding
<lgierth>
...
<lgierth>
whyrusleeping: ^ it's go1.5.1 now :(
<locusf>
:/
noob__ has joined #ipfs
<noob__>
can i get some help
<noob__>
can get init to run on mac
acidhax has joined #ipfs
fingertoe has joined #ipfs
<acidhax>
Hey everyone, I'm curious if ipfs has been implemented in C++
<noob__>
anyone?
charley has joined #ipfs
nicolagreco has quit [Quit: nicolagreco]
<sonatagreen>
Err: AttributeError: 'NoneType' object has no attribute 'group' in UiServer.py line 75 > UiRequest.py line 80 > UiRequestPlugin.py line 21 > UiRequest.py line 308 > UiRequestPlugin.py line 35
<sonatagreen>
errr, ignore that
<sonatagreen>
wrong channel
<sonatagreen>
sorry
<noob__>
TOOL COMMANDS config Manage configuration version Show ipfs version information update Download and apply go-ipfs updates commands List all available commands Use 'ipfs <command> --help' to learn more about each command. Mac-c42c0334a93c:ipfs pookielax31$ init -bash: init: command not found
<ion>
noob: ipfs init
<noob__>
did it and it doesnt work
<ion>
acidhax: There is no C++ implementation at the moment.
<noob__>
lol fuck me
<noob__>
thanks
FredEE has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<noob__>
i had tried it earlier
inconshreveable has quit [Remote host closed the connection]
nicolagreco has joined #ipfs
Guest73396 has quit []
noob__ has quit [Quit: Page closed]
<acidhax>
ion thank you
lazyUL has joined #ipfs
rehelmin has joined #ipfs
lazyUL has quit [Client Quit]
Guest73396 has joined #ipfs
<cryptix>
whyrusleeping: i think i fixed putHandler
<cryptix>
cc jbenet
pfraze has quit [Remote host closed the connection]
charley has quit [Remote host closed the connection]
inconshreveable has joined #ipfs
domanic has quit [Ping timeout: 272 seconds]
charley has joined #ipfs
tilgovi has quit [Read error: Connection reset by peer]
Guest73396 has quit [Ping timeout: 272 seconds]
inconshreveable has quit [Remote host closed the connection]
inconshreveable has joined #ipfs
fingertoe has quit [Ping timeout: 246 seconds]
rehelmin has quit [Quit: Leaving.]
_jkh_ has quit [Read error: Connection reset by peer]
_jkh_ has joined #ipfs
_jkh_ has quit [Changing host]
_jkh_ has joined #ipfs
rehelmin has joined #ipfs
dlight has quit [Ping timeout: 252 seconds]
charley has quit [Remote host closed the connection]
fingertoe has joined #ipfs
dlight has joined #ipfs
jacobpaul has joined #ipfs
charley has joined #ipfs
jacobpaul has quit [Remote host closed the connection]
charley has quit [Remote host closed the connection]
jacobpaul has joined #ipfs
dlight has quit [Quit: Leaving]
pfraze has joined #ipfs
bok is now known as poka
pfraze has quit [Remote host closed the connection]
Qwertie has quit [Ping timeout: 264 seconds]
pfraze has joined #ipfs
inconshreveable has quit [Remote host closed the connection]
<pinbot>
now pinning /ipfs/QmPkPwNx4rr9X9oma5BsMev6YT2GF9u6y1gDtLxG7NyPJq
<noob__>
Mac-c42c0334a93c:ipfs pookielax31$ ipfs daemon Initializing daemon... Swarm listening on /ip4/108.48.85.76/tcp/4001 Swarm listening on /ip4/127.0.0.1/tcp/4001 Swarm listening on /ip4/192.168.1.4/tcp/4001 Swarm listening on /ip6/::1/tcp/4001 API server listening on /ip4/127.0.0.1/tcp/5001 Gateway (readonly) server listening on /ip4/127.0.0.1/tcp/8080 Daemon is ready
<noob__>
im stuck here and cant do anything hel papreciated
<whyrusleeping>
noob__: please use pastbin or gists or something
<whyrusleeping>
whats wrong?
<noob__>
just hanging at daemon ready
<whyrusleeping>
so the daemon is ready
<pinbot>
[host 6] failed to pin /ipfs/QmPkPwNx4rr9X9oma5BsMev6YT2GF9u6y1gDtLxG7NyPJq: 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"
<noob__>
it wont let me do shit after
<whyrusleeping>
yeah, the dameon is running.
<whyrusleeping>
open a new terminal and do things
acidhax has quit [Remote host closed the connection]
<pinbot>
[host 0] failed to pin /ipfs/QmPkPwNx4rr9X9oma5BsMev6YT2GF9u6y1gDtLxG7NyPJq: 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"
<whyrusleeping>
pinbot: you really need to get your shit together
<lgierth>
sorry friends
ygrek_ has quit [Ping timeout: 250 seconds]
<lgierth>
i still haven't figured out nginx' timeouts
<achin>
noob__: in a new terminal, what does "ipfs cat QmT78zSuBmuS4z925WZfrqQ1qHaJ56DQaTfyMUF7F8ff5o" return?
<ion>
codehero: If you run systemctl --user kill -s QUIT ipfs, it will dump the stacks in the log and die. That’s what whyrusleeping is asking us to do when it gets stuck and needs a restart.
martinkl_ has quit [Read error: Connection reset by peer]
martinkl_ has joined #ipfs
chriscool has quit [Ping timeout: 264 seconds]
Guest73396 has quit []
Guest73396 has joined #ipfs
chriscool has joined #ipfs
dignifiedquire has joined #ipfs
<cryptix>
gmorning
<akkad>
morning
<dignifiedquire>
good morning
martinkl_ has quit [Ping timeout: 265 seconds]
martinkl_ has joined #ipfs
martinkl_ has quit [Max SendQ exceeded]
rendar has joined #ipfs
dignifiedquire_ has joined #ipfs
inconshreveable has quit [Remote host closed the connection]
chriscool has quit [Ping timeout: 264 seconds]
s_kunk has joined #ipfs
s_kunk has joined #ipfs
chriscool has joined #ipfs
martinkl_ has joined #ipfs
<akkad>
so if you push something into ipfs, is it only findable with the full hash? e.g. can someone say "show me all the files submitted by this node"
dignifiedquire_ has quit [Quit: dignifiedquire_]
<harlan_>
akkad: full hash for now... searching to come later, likely on top of IPFS, not built into its core.
trn has quit [Read error: Connection reset by peer]
konubinix has joined #ipfs
pod has quit [Ping timeout: 265 seconds]
felixn has quit [Ping timeout: 265 seconds]
spikebike has quit [Ping timeout: 265 seconds]
<pyron>
Question: is there an intent to possibly change the crypto on the software to newer curve/xsalsa20/poly1305 crpyto or NTRU quantum resistant cryptography?
<ploopkazoo>
does "Install Go 1.4" mean 1.4+ or do I have to downgrade?
<ploopkazoo>
or is it a python2/3 situation and I should have both?
jacobpaul has joined #ipfs
bret_ has joined #ipfs
Nitori_ has joined #ipfs
pinbot has quit [*.net *.split]
therealplato has quit [*.net *.split]
krl has quit [*.net *.split]
ike_ has quit [*.net *.split]
Nitori has quit [*.net *.split]
wking has quit [*.net *.split]
mafintosh has quit [*.net *.split]
danslo has quit [*.net *.split]
henriquev has quit [*.net *.split]
ansuz has quit [*.net *.split]
rollick has quit [*.net *.split]
M-mistake has quit [*.net *.split]
bret has quit [*.net *.split]
mappum has quit [*.net *.split]
daviddias has quit [*.net *.split]
M-oddvar has quit [*.net *.split]
RX14 has quit [*.net *.split]
kevin`` has quit [*.net *.split]
ekroon has quit [*.net *.split]
rossjones has quit [*.net *.split]
drathir has quit [*.net *.split]
cmeiklejohn has quit [*.net *.split]
kumavis has quit [*.net *.split]
cryptix has quit [*.net *.split]
true_droid has quit [*.net *.split]
mikee has quit [*.net *.split]
akaihola has quit [*.net *.split]
kitcambridge has quit [*.net *.split]
kitcambridge_ is now known as kitcambridge
drathir87 is now known as drathir
ed- is now known as ed
true_droid_ is now known as true_droid
pinbot has joined #ipfs
cryptix has joined #ipfs
kevin``_ is now known as kevin``
chriscool has quit [Client Quit]
ed is now known as Guest98324
<cryptix>
huh.. netsplit?
<ploopkazoo>
(*.net *.split)
rossjones has joined #ipfs
rollick has joined #ipfs
<M-davidar>
ploopkazoo: last i heard go-ipfs was broken on go1.5, so go1.4 would probably be safest
chriscool has joined #ipfs
mafintosh_ is now known as mafintosh
<cryptix>
M-davidar: you need 1.5 now
jatb has quit [Ping timeout: 244 seconds]
<M-davidar>
haha, I should really stop trying to answer these questions :p
jacobpaul has quit [Ping timeout: 246 seconds]
go111111111 has joined #ipfs
henriquev_ is now known as henriquev
chriscool has quit [Read error: No route to host]
chriscool has joined #ipfs
bret_ is now known as bret
chriscool has quit [Client Quit]
jatb has joined #ipfs
<pyron>
Strange, can't reach security@ipfs.io or dev@ipfs.io via gmail.
go1111111 has quit [Ping timeout: 256 seconds]
<M-davidar>
also goes to show how old my ipfs install is :/
M-mistake1 has joined #ipfs
chriscool has joined #ipfs
krl has joined #ipfs
kumavis_ is now known as kumavis
cmeiklejohn has joined #ipfs
<M-davidar>
which isn't that old really
<cryptix>
hehe
mappum has joined #ipfs
ekroon has joined #ipfs
ike_ has joined #ipfs
<kpcyrd>
go-ipfs works for me on 1.5.1
<ploopkazoo>
when I load something on https://ipfs.io/ipfs/ it persists even after I shut down my node. I assume that's someone on the network caching it?
jatb has quit [Ping timeout: 244 seconds]
lgierth has quit [Quit: WeeChat 1.0.1]
kpcyrd has quit [Read error: Connection reset by peer]
<cryptix>
ploopkazoo: https://ipfs.io is a bunch of public nodes
<cryptix>
ploopkazoo: you can install browser plugins so that these links get redirect to local nodes
<cryptix>
but if you access through https://ipfs.io, the pub nodes will fetch your content and persist it for some time
<gamemanj>
So if your content was popular, it would remain around?
<cryptix>
gamemanj: yup, thats the idea :)
<gamemanj>
Sounds like a useful feature.
<ploopkazoo>
can you add darknet nodes like freenet or is it all automatic like bittorrent?
jatb has joined #ipfs
<cryptix>
ploopkazoo: currently its swarming automatic and not anonymous
<cryptix>
but there are ideas to support tor and i2p but it needs more work
<ploopkazoo>
yeah, #freenet tells me the plan for anonymity on ipfs is "just connect to a public node over tor if you really need anonymity for some reason"
chriscool has quit [Ping timeout: 264 seconds]
<gamemanj>
And I suppose uploading would be handled by public uploading nodes?
<ploopkazoo>
it would be nice if it was designed for anonymity like freenet
<ploopkazoo>
not sure how practical that would be in practice though
<doublec>
ploopkazoo: the main issue with a tor ipfs node is you'd need some gateway that connects to both tor and clearnet so requests for data on tor nodes could be reached.
<doublec>
ploopkazoo: since ipfs add doesn't actually push data out
<ploopkazoo>
well tor can exit to clearnet so ipfs.io should be reachable, but a hidden service connected to both would be better, yes
<cryptix>
doublec: you advertise your local content on node startup
<cryptix>
so you want another flavor of ipfs node that acts a bit differently on the dht as well
<doublec>
ploopkazoo: I do like that ipfs lets you pin data vs freenet losing data. I have data in ipfs I pushed months ago and it's still around.
<mungojelly>
what happens on ipfs all depends on what clients pin, if we write programs with pinning behaviors that'll completely determine what's available
<pyron>
Any plans to use faster, more modern cryptography suites (other than RSA) like Curve25519(ECC 3072-bit RSA equiv)/XSALSA20 Stream Cipher/Poly1305 Message Authentication? Or are the newer suites not implemented in go?
<ploopkazoo>
pinning is nice, but freenet's ability to store data without anyone but the person who inserted it being able to access it is nice too
<mungojelly>
like i imagined a video game that instead of downloading all of the content at once, it downloads each block of the terrain as you're walking into it, and pins them
<mungojelly>
ploopkazoo: encrypt it?
<pyron>
NTRU is quantum resistant and is faster than ECC
<cryptix>
pyron: ed25519 is nearly done iirc
<pyron>
Word
<ploopkazoo>
mungojelly: how many people would store a large chunk of encrypted data though? the fact that freenet doesn't know or care what you inserted is what's nice
jatb has quit [Ping timeout: 244 seconds]
<mungojelly>
ploopkazoo: does freenet reward people at all for storing random data? last i knew freenet was slow and lossy aka random about what data it stores
<pyron>
Also I noticed that blocks saved in .ipfs/ are unencrypted -- so if storing other people's data it is easily readable.
<mungojelly>
ipfs just doesn't come with any algorithm at all determining what nodes pin, but we can give it different behaviors depending on what pinning strategies we all choose
<pyron>
There are other projects like MaidSafe that implement a currency into their framework
<doublec>
pyron: yes, that's a catch. Hope no one stores their keepass database or similar without encrypting.
<pyron>
They also have data-deduplication for chunks
gritzko_ has quit [Ping timeout: 268 seconds]
<pyron>
The code is GPLv3 so any forward progress for any project is good
<pyron>
any free software project minimally*
<ploopkazoo>
mungojelly: there isn't an incentive, but people still do it. I guess that's also possible on ipfs, but it just feels different somehow. I haven't used ipfs for more than 5 minutes so I can't really say
<pyron>
I like ipfs without a currency framework.
<pyron>
It's just about raw information from my perspective. That is kind of like how the Internet is supposed to be
chriscool has joined #ipfs
<pyron>
No participation paywall to access data-- it's linkable, etc.
<gamemanj>
there already is a currency, it's just hidden and exists only to stop a node from leeching too much... it's called "bitswap" :)
<pyron>
Ah. Well there needs to be mechanisms to prevent spam/abuse
<mungojelly>
currency is an easy but sloppy way of deciding things to pin, we could also come up with smarter ways
jatb has joined #ipfs
<pyron>
What if somebody is trying to download large files?
<pyron>
Does their bitswap prevent them if they never used ipfs before?
<pyron>
Or does it operate on a different level of the network
<cryptix>
i think gamemanj meant that bitswap isnt very efficient right now
<pyron>
Ah
<gamemanj>
?
<cryptix>
sorry you didnt?
<cryptix>
i mean.. it really is slow right now
<gamemanj>
I was just noting that it seems like a bartering mechanism. I don't know the exact details, though.
<pyron>
One thing I noticed is that there was an immediate ip v 6 setting when doing ipfs init. That could throw off new users quickly.
<pyron>
Some people disable ipv6 to prevent MAC leaks
<pyron>
Sometimes in GNU/Linux the spoofing techniques don't work right.
<pyron>
Or certain routers etc can prevent spoofing.
ei-slackbot-ipfs has quit [Remote host closed the connection]
ei-slackbot-ipfs has joined #ipfs
<cryptix>
pyron: true - a node also advertise itself with all possible addresses
<pyron>
Ah
<mungojelly>
so is the plan for making an ipfs web just to have ipfs.io links for backwards compatibility that are replaced by localhost links by clients?
<pyron>
Well at least ipfs works as intended. I am grateful to see a have something workable so readily.
<pyron>
see a project*
<mungojelly>
what about making a pure ipfs web too that's just ipfs native links?
ashleyis has quit [Ping timeout: 272 seconds]
<pyron>
I think it would be neat to have some type of means of getting human readable links to content-- should be easy enough to create at some point
<pyron>
Perhaps a search engine that allows users to submit directories, etc.
<doublec>
mungojelly: pure ipfs would probably be preferred. The ipfs.io way has no cross origin domain security.
harlan_ has quit [Quit: Connection closed for inactivity]
<mungojelly>
well we just have to decide on a naming agent to trust, easy enough except the "trust" part
<pyron>
Well Direct Connect has an inbuilt DHT search engine
Zuardi has quit [Remote host closed the connection]
<pyron>
There are lots of algos available for distributed search technologies
Zuardi has joined #ipfs
<pyron>
and distributed databases
<pyron>
I think it might be neat if people hashtagged hashes of companies and entities in their tweets.
<pyron>
to hide who they are referring to
<mungojelly>
namecoin seems like such a nice idea but then when i looked at the actual processes for registering and renewing names it seems to have hardcoded a strategy that's at best mediocre :(
<doublec>
mungojelly: what don't you like about their strategy?
<mungojelly>
doublec: i forget the specific details but it just seemed sloppy and random and arbitrary to me
<rendar>
pyron: hmm, how it can hide them? the hash will map to the entity you "tag" ...
<pyron>
Not hide per se
<pyron>
Just obfuscate slightly
<rendar>
i see
<pyron>
I have another idea-- create a thing that allows people to create their own comment threads around particular URLs by hashing and subtracting http:// and https://
<pyron>
For live communication on a news article, etc. One could have a KiwiiIRC link to an IRC with a channel of the hash of the webURL
<mungojelly>
doublec: ok well i guess i just don't get the POINT of the strategy. you can sit on anything, anything. and then you have to renew every six months-- why?! how does that help anything.
<pyron>
Namecoin I felt kind of turned off to
<rendar>
pyron: i don't know, some URLs have variables into them, which will change hash, but they will point to the same news/article/whatever
<pyron>
Serves a purpose
<pyron>
I like things that encourage people to learn and develop skills like dn42 etc
edsilv has joined #ipfs
<rendar>
pyron: so hashing URLs per se, would be like hashing file names, you have an hash of something, but not of the *content* of that something
<pyron>
Those variables can be subtracted
<rendar>
pyron: hmm, and how you can know the static variables needed to index the file/article, and those not?
<pyron>
Things like # and ? etc
edsilv has quit [Client Quit]
<rendar>
pyron: e.g. my_newspaper.com/get_article?id=45&conect_from=facebook
<pyron>
Those can be subtracted and whatever comes after them too
<rendar>
pyron: ^ in that example you need 'id' to refer the correct article, but you don't need 'connect_from'
<cryptix>
pyron: like that idea as well - have been experimenting with httrack and youtube-dl
<mungojelly>
the easy thing to do is just to bring in the article into your own space, hash the article itself, but copyright so we're pretending computers work some unexplained impossible way
<rendar>
which will change
<pyron>
I can use sed to erase whatever comes after a ? or # or &
<rendar>
pyron: but you need id=45
<pyron>
That's true. Perhaps there would be a way to isolate how those things are typically handled
cemerick has quit [Ping timeout: 268 seconds]
<pyron>
I am sure Google has to deal with the fine details of URLs- there are ways to get it to work.
<rendar>
well
<rendar>
that would be impossibe, with the normal http and URI syntax
<pyron>
But I do appreciate the input
<rendar>
you won't know what variables you need and those you don't need
<cryptix>
you need some kind of (url,time) index like the wayback machine
<rendar>
cryptix: time?
<gamemanj>
you're not thinking thre-dimensionally! :)
<pyron>
There may be a way to create a new temporary identifier for each site encountered, store in distributed database then delete after so long
<cryptix>
if you link to the classic web - you just said the url params are important
<pyron>
Yes that is true
<mungojelly>
well, won't it be nice to move from that confusion to hashes! very predictable behavior
<pyron>
I didn't consider that originally
<mungojelly>
you can always say, do this with this hash-- this. is. what. will. happen.
kpcyrd has joined #ipfs
<rendar>
the point is
<rendar>
the URL *is already* an hash to some content
<mungojelly>
we're so used to the messy mutable web but i think we should just make an immutable hash web. links only go backwards and never change, that's all.
<cryptix>
with added content smarts, like reader view on blog/news sites
<cryptix>
mungojelly: yup - transform rss feeds into blockchains <3
<cryptix>
each post having all the content it needs as merkle links
<gamemanj>
advantages to a "backwards/never change" web: you can guarantee that what someone is talking about is what they were talking about at the time... disadvantages: if there are updates to it that are of use to the user, how do you know?
<mungojelly>
blockchains are a subset i guess of chains of hashes. any chain of hashes is a secure ordering, combined they can paint accurate histories.
<mungojelly>
gamemanj: you can ask in a different channel "hey, any updates i should know about this hash?!"
<pyron>
Problem with static web is that every change to a file makes to have to reupload and give new URL
<pyron>
After changes made
<pyron>
But is fine for archival / file distribution activities
<mungojelly>
you could also post just a patch and a hash of what to patch and make patch chains
<Stskeeps>
idle question, put up ipfs daemon on my hosting provider and soon after i got a abuse report about it scanning 10.* 169.254.* 172.* 192.168.* 4001 ports.. i presume it's fine to attempt to connect to the DHT in general, but is this expected behaviour when only interface on server is public-IP?
<pyron>
I noticed on second ipfs add, that the speed is much greater.
<pyron>
(doing the same directory)
<cryptix>
pyron: well - if you use the same chunking for the content, there wouldnt be much change ideally
<cryptix>
Stskeeps: are you with a german hoster starting with the latter H?
<pyron>
I just launched baobab right before you shared that
<pyron>
Running low on space.
<cryptix>
hahaha quantum entanglement? :P
<pyron>
I should suppose
<pyron>
zero dimensional phase space singularity
ilyaigpetrov has joined #ipfs
<ilyaigpetrov>
How do you plan to implement streaming media in ipfs? If ipns pints to some ipfs object then how would you go -- repoint ipns over time?
<cryptix>
ilyaigpetrov: talked about with ion about that last night
gamemanj has quit [Remote host closed the connection]
Guest73396_v has quit [Ping timeout: 240 seconds]
hellertime has quit [Quit: Leaving.]
jatb has joined #ipfs
hellertime has joined #ipfs
gamemanj has joined #ipfs
<ilyaigpetrov>
If I get it right this algo is about backward rewinding (backwards in history), but my question may be simpler -- how to get the next chunk of content if I have one already?
<achin>
i believe there has to be an OOB channel for that
<pyron>
Another issue I am encountering is that whenever I add something to ipfs, it duplicates the data in the form of blocks. So if I want to share 1 TB on IPFS I will need a 2TB drive.
<pyron>
Is there a way to do some type of metainformation db that eliminates that need like torrents?
<achin>
yeah, that is a little annoying. i think it's something that someone will eventually try to solve
<ilyaigpetrov>
achin: what OOB stands for? out of browser?
<achin>
out-of-band
<achin>
so you might need to get the hash of the latest content via something that's not ipfs
<achin>
(ipns might work in theory, but could be too slow)
<mungojelly>
hashes inherently point backwards in time
elasticdog has quit [Ping timeout: 244 seconds]
<Stskeeps>
so anyway, i have a ipfs daemon behind NAT, but a forwarded 4001 port and ipfs notices the public ip as part of swarm, i ipfs add a file, check that ipfs cat works, and go to a remote server with a public ip, and ipfs cat with that same hash.. and it says Error: merkledag: not found .. did i go wrong somewhere? telnet'ing to the daemon's public ip port 4001 works fine
<ilyaigpetrov>
I think one could try ipfs pub/sub for streaming if it wouldn't be slow as well.
<ilyaigpetrov>
ion: yep, but yet I think it would be nice to have linked lists which you may rewind forward, I think there should be write-only ipns for this which is decentralized
<ion>
You can easily reverse a linked list locally.
Guest73396 has quit [Ping timeout: 250 seconds]
<ilyaigpetrov>
having one-way-directed lists means you have only one head which you can't go beyond if new data becomes available (E.g. I've paused live video streaming and then resumed it)
<cryptix>
/ipfs/QmTrRng9EWBWCLmpHVgKx3NDzHG8J8Y76Vuo21eYqKbyid < was recorded a couple of weeks ago by me and two friends
pyron has joined #ipfs
<whyrusleeping>
half a gig, lol
<cryptix>
5.5hrs ^^
<whyrusleeping>
damn dude
<whyrusleeping>
i was thinking like 'uncompressed wav @ 192khz' or something
<cryptix>
5.25, okay
<ilyaigpetrov>
how does bitcoin publishes the head of its ledger?
<whyrusleeping>
typing 'bitcoin head of' completes to 'head of bitcoin exchange found dead'
<whyrusleeping>
thats a bit too morbid for me this early
<cryptix>
hahaha
<whyrusleeping>
youre on your own ilyaigpetrov
<ilyaigpetrov>
whyrusleeping: maybe ipfs may use the same technique for publishing volatile things, so I guess you are interested too
<whyrusleeping>
we use ipns
<whyrusleeping>
publish records to the dht
<ilyaigpetrov>
it's slow
<whyrusleeping>
implementation details, it will be faster soon
<whyrusleeping>
especially once 1887 lands
<whyrusleeping>
(also, who says bitcoins not slow too ;) )
<whyrusleeping>
bitcoin uses a gossip protocol to distribute blocks in the chain throughout the network
<whyrusleeping>
(i think?)
<ilyaigpetrov>
I don't like ipns because it belongs to some node you have to trust to
<cryptix>
i heared they used irc for comms and thought "that cant be right"
amade has quit [Ping timeout: 256 seconds]
amade has joined #ipfs
nicolagreco has joined #ipfs
nicolagreco has quit [Remote host closed the connection]
Oatmeal has joined #ipfs
NeoTeo has quit [Quit: ZZZzzz…]
cemerick has joined #ipfs
Guest98324 is now known as ed__
<clever>
whyrusleeping: when a bitcoin client has verified a new block or pending transaction, it will send an invite command to every connected peer, with the type and hash of the object
<clever>
those peers are then free to request a copy if they lack it, and repeat the same proces
<clever>
so its pretty much just a broadcast storm of object hashes, and the occasional request for object you dont have yet
<whyrusleeping>
yeah, thats the gossip protocol if i understand it correctly
<ion>
ilyaigpetrov: Sorry, I didn't get what you meant wrt. pausing and then resuming.
<clever>
in the rare case where there are 2 different solutions for a block at a given depth, nodes will ignore the duplicate, causing the network to temporarily fork, based on how much heard about each block first
<clever>
and then whichever fork has more hashing power will add to their variant
<clever>
once one fork is longer, the other will drop the dup and switch over
<ion>
ilyaigpetrov: You just traverse the spine of the data structure from the IPNS pointer until you find the next chunk to proceed.
<clever>
so the network sort of relies on the spread of a block to take less then 20mins (the block interval)
NeoTeo has joined #ipfs
<victorbjelkholm>
is there a way a can see the filesize of pinning a hash before actually pinning it?
<ion>
ilyaigpetrov: With a linked list that's O(n), with the data structure in the link that's O(log n).
<achin>
ion: in theory, you don't need a linked list, if the head node contains links to every child block. maybe you want to add a brief note about why this is less desirable than your proposed structure
<achin>
(the primary one in my view is that the size of the head block grows much more slowly if you don't link to all past blocks)
<achin>
victorbjelkholm: ipfs object stat <hash>
Eudaimonstro has quit [Remote host closed the connection]
acidhax has joined #ipfs
<fiatjaf>
I have a bunch of somewhat big files (in a folder) in a machine with IPFS. I want to get them into other machine. what is the sane way to do it?
zz_r04r is now known as r04r
<fiatjaf>
I did: "ipfs add <folder>", "ipfs pin add <hash of folder>", ipfs daemon
<fiatjaf>
then in the other computer "ipfs daemon", "ipfs get <hash of folder>". it took a long time without any output, then the daemon crashed with a lot of errors.
acidhax has quit [Remote host closed the connection]
<ion>
achin: The overhead of the number of links might become too great at some point. A good compromise would be to retain the first n ancestors and then follow the algorithm for the rest.
<achin>
fiatjaf: that's the right thing to do. can you open a bug on github with the errors from the daemon?
<achin>
after doing that, you might be able to retry and have it work
<fiatjaf>
then I tried "ipfs ls <hash of the folder>" (in the second computer) and it shows a list of the files in the folder with hashes and names. does IPFS store filenames along with the hashes? that doesn't seem to make sense. how can I know which of these files were downloaded (assuming some were, before the error)?
cemerick has quit [Ping timeout: 240 seconds]
<achin>
if a hash is a directory, it will include links to other directoreis or files. these links will have names
<ion>
achin: That would mean bounded size for the recent parents and O(log n) size for the rest instead of O(n) for everything.
<achin>
fiatjaf: hashes themselves don't have names, as you correctly said
<achin>
to know what you've already downloaded, that's a little hard. "ipfs refs local" will show you all hashes you've downloaded, but mapping that back to your original directory structure will require quite a bit of postprocessing
<achin>
so i recommend just trying "ipfs get Qm..." again
<achin>
ion: right, i agree. to be complete, your webpage could mention those thoughts
gamemanj has joined #ipfs
<achin>
also, did you build that with elm?
<victorbjelkholm>
achin. thanks!
<achin>
victorbjelkholm: to be pedantic, it should probably be mentioned that the filesize reported by "object stat" could be wrong if the object itself mis-represents the size of its links
<victorbjelkholm>
achin, hmm, I see. I would have to repeat object stat on every links if I care about precision?
<achin>
yes, but in that point, you might have come pretty close to downloading everything :)
<achin>
also, the size reported by "object stat" is the size of the raw merkledag nodes. if you have a hash to a file, for example, you won't get the exact size of the file if you "ipfs object get"d it
<victorbjelkholm>
achin, hm, I see. So it would be a approx size
<victorbjelkholm>
good enough, as long as it's not completely wrong
<achin>
right. it should never be less than the filesize, always a little bit greater
<victorbjelkholm>
ah, that's good to know. Thanks for helping me out :)
<achin>
actually, on that last point.... that might not always be true. for example, if ipfs ever compressed its merklenodes, then the object size of the node might be smaller than the uncompressed file on the filesystem
<achin>
anyway, for right now, it's the best you've got, and in the current ipfs impl, it should always be "close enough"
<whyrusleeping>
victorbjelkholm: 'ipfs file ls' will give you actual file sizes
rehelmin has joined #ipfs
<achin>
does that only work on files, though?
<whyrusleeping>
not in the normal cli output, but if you pass --enc=json, or use the http api
<whyrusleeping>
yeah, its for unixfs files
<whyrusleeping>
the 'cumulative size' in object stat shows the total on disk size of everything below a node
<achin>
and is it smart enough to get the accurate filesize without downloading all the blocks?
<whyrusleeping>
yeah, we cache that info in the root file blocks
<whyrusleeping>
makes filesystems a little easier
<whyrusleeping>
lol
<achin>
oh yeah, via the 'filesize' attribute on the Data object?
<whyrusleeping>
yeap!
<achin>
in the "worst case", you might actually download the whole file if it wasn't chunked into smaller bits
<whyrusleeping>
right
<achin>
hmmmm
<achin>
hmMMM there is a javascript impl for protobufs
gritzko_ has joined #ipfs
<achin>
this might be useful for writing a small utility (distributed on IPFS!) to explore the PBNode, PBLink, and Unixfs strucurues
<whyrusleeping>
do it!
* achin
adds it to page 44251 of his TODO list
jatb has quit [Ping timeout: 244 seconds]
<achin>
right now (as far as i can see), to figure out if a PBNode's payload is a unixfs object, you have to try to decode it as such, and check for errors. has there been any thought to extending PBNode to include some hints about how to decode its payload?
<ion>
achin: Yeah, I used Elm.
__jkh__ has quit [Ping timeout: 244 seconds]
<achin>
ion: i've love to see the source. and hear about your experience with it
_jkh_ has joined #ipfs
jatb has joined #ipfs
Guest73396_h has joined #ipfs
<ion>
achin: I'm going to share it, I just didn't get around to adding the COPYING boilerplate yet.
<achin>
no rush (especially not now, i can only devote a few lines of my terminal to IRC)
Guest73396 has quit [Ping timeout: 240 seconds]
<ion>
achin: The experience was wayyy nicer than the JavaScript experience. But not as nice as Haskell, as Elm lacks some of its niceties.
<achin>
when you compare to haskell, are you comparing to having haskell generate javascript?
cemerick has joined #ipfs
<ion>
achin: There is ghcjs which compiles GHC Haskell into JavaScript but the size of the generated code is an order of magnitude greater than what Elm generates.
<achin>
for things that both elm and emscripten would be suitable for, i wonder how they compare (in terms of size of generated code)
<ion>
Elm’s built-in virtual-dom support is definitely nice.
sonatagreen has joined #ipfs
Encrypt has joined #ipfs
<fiatjaf>
achin: thank you, that was very clarifyfing.
<fiatjaf>
achin: I'll report the error if I can reproduce it.
Guest73396_h has quit [Ping timeout: 272 seconds]
compleatang has quit [Quit: Leaving.]
acidhax has joined #ipfs
lgierth has joined #ipfs
acidhax has quit [Remote host closed the connection]
acidhax has joined #ipfs
compleatang has joined #ipfs
acidhax has quit [Remote host closed the connection]
ygrek_ has joined #ipfs
<gamemanj>
I now have quite an incentive to try running IPFS i
<gamemanj>
On android.
<gamemanj>
...so getting a go compiler needs a go compiler. My plan is foiled...
<whyrusleeping>
gamemanj: lol, they have prebuilt
<whyrusleeping>
and you can cross compile
<whyrusleeping>
GOOS=android GOARCH=arm go build ./cmd/ipfs
<gamemanj>
Computer is currently memtesting
acidhax has joined #ipfs
<gamemanj>
But I could install go on a Pi and crosscompile here
<lgierth>
someone got it to work a couple of months ago
<lgierth>
gamemanj: raspi is sloooow
<gamemanj>
Exactly, so i'd use it to crosscompile ipfs to here
jamescarlyle has joined #ipfs
<gamemanj>
Where a relatively fast tablet is
acidhax_ has joined #ipfs
<gamemanj>
Using the Pi to get a binary usable from a terminal e!u
<gamemanj>
*emulator
<lgierth>
why not simply crosscompile on your desktop/laptop?
<gamemanj>
As i said, computer is memtesting - hence I wanted to do stuff on a tablet :)
<lgierth>
aaah
acidhax_ has quit [Remote host closed the connection]
<lgierth>
:)
<lgierth>
missed that line
<gamemanj>
But that won't stop me having IPFS-related fun
acidhax_ has joined #ipfs
<lgierth>
right on
_jkh_ has quit [Changing host]
_jkh_ has joined #ipfs
acidhax has quit [Ping timeout: 246 seconds]
mildred has quit [Ping timeout: 272 seconds]
<gamemanj>
Now I wish i hadn't just started updates on the pi...
sseagull has joined #ipfs
acidhax_ has quit [Remote host closed the connection]
acidhax has joined #ipfs
acidhax has quit [Remote host closed the connection]
acidhax has joined #ipfs
mikee_ is now known as mikee
<victorbjelkholm>
looking for some examples of using ansible and ipfs together, anything out there?
captain_morgan has quit [Ping timeout: 246 seconds]
<victorbjelkholm>
seems like solarnet is using ansible... looking for some other, smaller examples as well
mrrrgn has quit [Ping timeout: 240 seconds]
acidhax has quit [Remote host closed the connection]
acidhax has joined #ipfs
gritzko_ has quit [Ping timeout: 265 seconds]
mrrrgn has joined #ipfs
* gamemanj
does a go get -u, and waits
fingertoe has joined #ipfs
<gamemanj>
go get may want to give a bit more feedback
<sonatagreen>
ansible: link?
voxelot has joined #ipfs
voxelot has quit [Changing host]
voxelot has joined #ipfs
acidhax has quit [Remote host closed the connection]
acidhax has joined #ipfs
<jamescarlyle>
have done a minor update to the Interplanetary Wiki, built on IPFS
acidhax has quit [Remote host closed the connection]
mildred has joined #ipfs
acidhax has joined #ipfs
<whyrusleeping>
gamemanj:
<whyrusleeping>
go get is part of go, not ipfs
<gamemanj>
Yep, ik - just noting
harlan_ has joined #ipfs
Tv` has joined #ipfs
<gamemanj>
Ipfs-go is either quite big, or the pi is quite slow... Either makes sense
<ianopolous_>
gamemanj: I'll fix that api issue tonight..
<achin>
the first gen pis are super slow, (and only having 256mb memory doesn't help!)
gaboose has joined #ipfs
domsch has joined #ipfs
<victorbjelkholm>
gamemanj, use go get -v and you'll see the progress a bit better
<gamemanj>
Well, it's a bit late to do that, but I'll note it in future
<domsch>
hey everyone, I'm currently working on a revision of my Wikipedia project and want to integrate IPFS to store all of the page summaries (for now).
<domsch>
I was wondering if the most efficient to do this is to basically create a unique object for every page summary, and then store the name of the page + hash of object in a JSON which contains name:hash pair of all currently backed up pages
<domsch>
or any suggestions on how to set this up?
<gamemanj>
Aaaand the linker ran out of memory...
pfraze has joined #ipfs
<ipfsbot>
[go-ipfs] cryptix force-pushed fix/ThatPut from eb3d2aa to 5b47bb4: http://git.io/vW84K
<ipfsbot>
go-ipfs/fix/ThatPut 5b47bb4 Henry: putHandler: applied review comments from @whyrusleeping...
niekie has joined #ipfs
mildred has quit [Ping timeout: 240 seconds]
<achin>
hehe
<acidhax>
domsch: the problem with pages that can support changes is that ipfs is heavily driven by permanent state. :-/
<gamemanj>
Turns out gobuilder.me had an arm binary. It worked fine on android, somehow.
border0464 has joined #ipfs
warner has joined #ipfs
<codehero>
wait really?
<codehero>
hm. actually. my phone can do x86, so would the regular ipfs build work?
<codehero>
but then... it probably eats battery like crazy
Trogones has quit [Killed (Sigyn (Spam is off topic on freenode.))]
<RX14->
fancy meeting gamemanj here...
RX14- is now known as RX14
gritzko_ has joined #ipfs
pyrophorous has joined #ipfs
<gamemanj>
it might be the terminal emulator, it came with a toolchain - not entirely a stretch to assume that it was all normal-abi so more stuff would work...
<codehero>
heh. the stanford video is still downloading. really needs more seeders or should i say pinners? lol
<domsch>
@acidhax: I was actually thinking of running a script once a day that compares the objects to the current page, if there is a change, it will update the object and re-reference the hash to the name
martinkl_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<jamescarlyle>
domsch, i have an updatable wiki on ipfs
<gamemanj>
Looking at the go package "syscall"... Does go code even use glibc???
<gamemanj>
If not, that explains why the android abi wasn't a problem...
<domsch>
ahh this is interesting
<domsch>
and @jamescarlyle no I'm not, I just intend to collaborate with them on some projects (social tipping & identity)
<jamescarlyle>
that link i posted - my browser doesn't seem to load the /lib folder items, for some reason - would be interested to hear if it works for you if you are running the IPFS daemon locally
<domsch>
I'm running it locally right now, says it's missing angular-animate.min.js
trissypissy has joined #ipfs
lazyUL has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
rehelmin has quit [Quit: Leaving.]
<gamemanj>
The reason the ipfs arm binary just works on android seems to be: go stuff is statically linked. No dependence on Android ABI libs, no ABI problem...
Matoro has joined #ipfs
<domsch>
alright seems to be working now
<Stskeeps>
might for >5.0 need the -pie linking flags
<Stskeeps>
but again, does that matter for static binaries. hmm
<gamemanj>
I'm on Android 4.4.3, not a matter
Matoro has quit [Ping timeout: 272 seconds]
se3000 has quit [Quit: se3000]
<codehero>
there we go. pinned the stanford video
<jamescarlyle>
@domsch, if you are running locally, you need to set cors
adsadwga has joined #ipfs
adsadwga has quit [Client Quit]
jamescar_ has joined #ipfs
jacobpaul has joined #ipfs
chriscool has quit [Read error: No route to host]
Encrypt has quit [Quit: Quitte]
Encrypt has joined #ipfs
Encrypt has quit [Read error: Connection reset by peer]
<sonatagreen>
how naughty is this: /ipfs/QmPBLZpchvTRvPbophp2yEtAh6rDNAScnYK9vwxgE4kjmA [sfw]
gamemanj_ has joined #ipfs
gamemanj has quit [Quit: A IRC Client Of Some Form. Not Sure Though - Might Be Turkey Sandwich.]
gamemanj_ is now known as gamemanj
<sonatagreen>
(you may need to do `ipfs object get QmPBLZpchvTRvPbophp2yEtAh6rDNAScnYK9vwxgE4kjmA`)
cemerick has quit [Ping timeout: 240 seconds]
inconshreveable has joined #ipfs
<codehero>
huh? ipfs object get?
<codehero>
what does that do?
<sonatagreen>
shows the structure of the object, pretty-printed but not interpreted
Encrypt has quit [Read error: Connection reset by peer]
<ion>
Shows a JSON (by default) representation of an IPFS object.
Encrypt has joined #ipfs
<M-matthew>
jbenet: if you're out there... fwiw we're finally implementing anonymous access on matrix at last (for all your decentralised messaging scratchpad needs)
Encrypt has quit [Read error: Connection reset by peer]
* M-matthew
spots stskeeps and gets flashbacks to a past life playing with the unrealircd codebase :)
lazyUL has joined #ipfs
Encrypt has joined #ipfs
border0464 has quit [Ping timeout: 256 seconds]
border0464 has joined #ipfs
<gamemanj>
"anonymous access on matrix"? Lot of branches and repos, none of them are called matrix, haven't seen anything called matrix in commands...
Encrypt has quit [Read error: Connection reset by peer]
Encrypt has joined #ipfs
zeroish has joined #ipfs
vanila has joined #ipfs
<vanila>
hello
<vanila>
is there an ipfs based youtube replacementL?
<codehero>
i don't think there is, but it shouldn't be too hard to make one
<Stskeeps>
M-matthew: hehe ;) still on to having people communicate with eachother
rehelmin has quit [Quit: Leaving.]
<codehero>
sonatagreen: ah, okay. cool
* gamemanj
thinks about how a youtube replacement in IPFS would work without a way to recommend videos
<vanila>
lets do it!
<gamemanj>
Hmm, let me think
<sonatagreen>
I really feel like bare IPFS isn't enough for a dynamic web
<sonatagreen>
IPFS is the CDN
<gamemanj>
There is a problem with it
<gamemanj>
The pages can't do uploads or IPNS sets
<codehero>
that's why people use ethereum + ipfs
<gamemanj>
If a page was given the ability to do an IPNS set if it gave the private key and public key
<gamemanj>
that would do everything
<gamemanj>
well, that and uploads ofc
<gamemanj>
:)
<gamemanj>
thing is, there's a really simple solution to people being able to publically add content to an index... just post a keypair only used for that IPNS index
<gamemanj>
but there has to be a way to use that keypair
<gamemanj>
else there's no point
<sonatagreen>
so the API has to accept keypair data as an argument?
<sonatagreen>
although
<gamemanj>
Yep.
<sonatagreen>
allowing that has potential privacy issues i think
<gamemanj>
Public indexes are the only reason it should ever exist or be used.
trissypissy has quit [Quit: Konversation terminated!]
<gamemanj>
But that's a pretty important thing for a web.
<sonatagreen>
if an ipfs-loaded site can send data without explicit authorization
<gamemanj>
Then ask the user for authorization.
<sonatagreen>
i mean, i guess it kind of can already, if you allow cross-origin requests?
<sonatagreen>
but
rehelmin has joined #ipfs
<sonatagreen>
i'd like to have a web that doesn't spy on me
<gamemanj>
Like, have a page which an IPFS site can redirect you to, where it asks you if you want it to do this
<codehero>
what about people just messing up the ipns?
s_kunk has quit [Ping timeout: 240 seconds]
<gamemanj>
codehero: That's the nature of public lists, really... but there are mitigation methods
<gamemanj>
Have each entry link to the previous one, and if the tree is broken, the clients remember what it's supposed to be like and fix it
<codehero>
that would help, i suppose
<gamemanj>
kind of like a blockchain on IPNS, really
wopi has quit [Read error: Connection reset by peer]
wopi has joined #ipfs
<sonatagreen>
this feels like reinventing the wheel
<codehero>
:D
<codehero>
blockchainception
<gamemanj>
Isn't that the point? Except this time it's a bicycle wheel, which is better by design
<sonatagreen>
mmh
charley has joined #ipfs
<sonatagreen>
look at how much work namecoin has had to do to compete with ugly hacks for storing data in the bitcoin chain
<M-matthew>
gamemanj: matrix isn't an IPFS project (although we're planning to use IPFS for our distributed content repository needs) - it's a decentralised comms standard; kinda what you'd get if git & IRC had babies. matrix.org fwiw.
<sonatagreen>
using something that people have already bought into is a *massive* advantage
<M-matthew>
stskeeps: the more things change the more things stay the same :D
<kyledrake>
Someone mentioned security@ipfs.io may not be working
<kyledrake>
Last nighy
<kyledrake>
Night
<cryptix>
hrm..
Not_ has joined #ipfs
<kyledrake>
pyron do you need to talk with someone about a security issue?
<gamemanj>
sonatagreen: IPFS will only be a viable web replacement if there's a way to do things the centralized people do, preferably in as decentralized a manner as possible. That means public indexes :)
jamescarlyle has joined #ipfs
<sonatagreen>
hmm.
<gamemanj>
Without public indexes, finding content becomes harder, and without the ability to find content, no matter how much content you have, it's essentially worthless.
jamescar_ has quit [Ping timeout: 268 seconds]
<Stskeeps>
M-matthew: indeed
<fingertoe>
Couldn't each user Identity maintain a tree of their comments and somehow "ping" the indexer of the website those comments are relevant to?
<gamemanj>
The indexer... would that be a centralized peer?
gritzko_ has quit [Ping timeout: 268 seconds]
rehelmin has quit [Quit: Leaving.]
<fingertoe>
ah, I see...
charley has quit [Remote host closed the connection]
<cryptix>
gamemanj: not sure those indexes need to be central
<fingertoe>
Would have to be some kind of blockchainish solution.
gritzko_ has joined #ipfs
<gamemanj>
cryptix: Well, how do you "ping" the indexer?
rendar has quit [Quit: std::lower_bound + std::less_equal *works* with a vector without duplicates!]
<JasonWoof>
It's javascript, produces svg
<JasonWoof>
I'm trying to figure out if/how I want to save the data produced
<JasonWoof>
it needs a similar backend to a pastebin
<JasonWoof>
would be cool if people could get: 1. A url to the mockup editor, with their mockup loaded. and 2. a url that shows their mockup without the editor
<JasonWoof>
If I set this thing up to be a ipfs page/app, can it (from javascript) save these values (the encoded mockup) into ipfs?
<JasonWoof>
does it need to be a "trusted app" to do that?
<cryptix>
gamemanj: JasonWoof /ipns/crayons.jsonwoof.com/ < latest version of the editor /ipns/crayons.jsonwoof.com/?edit=$ipfsHash and /ipns/crayons.jsonwoof.com/?show=$ipfsHash
<cryptix>
- gamemanj , sorry
<gamemanj>
cryptix: I see how that's relevant :)
<JasonWoof>
cryptix: exactly. my question: can the javascript on that page inject data into ipfs, so that $ipfsHash points to a real thing?
<cryptix>
JasonWoof: right now the public gateways arn't writable - i wonder where we would coordinate this discussion
<cryptix>
i pushed a PR tomorrow which fixes 'ipfs daemon --writable'
<cryptix>
which allows for PUT uploads to :8080
<gamemanj>
cryptix: Timetravelling? You mean yesterday, which is somewhat funny...
<gamemanj>
Or do you mean today?
<cryptix>
if i understood jbenet correctly, those will also be enabled on the public gateways but I imagine in a reduced/limited fashion
<cryptix>
lol
<cryptix>
s/tomorrow/this morning/
<multivac>
cryptix meant to say: i pushed a PR this morning which fixes 'ipfs daemon --writable'
Matoro has joined #ipfs
<gamemanj>
JasonWoof: In the meantime, you could always run an IPFS node, and add a serverside script to your website to 1. Check that the data was valid, then 2. Upload it to IPFS and give back the hash
<ion>
cryptix: Would it be useful to also support POST in case the uploader doesn’t know the hash in advance?
<JasonWoof>
so it's user-configurable whether javascript can create ipfs files?
nicolagreco has joined #ipfs
inconshreveable has quit [Remote host closed the connection]
<cryptix>
ion: you dont have to know it in advance
<ion>
cryptix: How does PUT work then?
<JasonWoof>
gamemanj: I'm thinking of doing that, though this doesn't have much advantage over stashing stuff on my server in MySQL
<gamemanj>
JasonWoof: Well, you can make it so you read it all from ipfs.io >:)
<cryptix>
you can PUT to /ipfs/QmNullObj to just create the file or /ipfs/QmRoothash/update/or/create/on/path
rendar has joined #ipfs
<cryptix>
i think mildred came up with it - i just fixed it to work again
<gamemanj>
JasonWoof: That way it would, presumably, cache it, so you don't use the bandwidth
<cryptix>
ion: you might be right - not that restful, i guess
<JasonWoof>
is it hard for js to figure out what the filename will be?
<JasonWoof>
sorry if my terminology is off, I mean /ipfs/THIS_PART
<gamemanj>
JasonWoof: Maybe, IDK. But you can just have the server return the hash.
<cryptix>
JasonWoof: you dont bother - the ipfs node will tell you the hash after the upload
<JasonWoof>
cryptix: cool, that's easy
<ion>
cryptix: Oh, I see. Using PUT to “mutate” an existing object (receiving a 301 Moved Permanently response pointing to the mutated object) *is* a nice idea. But “POST /ipfs” seems more standard than “PUT /ipfs/QmNullObj” for just creating a new object.
<JasonWoof>
probably best to get confirmation that it's saved off-browser before giving the user a url to it
<nicolagreco>
do http methods work / is there an adapter?
<sonatagreen>
jamescarlyle, the wiki doesn't seem to be working for me? is there like a test page i can try, or a series of troubleshooting steps
<JasonWoof>
gamemanj: you meant when people make stuff, hand out urls like http://ipfs.io/ipfs/<hash of your rendered mockup> ?
<jamescarlyle>
@sonatagreen, just on a call, will be off in 15 mins
<sonatagreen>
thanks
<ion>
cryptix: Given the idea of allowing you to mutate objects with PUT, it would also be appropriate to be able to DELETE files.
<cryptix>
ion: agreed - there might be a postHandler already, actually didnt check
<cryptix>
ion: i saw the old deleteHandler and also already saw the new api to do it properly
grahamperrin has joined #ipfs
<JasonWoof>
is it expected that --writable will be normal/default at some point?
<gamemanj>
JasonWoof: Well, you could do that, or you could just put all of the app except the uploader onto IPFS
<cryptix>
i should open a tracking issue to get feature parity of 'ipfs object patch' and the writable gateway
<gamemanj>
JasonWoof: So that way bandwidth is only used when A. stuff (including uploads) haven't been cached, or B. someone is uploading something
<cryptix>
JasonWoof: my front-end wizardry is lacking - i guess CSRF might be a problem?
charley has joined #ipfs
<gamemanj>
cryptix: relatively simple, as I said before - verify that it's actually an upload from your app. You can't trust people not to try uploading non-crayon stuff, no matter if you're using IPFS or not :)
elasticdog has joined #ipfs
<JasonWoof>
yeah, this is what I want to avoid
<gamemanj>
cryptix: It doesn't matter if it's a valid upload but it's not technically from your app, though, as long as they aren't trying to put MP4s or stuff through it :)
elasticdog has quit [Changing host]
elasticdog has joined #ipfs
<cryptix>
so you would funnel the uploads through a central crayon.com instance?
<JasonWoof>
I was hoping people could upload into ipfs on their own nodes
<cryptix>
JasonWoof: i'd prefer that as well
<gamemanj>
I hope that becomes a thing, but until then, this solution still saves you bandwidth
<cryptix>
JasonWoof: also think it _should_ be possible somehow
<JasonWoof>
gamemanj: it saves me bw because I don't have to distribute the app?
<gamemanj>
JasonWoof: Yep. You just put the app on IPFS, and have it only contact you directly for uploads.
acidhax has quit [Remote host closed the connection]
<gamemanj>
Downloading uploaded stuff can go through the gateway that the user is using to get at the app, so that'll be cached, by either them or one of the nodes in-between.
<cryptix>
pretty boring though and def not offline-first
<JasonWoof>
gamemanj: I don't understand how I'd know if stuff was from my app or not
<cryptix>
think conference network or flacky bus - we need to get away from those access strategies
<gamemanj>
JasonWoof: Well, you couldn't really verify that exact property, but checks like "Is this file actually a crayon mockup file" and "is the filesize not ridiculous" can go a long way
<JasonWoof>
also, it would be very cool if you could put images into the mockup. But I'm not hosting images for people, this would be a feature I'd only do if it would work completely in ipfs
Encrypt has quit [Read error: Connection reset by peer]
<JasonWoof>
the mockup thing is reasonably difficult to put porn into, but not so with images
Encrypt has joined #ipfs
<nicolagreco>
I think that ipfs on the browser is a good way to go, but my question is, wouldn't that be on a different network? webrtc?
<JasonWoof>
gamemanj: yeah, I can check formatting and filesize
<cryptix>
nicolagreco: yup - i think webrtc is the targeted transport for browser
harlan_ has quit [Quit: Connection closed for inactivity]
<cryptix>
nicolagreco: which is why we need protocol relay - we already need it between ip4<>ip6
<nicolagreco>
but then you have the problem of webtorrent
<nicolagreco>
you either have hybrid clients or you have two parallel networks
<sonatagreen>
feature request: when getting stuff recursively, have an optional limit on how much datasize to grab
<ion>
cryptix: Hmm, reading the RFC, a 301 response to PUT doesn’t mean “okay, your modified resource is now at <location>” but “okay, please PUT the same thing again to this location” (e.g. when the gateway knows which path the mutated object *would* get based on the first request) :-\. Now, if the user were to PUT /ipfs/<newhash>/foo/bar based on the 301, the information about which object to mutate is
<ion>
lost. It looks like PATCH would be the proper method for this.
<sonatagreen>
like, "I want to download the image/video at [hash], unless it's bigger than [size], in which case nevermind"
<sonatagreen>
and, maybe sensible defaults?
<cryptix>
ion: hmmm can i x-post that on the GH issue?
<cryptix>
ion: or do you want to?
<gamemanj>
JasonWoof: Also, you should probably try retrieving the file just after the upload succeeds, as an evil plan to get the user's node to cache the file :)
<ion>
cryptix: I’m fine with both. I can write a post (no pun intended) on the issue, quoting the relevant parts from the RFCs. Would you mind linking the issue?
<ion>
s/both/either/
<multivac>
ion meant to say: cryptix: I’m fine with either. I can write a post (no pun intended) on the issue, quoting the relevant parts from the RFCs. Would you mind linking the issue?
<ion>
multivac: THANK YOU
<cryptix>
hahaha
<JasonWoof>
hmmm... allowing javascript to insert is a tricky problem. the main danger I see is that javascript could fill your cache space, pushing everything else out, which severely loweres the likelyhood that stuff written to your local cache will survive long enough to be useful... I guess this is why browsers implement separate storage limits per domain
<cryptix>
s/ha/haha/
<multivac>
cryptix meant to say: hahahaha
<cryptix>
s/ha/haha/
<multivac>
cryptix meant to say: hahahahaha
<cryptix>
s/ha/haha/
<multivac>
cryptix meant to say: hahahahahaha
<cryptix>
s/ha/haha/g
<multivac>
cryptix meant to say: hahahahahahahahahahahaha
<cryptix>
s/ha/haha/g
<multivac>
cryptix meant to say: hahahahahahahahahahahahahahahahahahahahahahahaha
gaboose has quit [Quit: Connection closed for inactivity]
<ion>
thanks
acidhax_ has joined #ipfs
<gamemanj>
JasonWoof: What I mean is: from the javascript, download the file from the gateway the rest of the app is coming from after uploading it via your site. Unlikely to save you bandwidth, but you can always try :)
nicolagreco has quit [Read error: Connection reset by peer]
<cryptix>
ion: since it was disabled for so long, we might as well get the semantics right before re-enabling it
nicolagreco has joined #ipfs
<cryptix>
gamemanj: please always think how that would work if your friend comes over, shows you this awesome app and then your router goes up in flames
inconshreveable has joined #ipfs
<gamemanj>
For a large file, it would be a terrible idea.
lazyUL has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
acidhax has quit [Ping timeout: 264 seconds]
dignifiedquire has quit [Quit: dignifiedquire]
<JasonWoof>
gamemanj: yeah, this isn't exciting me. if the --writeable becomes standard, then an ipfs-only (offline-first) mode would be possible
domanic has joined #ipfs
<JasonWoof>
I wish I had time to implement a thing where you teach your computer who your friends are, and you all auto-pin stuff the others care about
gritzko_ has quit [Ping timeout: 252 seconds]
rehelmin has joined #ipfs
<JasonWoof>
this way you can get reasonable uptimes and backup/redundancies even if people are mostly just on laptop
<JasonWoof>
s
<JasonWoof>
maybe somebody will implement it before I get around to it
<JasonWoof>
I gotta finish another project first, which will probably take another year
jamescarlyle has quit [Remote host closed the connection]
<sonatagreen>
step 1: subscribe to an ipns address to automatically get pinged when it updates
<whyrusleeping>
sonatagreen: on my todo list!
<sonatagreen>
eeeeeexcellent
* sonatagreen
should learn go so as to contribute more directly than cheerleading
<whyrusleeping>
probably wont happen until around january though, my todo list is pretty big
chriscool has quit [Ping timeout: 255 seconds]
lazyUL has joined #ipfs
lazyUL has quit [Client Quit]
jamescarlyle has joined #ipfs
jamescarlyle has quit [Remote host closed the connection]
jamescarlyle has joined #ipfs
jamescarlyle has quit [Remote host closed the connection]
wopi has quit [Read error: Connection reset by peer]
inconshreveable has quit [Remote host closed the connection]
wopi has joined #ipfs
lazyUL has joined #ipfs
Nitori_ is now known as Nitori
rehelmin has quit [Quit: Leaving.]
charley has quit [Read error: Connection reset by peer]
<geir_>
It's inside docker but i've exposed 5001 and 8080
<codehero>
ah
<ion>
codehero: Add a directory and the hash will point to it while everything under it retains its original filename. You can add a single file with a directory wrapper, ipfs add has a parameter for that.
<codehero>
hm
border0464 has quit [Ping timeout: 256 seconds]
<codehero>
ion: oh. okay
<codehero>
well. not perfect, but it works, i suppose
Encrypt has joined #ipfs
border0464 has joined #ipfs
inconshreveable has joined #ipfs
inconshreveable has quit [Remote host closed the connection]
hellertime has quit [Ping timeout: 264 seconds]
<jamescarlyle>
@cryptix, can I ask a question on writable gateways
mugginess has joined #ipfs
mugginess has quit [Excess Flood]
wopi has quit [Read error: Connection reset by peer]
noob__ has joined #ipfs
wopi has joined #ipfs
<noob__>
im back hahaha
<ipfsbot>
[go-ipfs] MichaelMure opened pull request #1892: Add log events when blocks are added/removed to/from the blockstore (master...blockstore_log) http://git.io/vW0e7
<noob__>
sorry for the dumb questions
<noob__>
how do i kill the daemon
<noob__>
i thought i did but it says
<noob__>
error: ipfs daemon is running please stop it to run this command
<codehero>
killall ipfs
<codehero>
that should do the job
<noob__>
thanks for the quick response
<codehero>
np
<noob__>
boom* thanks
<codehero>
^^
<noob__>
still learning
<noob__>
love the file system though
<codehero>
yeah. it's pretty sweet
<noob__>
game changer :)
noob__ has quit [Client Quit]
<vanila>
yeah game changer is what I'm thinking too
<codehero>
and he's gone
<codehero>
huh
<vanila>
so much potential
<codehero>
yupp
<codehero>
i'm just not sure how i'm supposed to make the fuse mounting work. lol
nicolagreco has joined #ipfs
mono is now known as monokles
<mungojelly>
i haven't gotten fuse mounting to work yet
<codehero>
heh. well. at least i'm not the only one
voxelot has joined #ipfs
nicolagreco has quit [Client Quit]
<mungojelly>
"Error: fuse failed to access mountpoint /ipfs
<whyrusleeping>
jamescarlyle: i dont think cryptix will be available for a little while
<whyrusleeping>
mungojelly: check the fuse docs in our docs folder on go-ipfs
<ion>
whyrusleeping: That sounds ominous.
<mungojelly>
ok will do :)
<jamescarlyle>
thanks, just wanted to find out more about timing of writable gateway
<whyrusleeping>
ion: haha, he might be... preoccupied
<whyrusleeping>
jamescarlyle: timing?
<jamescarlyle>
i thought he said earlier that he had submitted a PR for this..
<jamescarlyle>
i wanted to find out what capability was being addressed, and when
<codehero>
you need to create the /ipfs folder and chown it
nicolagreco has joined #ipfs
<mungojelly>
codehero: chown them to my user account? i haven't bothered to make ipfs an account, i guess that would be better eh
nicolagreco has quit [Client Quit]
<codehero>
yeah. to your user
<codehero>
oh. you use an extra account for ipfs?
<mungojelly>
no i just always run everything on the same account while also knowing that's bad practice :)
<whyrusleeping>
me too
<whyrusleeping>
but i'm fairly confident in the code, worst its gonna do is eat my ram
<achin>
om nom nom
nicolagreco has joined #ipfs
<ion>
alt-sysrq-f
charley has joined #ipfs
gamemanj has quit [Ping timeout: 268 seconds]
nicolagreco has quit [Client Quit]
<M-amstocker>
hey guys, if anyone is interested in a python/ipfs dev discussion on mondays, respond to this thread on github: https://github.com/ipfs/ipfs/issues/112
<mungojelly>
yay now my fuse mount works thanks :)
<M-amstocker>
i'm just doing a survey to see who's interested
<codehero>
mungojelly: heh. the only thing i have on a different account is steam
<ion>
To me, it would seem pertinent to work on a C implementation for FFI purposes. That would benefit Python, Haskell, Rust, D, browser vendors, everything.
<mungojelly>
i like the web of just ipfs-cat-ing text files suggested by the readme at /ipfs/QmVtU7ths96fMgZ8YSZAbKghyieq7AjxNdcqyVzxTt3qVe/readme , it's sorta gopheresque, there isn't any more web like that yet is there?
<codehero>
i'd love to see a rust implementation
<codehero>
and a haskell one perhaps
<whyrusleeping>
mungojelly: theres a bunch out there, check ipfs/archives
<codehero>
but yeah, a c implementation would help every language
<M-amstocker>
ion, i definitely agree
<M-amstocker>
I was gonna add that to my post but its not really relevant to just python and I also don't have those programming chops
jacobpaul has quit [Remote host closed the connection]
<M-amstocker>
ion, did you mean a c impl of the http api bindings or of the entire ipfs protocol?
<codehero>
i think he means the entire ipfs protocol
<ion>
M-amstocker: A full C implementation
<M-amstocker>
ion, yeah
<M-amstocker>
i think that would be awesome, I know some C but I have zero experience doing C network programming
<mungojelly>
whyrusleeping: ipfs/archives seems to be trying to get together hashes of various important things, which makes sense, but i'm just talking about hashes that link to a few curated other hashes, a new little web in ipfs
<codehero>
M-amstocker: i made an irc bot in c++, and i can tell you that even that little network stuff was painful
<whyrusleeping>
mungojelly: oh, i've seen a bunch around. But nothing i've written down
<codehero>
also. revel at my beautiful commit messages
<ion>
codehero: Yeah, it will certainly be painful and also scary. It’s so easy to create vulnerabilities.
<codehero>
yeah
<codehero>
very easy
<mungojelly>
whyrusleeping: i'll start a little page of my own but all i'll have to link to is that readme. then someone can link to my page and the readme and we're off to the races!? :D
<M-amstocker>
maybe this is a naive thought, but there must be tons of existing networking/code libraries to use right?
<ion>
What is SupernodeRouting? Why are the key hashes different from Bootstrap entries for the same nodes?
<whyrusleeping>
supernode routing is dead, needs to be removed
rendar has quit [Ping timeout: 260 seconds]
<whyrusleeping>
but it was an interim replacement for the dht, where all records would be served by central nodes
<ion>
ok
<whyrusleeping>
mungojelly: do it!
<whyrusleeping>
M-amstocker: i feel confident in the go impl of ipfs because the potential of having any sort of memory issues is almost entirely removed
<mungojelly>
i'd like to pin the works of shakespeare, do you think there's one added already or should i find one
pfraze has joined #ipfs
pfraze has quit [Remote host closed the connection]
<whyrusleeping>
mungojelly: not that i've seen
<mungojelly>
i guess /ipfs/Qmanch,.nucho.u is the standard way of linking? i don't want to make things harder for future crawlers than necessary
<M-amstocker>
whyrusleeping: yeah that is definitely worth something
pfraze has joined #ipfs
<M-amstocker>
whyrusleeping: i mean this discussion is really out of my depth TBH :)
<whyrusleeping>
i would feel comfortable with a rust implementation
CarlWeathers has quit [Remote host closed the connection]
<fingertoe>
How far along is the Rust implementation?
<codehero>
whyrusleeping: does it only work on chrome?
<whyrusleeping>
well, i only have chrome
charley has quit [Remote host closed the connection]
<whyrusleeping>
so....
<whyrusleeping>
yes?
<whyrusleeping>
fingertoe: theres not an impl that i know of
<codehero>
oh. okay
<achin>
ion said earlier that he didn't test firefox or other browsers
<codehero>
:/ well
<M-amstocker>
fingertoe: not at all I don't think?
<fingertoe>
I saw one in Git, but it looks fairly sparse..
<ion>
In any case, the navigator.registerProtocolHandler API will not let you register a handler in a nice manner (fs:/ipfs/… going straight to https://ipfs.io/ipfs/… instead of through a page that receives the fs: URI as an encoded query parameter). The proper solution is for a piece of software outside the web to install a protocol handler to the OS.
CarlWeathers has joined #ipfs
jamie_k_ has joined #ipfs
jamescarlyle has quit []
domanic has joined #ipfs
rehelmin has joined #ipfs
<achin>
fingertoe: do you happen to remember what you saw?
<whyrusleeping>
fingertoe: thats just api bindings
<fingertoe>
ah okay.
<ion>
achin: I've seen things you people wouldn't believe. Attack ships on fire off the shoulder of Orion. I watched C-beams glitter in the dark near the Tannhauser gate. All those moments will be lost in time... like tears in rain...
<achin>
... so beautiful
<M-amstocker>
time... to die...
<whyrusleeping>
ion, if you could come with me, i need you to take a test
wopi has quit [Read error: Connection reset by peer]
wopi has joined #ipfs
<ion>
whyrusleeping: Replicants are known to be great programmers, right?
inconshreveable has joined #ipfs
<whyrusleeping>
lol
<M-amstocker>
I want to make a joke about replicants and replicated datatypes
<M-amstocker>
but I can't figure it out
* whyrusleeping
has faith
charley has joined #ipfs
acidhax_ has quit [Ping timeout: 252 seconds]
dignifiedquire has quit [Quit: dignifiedquire]
charley has quit [Remote host closed the connection]
hoboprimate has joined #ipfs
<mungojelly>
oh ok i need -H for hidden files.. ugh wait to script publishing the hash i just added, do i have to scrape the hash out of this stdout-put? :/ sigh, "add recursively and publish" option maybe
<hoboprimate>
Hi, playing with ipfs. Should 'localhost:5001/ipfs/hash' work? I get a '403 - Forbidden' text in the browser
<hoboprimate>
thank you! so localhost:5001 is just for the webui?
<achin>
the server on port :5001 is the API server (the thing the IPFS command line tools interact with). you can use it to get tfiles, but with a different set of URLs
<hoboprimate>
'localhost:5001/webui' I mean
<achin>
yeah, the webui too, i forgot about that
<achin>
the gateway on localhost port 8080 is pretty must identical to what you get with the public gateways http://ipfs.io/ipfs/hash
<hoboprimate>
ok
<hoboprimate>
no need to go aroun a loop, I can just fetch it from my node :)
rehelmin has quit [Quit: Leaving.]
<achin>
yep
tinybike has quit [Quit: Leaving]
<hoboprimate>
so this is pretty amazing stuff. Watched the latest talk posted on youtube, at Stanford University
<hoboprimate>
keep up the good work
hi has joined #ipfs
hi is now known as Guest22406
<achin>
i'm sure the IPFS team is quite happy to have another happy user!
<Guest22406>
I am happy to join
charley has joined #ipfs
<achin>
users! users everywhere! \o/
<hoboprimate>
:) In the talk, was mentioned about 10,000 users right now
<achin>
i'm not quite sure how that number of derived
domanic has quit [Ping timeout: 246 seconds]
harlan_ has joined #ipfs
<achin>
there are probably less than 300 nodes on the network, but that's not a very good measure of "users"
<ion>
achin: It’s probably just the number of sites Neocities puts on IPFS.
charley has quit [Remote host closed the connection]
<ion>
The number of IPFS nodes would have been a better metric to mention IMHO.
<achin>
the number of "users looking at content served over IPFS" could for sure be in the tens of thousands
<kyledrake>
ion We're using ipfs for archives right now and it updates for the latest versions of the sites, but we're not running Neocities -from- IPFS yet. Want to get the IPNS stuff town solid before we start looking into that.
<mungojelly>
oh i think i just solved it with "fusermount -u" perhaps
<codehero>
oh
<codehero>
k
<mungojelly>
what's the difference between chown user and chown user:user
<ion>
chown user might not affect the group.
<ion>
>If a colon but no group name follows the user name, that user is made the owner of the files and the group of the files is changed to that user's login group.
<ion>
Good to know.\
<codehero>
oh mah goodness. xkcd archive on ipfs
<codehero>
hell yes
Guest22406 has quit [Ping timeout: 246 seconds]
epitron is now known as ^_^
^_^ is now known as epitron
ashark has quit [Ping timeout: 250 seconds]
danielrf has joined #ipfs
charley has joined #ipfs
rehelmin has joined #ipfs
wopi has quit [Read error: Connection reset by peer]
fiatjaf has left #ipfs ["undefined"]
<NeoTeo>
Hi ianopolous :) I'm working on the Swift version of the API and I'm wondering if you had any issues with getting invalid json back from the add command. I'm getting multiple objects in the root and am having to manually wrap and comma separate the result.
charley has quit [Remote host closed the connection]
bedeho has quit [Ping timeout: 256 seconds]
charley has joined #ipfs
qgnox has joined #ipfs
<kpcyrd>
is there a way to get data if you don't know the hash?
<kpcyrd>
like, discover hashes in dht, request them
jhulten has joined #ipfs
<codehero>
i hope not
<ianopolous>
neoteo: Hey there! I didn't have any invalid json from the go ipfs implementation. Which implementation are you using?
charley has quit [Ping timeout: 260 seconds]
<vanila>
hey
<vanila>
we should get all these talks and lectures about IPFS
<vanila>
and host them ON IPFS!
<ianopolous>
neoteo: I think you are referring to the JSON stream some commands return.
<vanila>
and then have a nice viewer that lets people watch them
<ianopolous>
jsut a stream of JSON objects
<ion>
vanila: There’s a copy of jbenet’s Stanford talk on IPFS: /ipfs/QmPkPwNx4rr9X9oma5BsMev6YT2GF9u6y1gDtLxG7NyPJq
jacobpaul has joined #ipfs
jacobpau_ has joined #ipfs
<vanila>
great :D
jacobpaul has quit [Ping timeout: 240 seconds]
<cryptix>
kpcyrd: there is
<cryptix>
ipfs log tail | jq
<codehero>
vanila: you should consider pinning them
<codehero>
because there's no point in hosting stuff on ipfs without seeding :P
<vanila>
yeah I need to set up IPFS on a VPS to do that
<codehero>
i just pinned it on my laptop. no worries :P
<vanila>
i didn't do it yet because I had a problem with go lang versions
<vanila>
easy to sort, just gotta put teh time in
<cryptix>
vanila: you dont need go on the box to deploy it
<codehero>
btw. is there any upper limit on how much upload ipfs uses?
<mungojelly>
yeah i want to token bucket limit it pls
atrapado has quit [Quit: Leaving]
mercora has quit [Read error: Connection reset by peer]
mercora has joined #ipfs
Keiya has quit [Ping timeout: 268 seconds]
<cryptix>
yup, that would be nice
<NeoTeo>
ianopolous Soz, was afk for a sec. Yes, i'm referring to the JSON returned.
Keiya has joined #ipfs
<ianopolous>
neoteo: yeah it returns a stream of json objects, not a single json object
<NeoTeo>
Indeed. It seems Java just deals with it whereas Swift/Cocoa's JSON handler chokes.
<mungojelly>
json doesn't actually support streaming at all
<ianopolous>
Yeah I had to modify my Java JSON parser to handle it
<NeoTeo>
ah, ok. Makes me feel better about having to "bodge" it myself.
Encrypt has quit [Quit: Sleeping time!]
<codehero>
actually. it'd be cool if a few people could 'ipfs get' the video, so i could see how much bandwidth it uses
<ianopolous>
just repeatedly read an object until end of stream
<codehero>
also. when someone gets a file from ipfs. does ipfs get small parts from each node?
<codehero>
like with bittorrent
<codehero>
or is it different?
Matoro has quit [Ping timeout: 272 seconds]
captain_morgan has quit [Ping timeout: 246 seconds]
<NeoTeo>
I'm actually getting back a single object but I'm having to separate the sections manually.
<ion>
codehero: It’s like BitTorrent, although the current go-ipfs implementation is less smart about transferring blocks (and not doing redundant work) than many BitTorrent implementations. That will change sooner or later.
<kpcyrd>
cryptix: thanks
niekie has quit [Ping timeout: 246 seconds]
<codehero>
ion: oh, okay. that's cool. i suppose those optimizations will have to happen anyways if we want peercasting with ipfs
<codehero>
i really hope this will become a thing
<codehero>
peercasting is such an awesome idea
rehelmin has quit [Quit: Leaving.]
vanila has quit [Ping timeout: 268 seconds]
inconshreveable has quit [Remote host closed the connection]
<ion>
codehero: They will have to happen, period. And the developers are determined to make it happen. The daemon can transfer almost an order of magnitude more data than the content you’re downloading consumes at the moment.
<codehero>
so what does bitswap do right now? and what is it supposed to do in the future
<achin>
it swaps bits!
sonatagreen has quit [Ping timeout: 250 seconds]
inconshreveable has quit [Remote host closed the connection]
<ion>
I’m under the impression that when you ask the network for the object QmFoo, pretty much everyone who has it will rush to just shove it down your pipe at the same time at the moment.
<codehero>
achin: lol
<codehero>
here. have a 1
<codehero>
i'll take your 0
<achin>
pretty much!
<codehero>
well, but what dictates how data is swapped
charley has joined #ipfs
<codehero>
is it just random?
<ion>
What do you mean by random?
<codehero>
choosing arbitrary nodes to swap data with
<ion>
I’m not sure what go-ipfs does, but i wouldn’t be surprised if it just tried to dial everyone who says they can provide an object you want.
<codehero>
oh. so as of now it only gets data that you request?
charley has quit [Ping timeout: 250 seconds]
domanic has joined #ipfs
<ion>
I’m under the impression that IPFS will never, ever download data that you didn’t request.
<codehero>
but....but filecoin?
<codehero>
i thought..
<ion>
That’s a thing that will request data from IPFS for storing on behalf of others.
<codehero>
oh, i see
<codehero>
yeah
<codehero>
makes sense
<codehero>
btw. how will filecoin be verified? will there also be gpu mining?
<ion>
I’m afraid I don’t know anything about that.
<codehero>
ah, k
<ion>
Although there will be no GPU mining since the mining is not about producing hashes.
<codehero>
oh. hm
<codehero>
well, i thought that bitcoin mining was an important part of preventing double spending and fake bitcoins
<codehero>
so i expected that filecoin would also need that
<ion>
Other than that, i don’t know how it will work.
niekie has joined #ipfs
<ansuz_>
mining really just means work as proof of honesty
<ansuz_>
in a filesystem setting, that could mean accurately serving up the correct content
<ansuz_>
which you can check via its hash
<codehero>
oh, okay
ansuz_ is now known as ansuz
<ansuz>
(which isn't to say that's how filecoin will do it)
<ansuz>
((I don't work here))
<codehero>
yeah. well, i'll just wait and see
<ansuz>
:)
<codehero>
maybe i can make enough money to get tons of storage and mine filecoins and become filecoin rich
<fingertoe>
Any thought been given to a self encryption system like MaidSAFE is using?
<mungojelly>
why doesn't filecoin work yet? is development private at the moment?
<codehero>
fingertoe: good question
<codehero>
mungojelly: seems like it
inconshreveable has joined #ipfs
<fingertoe>
Probably best that the crytpo be quiet till it's ready -- Folks tend to get hyper about that when it probably isn't really what things oughta be about..
<ion>
mungojelly: jbenet is prioritizing IPFS, Filecoin will be implemented on top of a robust IPFS.
<codehero>
makes sense
qgnox has quit [Ping timeout: 272 seconds]
Johannean has joined #ipfs
bedeho has joined #ipfs
<mungojelly>
i just skimmed the filecoin paper, it seems workable but also it smells like a solution in search of a problem, can't you just pay a smaller amount directly to multiple storage providers, the economics tilts towards that in my head
charley has joined #ipfs
<fingertoe>
The advantage of the Self encryption is that the chunks are total gibberish without the other 3 parts -- Thus you could store data without as much risk regarding what it is -- It's meaningless until a client re-assembles it.
sonatagreen has joined #ipfs
charley has quit [Remote host closed the connection]
<mungojelly>
if trends hold and storage keeps outpacing transfer then we should see nodes that store everything selling *bandwidth*, rather, that being the bottleneck
bedeho has quit [Ping timeout: 255 seconds]
charley has joined #ipfs
<mungojelly>
it's more like "we of course have all meaningful data that's ever been published, but how many terabyte/second can we afford to stream to you"
charley has quit [Remote host closed the connection]
qgnox has joined #ipfs
inconshreveable has quit [Remote host closed the connection]
domanic has quit [Ping timeout: 240 seconds]
<codehero>
hm
charley has joined #ipfs
<codehero>
if net neutrality gets destroyed that could become a problem
charley has quit [Remote host closed the connection]
harlan_ has quit [Quit: Connection closed for inactivity]