Hory_ has quit [Read error: Connection reset by peer]
fleeky_ has joined #ipfs
palkeo has quit [Quit: Konversation terminated!]
<kevina>
lgierth: I don't think there is proper symbolic link support in general in go-ipfs
<kevina>
I could be wrong though.
<Kubuxu>
thanks, I hope you figure it out
<kevina>
Kubuxu: so basically we just want symbolic links to work?
<Kubuxu>
lgierth: would you mind if I change the points in your issue with requirements to numbers? I will be easier for me to ref them in a comment?
b100d has quit [Quit: leaving]
<Kubuxu>
kevina: I think yes, they are already being added into ipfs, just gateway is unable to resolve them :|
fleeky has quit [Ping timeout: 252 seconds]
<lgierth>
Kubuxu: do it, thanks
<kevina>
Kubuxu: lgierth: got to run, I'll look into it later today or tomorrow
<lgierth>
kevina: yeah, the gateway just needs to understand what to do with symlink. similar to how we render a listing or resolve to index.html if it's a directory
<lgierth>
kevina: awesome thanks -- see you, i'll go to bed in a bit
<Kubuxu>
me the same :
<Kubuxu>
same :p
<lgierth>
o/
<kevina>
good night then :)
nunofmn has joined #ipfs
<jbenet>
haad: did you test it? how did you get around the TLS catch-22 problem?
ashark has joined #ipfs
<Kubuxu>
I think they used raw IPs
SuperPhly has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
espadrine has quit [Ping timeout: 260 seconds]
<lgierth>
yeah, no /dns and no /tls yet
ashark has quit [Ping timeout: 245 seconds]
anonymuse has quit []
infinity0 has quit [Ping timeout: 248 seconds]
infinity0 has joined #ipfs
<whyrusleeping>
alu: do you have any pretty demos of VR+ipfs laying around?
infinity0 has quit [Remote host closed the connection]
everywhen has joined #ipfs
Kane` has quit [Remote host closed the connection]
<Kubuxu>
yup, expect to be able to see what is covered by sharness and what is covered by unit tests
cemerick has quit [Ping timeout: 256 seconds]
sametsisartenep has joined #ipfs
shizy has joined #ipfs
vmatekole has quit []
pfrazee has joined #ipfs
eibriel has quit [Ping timeout: 245 seconds]
anonymuse has joined #ipfs
maxlath has quit [Ping timeout: 264 seconds]
ashark has joined #ipfs
maxlath has joined #ipfs
leeola has joined #ipfs
nunofmn has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
nunofmn has joined #ipfs
pringless has joined #ipfs
<pringless>
Anyone know what KDF these guys use? https://github.com/dchest/tweetnacl-js . Most nacl ports I found have a function to generate the shared secret key and stop at that, but this one, doesnt show the shared secret, it just encrypts directly
<Kubuxu>
pringless: all nacl impls will have box and secret box functions, this is nacl interface
<pringless>
Thing is I am trying to get an android version of this, but the one lib I found, stops at calculating a shared secret key : https://github.com/twystd/tweetnacl-android
<Kubuxu>
the android lib you send implements the box interface
<pringless>
but in this, during encryption, they must be calculating the shared secret with the public and private key, and then encrypting the message. So won't they have to use a KDF of some sort
<Kubuxu>
it is refereed in benchmarks
<Kubuxu>
KDF - key derivation function - is function that mixes keys/passwords to create new key
<pringless>
Kubuxu: yes I know what kdf is, but which one is used
<pringless>
is what im stuck at, tried to see source code, but its 3k lines
<Kubuxu>
it doesn't use KDF- it uses EcDH - to create shared secret from public and private keys
<Kubuxu>
SS := curve25519(pub, priv)
<Kubuxu>
then uses SS to encrypt and authenticate using Salsa20+Poly1305
<pringless>
but in ecdh, once the key exchange is done, the encryption is done not with the shared key but a KDF(sharedKey)
<pringless>
kdf like aes
robattila256 has quit [Ping timeout: 245 seconds]
nunofmn has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
maxlath has quit [Quit: maxlath]
Boomerang has joined #ipfs
<Kubuxu>
it doesn't have to be like that, it can be use with just the shared key
<Kubuxu>
there doesn't have to be any KDF
<Kubuxu>
especially with modern crypto primitives
<Kubuxu>
it was used when encryption algorithm is AES as AES has small nonce space and needs to be rekeyed every now and again
kenshyx has quit [Quit: Leaving]
nunofmn has joined #ipfs
pfrazee has quit [Remote host closed the connection]
nunofmn has quit [Ping timeout: 240 seconds]
pfrazee has joined #ipfs
pcre has joined #ipfs
pfrazee has quit [Ping timeout: 248 seconds]
treaki_ has quit [Ping timeout: 258 seconds]
Hory_ has quit [Read error: Connection reset by peer]
<voker57>
I have a directory to which I add new files and republish, any way to avoid rehashing existing files each time?
brianhoffman has joined #ipfs
<voker57>
(more automatic than manually assembling directory object
<Kubuxu>
voker57: not really, if you know the delate that was done on the directory you can use files API `ipfs files --help` to do the same delate in virtual IPFS fs.
<Kubuxu>
brianhoffman: no, but I will take a look at it later today and if isn't a lot of work (no refactoring) I will fix it
<brianhoffman>
cool. thanks. yeah we’re seeing 10s for every single IPNS call. Is that normal?
Boomerang has quit [Read error: Connection reset by peer]
<kenshyx>
victorbjelkholm: nvm it was from ipfs bin
<kenshyx>
so weird
<kenshyx>
{"level":"error","message":"\u001b[0;37m18:55:02.216 \u001b[31mERROR \u001b[0;34m mfs: \u001b[0mroot not a dir: &mfs.File{parent:(*mfs.Root)(0xc4220a4c40), name:\"QmdfTbBqBPQ7VNxZEYEj14VmRuZBkqFbiwReogJgS1zR1n\", desclock:sync.RWMutex{w:sync.Mutex{state:0, sema:0x0}, writerSem:0x0, readerSem:0x0, readerCount:0, readerWait:0}, dserv:(*merkledag.dagService)(0xc420124070), node:(*merkledag.Node)(0xc422089f80), nodelk:sync.Mutex{state:
<kenshyx>
0, sema:0x0}} \u001b[0;37mops.go:168\u001b[0m\n\u001b[0;37m18:55:02.216 \u001b[31mERROR \u001b[0;34mcommands/h: \u001b[0merr: root was not a directory \u001b[0;37mhandler.go:288\u001b[0m\n","timestamp":"2016-12-13T16:55:02.216Z"}
<kenshyx>
I have a full log of this stuff
<victorbjelkholm>
kenshyx: that doesn't seem very good. Unfortunately, I don't have lots of experience with electron and ipfs binary working together, but if you open an issue in the appropriate repository, I'm sure someone knows
se3000 has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
galois_d_ has joined #ipfs
Sophrosyne has joined #ipfs
Sophrosyne_ has joined #ipfs
Sophrosyne_ has quit [Remote host closed the connection]
galois_dmz has quit [Ping timeout: 258 seconds]
gavinmcdermott has joined #ipfs
dmr has quit [Quit: Leaving]
s_kunk has quit [Read error: Connection reset by peer]
s_kunk has joined #ipfs
bwerthmann has joined #ipfs
corvinux has joined #ipfs
<Kubuxu>
brianhoffman: yeah unfrotunatelly anything from 1 to 15s is normal currently
<brianhoffman>
is that ever going to change or just the nature of things?
gavinmcdermott has quit [Read error: Connection reset by peer]
bwerthmann has quit [Ping timeout: 246 seconds]
s_kunk has quit [Ping timeout: 250 seconds]
bwerthmann has joined #ipfs
bastianilso_____ has quit [Quit: bastianilso_____]
corvinux has quit [Ping timeout: 250 seconds]
<lgierth>
Kubuxu whyrusleeping: i'm refactoring all of gc and ripping out corerepo/gc.go, just so you know :)
* lgierth
boy scouts programmer
<lgierth>
so much code, so little use
konubinix has quit [Ping timeout: 258 seconds]
<dignifiedquire>
lgierth: sounds like you are having fun =)
<lgierth>
i am
<Kubuxu>
lgierth: do you want to get it into 0.4.5?
<Kubuxu>
brianhoffman: it will for sure change, we just haven't focused much on optimizing this subsystem yet
<lgierth>
i think so -- my assessment right now is that the net gain is positive -- we get the gc tests back, and cleaner gc code
ygrek_ has joined #ipfs
<lgierth>
the timeouts-related fix can go in independent of this
corvinux has joined #ipfs
gavinmcdermott has joined #ipfs
Encrypt has joined #ipfs
SuperPhly has joined #ipfs
infinity0 has quit [Ping timeout: 252 seconds]
cemerick has joined #ipfs
<whyrusleeping>
lgierth: do it
SuperPhly has quit [Client Quit]
gavinmcdermott has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<Kubuxu>
whyrusleeping: I have sharness successfully just run from make with -j 12
<Kubuxu>
from new make
rendar has quit [Ping timeout: 240 seconds]
<whyrusleeping>
Kubuxu: that scares me
<whyrusleeping>
but good job anyways :D
<Kubuxu>
chris during the Lisbon workshop said to me: don't even try it. my thought was: there is no reason it shouldn't work, apart from someone statically allocating port 5050 two times somewhere in the makefile (this isn't the case but it would probably be the reason it would break).
ygrek_ has joined #ipfs
bsm117532 has quit [Ping timeout: 256 seconds]
<Kubuxu>
running in parallel reduces the time by factor of 3.
se3000 has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<Kubuxu>
voker57: it does not, currently unfortunately it gets slower and slower as you have more pins
<ianopolous_>
rfc4648 defines the alphabet as A-Z and 2-7, the highest char of which would be Z
<Kubuxu>
ianopolous_: hex is one that alphabet start as hex alphabet, base32 is the other defined in the RFC which starts with A
<Kubuxu>
base32hex is 0 to V and base32 is A to 7
<Kubuxu>
afaik
<ianopolous_>
so base32 or base32pad then, but they are listed with a highest char of B or C in that table, I get that Z is taken, just trying to understand why B and C
pp99s has joined #ipfs
<voker57>
Kubuxu: I have only few recursive pins but their size is very big
<Kubuxu>
no idea
<Kubuxu>
voker57: you can do ipfs refs -r HASH to first fetch it
<Kubuxu>
and ipfs pin add should be fast
M-mistake has quit [Quit: Client limit exceeded: 5000]
<lgierth>
while "dynamodb can't store empty strings" is on HN #1 -- empty unixfs file: /ipfs/QmbFMke1KXqnYyBBWxB74N4c5SBnJMVAiMNRcGu6x1AwQH -- empty unixfs dir: /ipfs/QmUNLLsPACCz1vLxQVkXqqLX5R1X345qqfHbsf67hvA3Nn -- empty block: /ipfs/QmdfTbBqBPQ7VNxZEYEj14VmRuZBkqFbiwReogJgS1zR1n
arpu has quit [Ping timeout: 256 seconds]
<ansuz>
OWNED
corvinux has quit [Ping timeout: 268 seconds]
dmr has joined #ipfs
<ianopolous_>
:-)
arpu has joined #ipfs
<des_consolado>
can someone tell me a thing about indexing/searching with this? I just installed this and now I'm like "okay... can I find some files from some strangers now?"
corvinux has joined #ipfs
eibriel has joined #ipfs
lemon_ has quit [Quit: Lost terminal]
<Mateon1>
des_consolado: There are projects that listen to DHT messages, such as ipfs-search.com, it's far from perfect, though. You can also look at ipfs.pics, which is an IPFS based image service.
warner has quit [Quit: ERC (IRC client for Emacs 25.1.2)]
bsm117532 has quit [Remote host closed the connection]
bsm117532 has joined #ipfs
bsm117532 has quit [Remote host closed the connection]
<des_consolado>
Is there something happening with this? Are there some projects or mailing lists or something, for development on solutions of this problem? I mean that's all that's needed now right?
bsm117532 has joined #ipfs
<des_consolado>
I'm just here trying to get hold of some stuff that used to be very easy to get hold of back in the day... it's really shocking what they've done with the file-sharing community this last few years, torrents and newsgroups are really just depressing now
<des_consolado>
I can see it being bloody awesome once this kind of distributed/decentralized/p2p/more secure technology comes in though
<des_consolado>
I'm a pretty noobish developer but if there's something happening in terms of bringing this up to speed with what we used to have... I'd at least like to be tuned in for updates, you know, heh...
<des_consolado>
maybe I should check #ipfs-dev
chris613 has joined #ipfs
patcon has joined #ipfs
Encrypt has quit [Quit: Quit]
<r0kk3rz>
des_consolado: realistically theres nothing stopping people from posting IPFS links on reddit or something similar
fleeky__ has quit [Ping timeout: 258 seconds]
<des_consolado>
reddit wouldn't like that if people are posting things that aren't above board
<Mateon1>
Well, people do in fact use ipfs (mainly ipfs.pics) on Reddit
<des_consolado>
besides, it would be cool to stick with the decentralized thing... surely it'd be possible to have some way to be able to do the same kind of shit you do on a website, like have a bunch of links, accounts and reputations, forums, etc etc etc, but not on a site, on the decentralized p2p platform thing...
<Mateon1>
It's still very small, though
<Mateon1>
(the use of IPFS)
Guest18 has joined #ipfs
patcon has quit [Ping timeout: 248 seconds]
<r0kk3rz>
des_consolado: of course, there are a few useful ways of doing that, nothing automagic though
Guest18 has quit [Client Quit]
robattila256 has joined #ipfs
<voker57>
ipfs pin add of 18GB has been working for several hours now
<voker57>
is it safe to ctrl-c it?
<lgierth>
voker57: yes
<lgierth>
it probably stalled while fetching -- this can happen at the moment if the node you're fetching from restarts
<voker57>
no fetching, everything is local
<lgierth>
does ipfs refs -r complete?
fleeky has joined #ipfs
corvinux has quit [Ping timeout: 258 seconds]
<voker57>
it's working, outputting 5kb/s of references into /dev/null, will probably take the same amount of time
<lgierth>
if you had all of that hash local, it should finish immediately
<voker57>
I'm sure it's 100% local. daemon is reading 22mbs from disk now
<lgierth>
mh. yeah it has to go through all objects in order to resolve links...
<lgierth>
slow spinning disks?
sametsisartenep has quit [Quit: zzz]
<voker57>
not bleeding age of hard drive technology, Seagate Barracuda 7200.14
<lgierth>
it's still annoyingly slow -- wanna file an issue so we can gather info to track this down?
<lgierth>
also include info about which filesystem your ~/.ipfs is on
slothbag has joined #ipfs
<voker57>
k thanks
<voker57>
refs finished in 9 minutes
PseudoNoob has quit [Remote host closed the connection]
cemerick has quit [Ping timeout: 248 seconds]
<voker57>
interestingly, pin add results in _lots_ of daemon threads being launched which do not appear to help much since they all wait for IO
<lgierth>
mh, that sounds like contention simply from the parallelism of pin add (cc whyrusleeping kevina)
<lgierth>
voker57: what you can try is `ipfs refs -r <hash> | ipfs pin add`
<lgierth>
(i hope that construction actually works)
<voker57>
eh, I don't want to directly add all the chunks
ulrichard has quit [Ping timeout: 264 seconds]
<lgierth>
yeah good point :) just proposing it to see if it gets faster if there is no pinning in parallel
<lgierth>
you can do the same but with pin rm, to reverse it
<voker57>
ok now pinning finished in 14m. caches something?
HostFat has joined #ipfs
HostFat has quit [Remote host closed the connection]
HostFat has joined #ipfs
<kevina>
voker57, lgierth: I don't know enough about that "parallelism of pin add" to really know what is going on, sorry
galois_d_ has quit []
bastianilso_____ has quit [Quit: bastianilso_____]
<OstlerDev>
Hey there! I am currently working with a web uploader that uploads directly to one of our IPFS nodes (using a multipart POST), I just was testing it out and attempted to upload 7 files into a folder and the upload seemed to go perfectly fine, however it only returned the hashes for 3 files instead of all 7 hashes and the folder hash. Any thoughts as to why this might be happening?
<lgierth>
whyrusleeping: for applying this pattern in other situations, there might have to be a way for the caller of the iterator to know when there's an error (e.g. a context.Done())
<lgierth>
e.g. = a kind of error that might have to be detected in the caller
anonymuse has quit [Remote host closed the connection]
ianopolous_ has quit [Remote host closed the connection]
<whyrusleeping>
Yeah, agreed
<whyrusleeping>
that func should be able to return an error as well
<whyrusleeping>
or a 'result' that contains either an error or a value
cubemonkey has quit [Remote host closed the connection]