Eudaimonstro has quit [Remote host closed the connection]
simonv3 has quit [Quit: Connection closed for inactivity]
rubiojr has quit [Ping timeout: 250 seconds]
ehd has quit [Ping timeout: 250 seconds]
rubiojr has joined #ipfs
Eudaimonstro has joined #ipfs
ehd has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Eudaimonstro has quit [Ping timeout: 245 seconds]
domanic has quit [Ping timeout: 264 seconds]
Leer10 has joined #ipfs
<daviddias>
jbenet: I think you are good :)
domanic has joined #ipfs
<M-davidar>
oops, filled sirius' drive
<M-davidar>
will transfer to pollux
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
voxelot has joined #ipfs
<M-davidar>
lgierth: tried freeing some space, but still getting no space left errors (can't even initiate transfer) - help?
<lgierth>
M-davidar: taking a look
M-davidar is now known as davidar
inconshreveable has quit [Ping timeout: 246 seconds]
* lgierth
deleting the incomplete openwrt mirror
<lgierth>
that's something like 220G
<davidar>
cool, thanks
<lgierth>
158G
<lgierth>
close enough :]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<davidar>
transferring to pollux now
<davidar>
hmm, the bandwidth between sirius and pollux isn't great
<lgierth>
meh
<lgierth>
what do you get?
<davidar>
1-2MB/s
<davidar>
much slower than it took to download from s3 in the first place
<datagrok>
I'm setting up IPFS for the first time behind a router with NAT. Which ports should I forward?
<datagrok>
(using the go-ipfs implementation)
<lgierth>
davidar: oh meh :/
<davidar>
lgierth: ETA ~1week to transfer ~700GB from sirius to pollux... :/
<lgierth>
i had about 11MB/s mirroring from downloads.openwrt.org
<lgierth>
the 158G more aren't enough i guess?
domanic has quit [Ping timeout: 245 seconds]
<davidar>
lgierth: perhaps temporarily, but I still need to transfer to pollux at some point to have room for adding to ipfs
domanic has joined #ipfs
<lgierth>
hrm right
<lgierth>
wanna try copying directly into a fuse mount?
<lgierth>
unless you need file timestamps
<davidar>
lgierth: if you think that would be faster than scp?
<davidar>
oh, you mean with ipfs
<davidar>
is that possible yet?
<lgierth>
it'd save you time and diskspace if it works (it should)
notduncansmith has joined #ipfs
inconshreveable has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
amstocker has joined #ipfs
<davidar>
hmm, might check with jbenet before I try anything
<davidar>
need to wait for it to finish downloading anyway
<davidar>
hopefully the extra 150GB will be enough :/
<jbenet>
hmmm wow 1-2MB/s? with rsync?
<jbenet>
sucks
<jbenet>
prob sirius outgoing bw sucks.
<davidar>
jbenet: scp, but i imagine rsync would be similar
amstocker_ has joined #ipfs
<jbenet>
davidar: rsync usually way faster than scp
<davidar>
jbenet: ok, i'll try that later
pfraze has joined #ipfs
<davidar>
jbenet: also, citeseer wants an aws acct to whitelist, i'll forward the email
amstocker has quit [Ping timeout: 255 seconds]
<jbenet>
davidar: give them yours maybe?
<jbenet>
should work
<jbenet>
(its under eng@ipfs.io)
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
jedahan has joined #ipfs
<davidar>
thanks
pfraze has quit [Remote host closed the connection]
<davidar>
i might leave it until arxiv is finished (hopefully soon), since we're only going to have access for a week
<whyrusleeping>
datagrok: if your router supports upnp or nat pmp, then you dont need to forward anything
Eudaimonstro has joined #ipfs
jedahan has quit [Remote host closed the connection]
fiatjaf has left #ipfs ["undefined"]
jedahan has joined #ipfs
<jbenet>
davidar: sgtm
Eudaimonstro has quit [Ping timeout: 245 seconds]
Eudaimonstro has joined #ipfs
amstocker_ has quit [Ping timeout: 246 seconds]
therealplato1 has quit [Quit: Leaving.]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
patcon has joined #ipfs
amstocker has joined #ipfs
<datagrok>
whyrusleeping: i've got upnp disabled, not sure if my router uses nat pmp. In that situation should I just forward the port used for the "swarm address"?
<drathir>
jbenet: sorry for that, need to lear how using that... thanks a lot for imforming... im pretty fresh with git plays... better is makin the web changes or local repository ones?
* cryptix
<3 podcasts
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<Bat`O>
jbenet: i feel like saying "ipfs is close to git" is misleading for most people because ipfs is close to git *internals*, and peoples usually think about git in term of branch, commit, push, repository
<pjz>
did you guys get the private sharing piece done yet? :)
<pjz>
also: don't pussyfoot around with pypi/npm repos... go straight for the big guys: ubuntu and debian repos on ipfs. apt-get is already set up for transport plugins; apt-p2p is a thing (uses its own DHT)
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<daviddias>
cryptix: me too :)
<daviddias>
wait, jbenet was this your important call? You could have told us, this is great! :D
<pjz>
given ipfs.js should go-ipfs be ipfs.go ? and py-ipfs become ipfs.py ?
<chriscool>
drathir: GitCop checks the commit message of the commits in the branches you send to GitHub
<chriscool>
it doesn't check the PR message
<daviddias>
whyrusleeping: do you have the new IPRS in go-ipfs already?
mildred has quit [Ping timeout: 244 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
daviddias: nope. we're waiting on the ipld stuff
<chriscool>
yeah, you probably have to modify your commit localy (using git commit --amend for example) and then force push the branch (git push -f <remote> <branch>)
<daviddias>
for the IPRS? How so
<daviddias>
?
<whyrusleeping>
daviddias: because jbenet wants to implement the record objects using ipld
<daviddias>
oh, the protobuf vs jsonld , right
<whyrusleeping>
yeap
<pjz>
IPRS could be bitcoin :)
<daviddias>
I guess my big question was: We use MerkleDAG objs for Records to chain the provides of the same blog, right?
<pjz>
I liek it
<pjz>
I typo it to, apparently
<whyrusleeping>
daviddias: not quite sure on that yet. we havent gotten to it yet
<daviddias>
whyrusleeping: it is on the spec, but I'm not 100% of the reason anymore
notduncansmith has quit [Read error: Connection reset by peer]
voxelot has quit [Ping timeout: 240 seconds]
<whyrusleeping>
we use merkledag objs for records so that we use merkledag objects for records
edrex has quit [Remote host closed the connection]
<daviddias>
watman :D
edrex has joined #ipfs
<daviddias>
whyrusleeping: I didn't understand
<whyrusleeping>
daviddias: the merkledag obj isnt used to chain provides records together
<daviddias>
not even for reprovides?
<whyrusleeping>
its just a merkledag obj so it can do things like link to a public key, or something
<daviddias>
Now I get it
<whyrusleeping>
what were you thinking for reprovides?
<whyrusleeping>
i'm curious
<drathir>
chriscool jbenet k i guess have locally setup that suggested configuration wonder if it will workin corectly...
<daviddias>
I was thinking that each block would have its own chain and we would just keep adding to the tip, each time we were reproving, getting an idea of where the blocks have been through time
border has joined #ipfs
<whyrusleeping>
daviddias: while i think that would be a cool idea, the storage space required for that on a large network would be monstrous
<daviddias>
true, if there is a lot of peers reproviding, the chain can get quick large
<whyrusleeping>
especially since, for any large file there will be hundreds to thousands of blocks
<daviddias>
well, makes sense we have on DAG node per block provided. Somehow I had thought the Chain idea was a thing now
voxelot has joined #ipfs
voxelot has joined #ipfs
<whyrusleeping>
we are going to do chains for ipns entries
<daviddias>
are we gc'inc on fetch? Like, instead of replying an expired block, just gc it and avoid putting that in the network?
* drathir
like git idea/behaviour of keep merging the same files and link only to correct file, that could be used in network to pararel download file from multiple sources have that correct file requested?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
drathir: that was an intense /me
<drathir>
whyrusleeping: in eg irssi it point to own username its /me = drathir
<whyrusleeping>
yeah, i know
<whyrusleeping>
but i normally see it used for things like:
* whyrusleeping
yawns
* whyrusleeping
needs coffee
<drathir>
oh in that case its more like the "free minds"...
<alu>
tfw all out of coffee
<drathir>
even not sure if technically possible to implement...
<whyrusleeping>
alu: :(
<alu>
I might have to make an emergency trip
* daviddias
misses vita
<alu>
btw I move out to cali in 2 weeks
<alu>
:)
<drathir>
alu: maybe tea help a little too?
<alu>
I got yyerba mate tea
<whyrusleeping>
daviddias: remember that convoy coffee place?
<alu>
so yeah I suppose
<daviddias>
cafe vita that is
<whyrusleeping>
that was gonna open in IH?
marianoguerra has quit [Ping timeout: 260 seconds]
<daviddias>
I remember, didn't have the chance to try it though
<ipfsbot>
go-ipfs/feat/mfs bef0bd8 Jeromy: make ipns fuse mount use mfs...
<ipfsbot>
go-ipfs/feat/mfs 8a67ecb Jeromy: move session option up to root...
<drathir>
idea is nice, wonder if ppl interested and buy from them...
<border>
Im always interested into buying coffee :)
<drathir>
;p
<daviddias>
whyrusleeping: do you have admin on https://github.com/ipfs/node-ipfs ? can you add me as colab? Want to deprecate that one and make a new one with roadmap
<drathir>
bc its really interesting looks and all move with bike...
<whyrusleeping>
daviddias: added the nodeipfs team
<daviddias>
thanks
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<ipfsbot>
[node-ipfs] diasdavid pushed 7 new commits to master: http://git.io/vs7Ir
<ipfsbot>
node-ipfs/master 8a9b111 David Dias: peer ID module
<ipfsbot>
node-ipfs/master e824612 David Dias: ipfs-logger, a wrapper around bunyan to create several loggers for each layer of IPFS
<ipfsbot>
node-ipfs/master 3bd3a0f David Dias: DHT, mDNS (need to break commits more from now on)
<daviddias>
github still doesn't let me change the name. whyrusleeping could you change https://github.com/ipfs/node-ipfs to node-ipfs-deprecated or node-ipfs-old
<daviddias>
so I get the node-ipfs name available
<whyrusleeping>
sure thing
<whyrusleeping>
donezo
<voxelot>
now your link to the new repo just links to the depraciated :p
akhavr1 has quit [Read error: Connection reset by peer]
<daviddias>
Thank you Sir! :)
<voxelot>
deprecated*
<daviddias>
not anymore
<voxelot>
:)
<daviddias>
magic! :D
akhavr has joined #ipfs
<voxelot>
voila
<drathir>
the/w 20
<drathir>
err...
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<rjeli>
what can i read to learn about merkle DAGs?
<rjeli>
what does "the hash(object) includes all hash(linked_object)" mean?
<mmuller>
yeah, sorry - didn't know about the distinction myself :-)
kord has joined #ipfs
<kord>
I'm writing up a spec sheet today for a distributed docker hub. The company I'm working for, Giant Swarm, is seriously considering building and supporting the project. I have a few technical questions.
<daviddias>
kord: make them come :) //cc jbenet whyrusleeping
therealplato has quit [Ping timeout: 260 seconds]
<kord>
As far as I can tell, the API will respond to a localhost connection. Given we'll be doing this in a containerized service (multiple containers) I need to know if an auth token can be used to authenticate another IP to access the IPFS API.
mildred has quit [Ping timeout: 244 seconds]
<kord>
Thanks daviddias!
<kord>
Specifically, we'll need to be able to write the layers that comprise a docker image into IPFS.
mildred has joined #ipfs
<daviddias>
We do intend to have other forms of authentication (one discussion here https://github.com/ipfs/notes/issues/13) and we've been recently talking about a capability system
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<daviddias>
Right now we still do filtering through host + port, not knowing your timeline, you could get it working today with a proxy that has your auth in front of it
therealplato has joined #ipfs
<kord>
That would work. Microservice requirements would be that I have a token I can pass to the proxy, which is running in a seperate container.
<kord>
I thought about making a fat container for it, but it would be better to keep the ipfs process inside it's own container.
mildred has quit [Ping timeout: 260 seconds]
Eudaimonstro has joined #ipfs
<rjeli>
so, the merkle dag
<rjeli>
there's no hierarchy, and each object's hash is a hash of its data and its links?
<rjeli>
if that's right then i get it
<drathir>
the connected each other peers have ability askin about what hashesh node have?
<drathir>
hashes mean file hashes which serve...
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
mildred has joined #ipfs
voxelot has quit [Ping timeout: 244 seconds]
voxelot has joined #ipfs
Encrypt has quit [Quit: Dinner time!]
Tv` has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
bedeho has joined #ipfs
danslo has quit [Remote host closed the connection]
danslo has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<lgierth>
and that little daemon would figure out which notice is relevant
<whyrusleeping>
oh, that works
<lgierth>
it needs to keep that data in memory anyhow
<whyrusleeping>
and the keys are just hashes, yes?
<lgierth>
yup, a list of refs with a $reason and $noticeText
<lgierth>
with reason = dmca | code-of-conduct
<lgierth>
can we have the gateway check against multiple denylists?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<lgierth>
it might make sense to respond with something different than 451 in case of a code-of-conduct violation (code-of-conduct would basically be a completely separate denylist)
<lgierth>
see discussion pad ^
<lgierth>
code-of-conduct might be out of scope for now
<whyrusleeping>
i could probably do multiple denylists
<whyrusleeping>
that adds a bit more complication to things though
<lgierth>
i'm fine with leaving the code-of-conduct denylist out for now
<whyrusleeping>
okay
<lgierth>
this has been sitting around for too long
<whyrusleeping>
once we get the system up and running it should be easier to tweak it
<whyrusleeping>
if i want to only serve certain files
<whyrusleeping>
for example, maybe i run a website where i host linux isos
<whyrusleeping>
and i want to host them through ipfs
<whyrusleeping>
but i dont want random people to download anything they want through my gateway
<lgierth>
kek
<lgierth>
sounds good
chriscool has quit [Ping timeout: 246 seconds]
<pjz>
whyrusleeping: I just used it to share a screenshot with some friends :) 'ipfs add screenshot.png' locally then point them at the gateway URL :)
<whyrusleeping>
pjz: noice. i find it much easier to use ipfs to share things with people than pretty much anything else
<whyrusleeping>
unless i'm running a development daemon that doesnt talk to the rest of the network.. then its hard
<pjz>
heh
patcon has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
domanic has quit [Ping timeout: 246 seconds]
mildred has quit [Ping timeout: 255 seconds]
amstocker has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
krl: ping
<whyrusleeping>
and/or daviddias
<amstocker>
if I insert an object manually into ipfs, when I set up the links will the network check that the size is correct?
<kord>
whyrusleeping sorry, been writing specs for this. Yes, the basic idea is to emulate a docker hub API locally in a container.
<kord>
Thinking we put a Go based API proxy in a container, link it to the IPFS container, and then have the proxy translate the relevant bits of the writes by the docker client into the IPFS API.
kord_ has joined #ipfs
<kord_>
I wasn't considering rewriting the graph functionality of docker though.
dlight has quit [Remote host closed the connection]
kord has quit [Ping timeout: 244 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<kord_>
I think we can do a mapping of layer file data to a dag. Make the layer data blocks children to a few blocks that list github repo+commit hashes so users can map (both ways) from a container image to a repo, or from the repo to a container image.
<jbenet>
kord-- i'm actually working on this at the moment. we should work together on it
<kord_>
Ok.
<kord_>
jbenet, you appear to be in PA, correct?
sharky has quit [Quit: Leaving]
<jbenet>
i'm mid something right now, but if you're around in 1hr, i can give you a rundown of the plan so far.
<jbenet>
kord_ on average. i'm in berlin right now
<kord_>
oh, ok. +9 then. I bounce from Moraga (east bay) to Cologne.
<kord_>
back in an hour and a half. riding. talk to you then.
amstocker_ has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
amstocker has quit [Ping timeout: 246 seconds]
<jbenet>
krd_ sgtm.
<jbenet>
catching up with irc <<<EOF
<jbenet>
pjz: yeah, can you help us with apt? would be awesome.
<jbenet>
daviddias: no it actually was a different important call. :)
<jbenet>
pjz: "given ipfs.js" -- no, "ipfs.js" will be a browser-specific bundling of "node-ipfs".
<jbenet>
whyrusleeping: the IPLD stuff is waiting on cbor and waiting on protobuf. i ran into a test bug i couldnt figrure out-- did you get a chance to look at it?
<jbenet>
daviddias: why deprecate a repo? just a commit that blows it all out. yay version history.
<jbenet>
on the merkledag vs merkletree distinction-- see https://github.com/jbenet/random-ideas/issues/20 -- i was tired of people getting confused when i said "merkle-tree" because people think of _specifically_ ralph merkle's signature revealing scheme.
<jbenet>
whyrusleeping: nice convoy promo. y u no ipfs hoodie!? those things are coveted!
<jbenet>
lgierth whyrusleeping: can we use maybe "ipfs.io/refs/lists/denylists" <-- and put a deny list there like "ipfs.io/refs/lists/denylists/dmca" ? that way we can have other lists like "ipfs.io/refs/lists/archives/archive.org/HEAD" or "ipfs.io/refs/lists/archives/arxiv.org/HEAD" -- can discuss on github.
<jbenet>
amstocker: unfortunately not. you can lie about the object size. i've considered removing it for this reason, and have nodes calculate it. krl brought up some other good reasons to do similar things. not sure yet!
<jbenet>
EOF
chriscool has joined #ipfs
<jbenet>
daviddias: a bunch of people were watching the node-ipfs repo-- may be worth just blowing out the code in the original repo for the watchers' sake
* Bat`O
caught another bug
<Bat`O>
sometimes, i get erroneous api reply for ipfs swarm peers
<daviddias>
jbenet: oh, didn't thought of that. I thought of deprecating because it changes completely the code base
<daviddias>
but we can revert
<daviddias>
will take me just a couple of mins
<daviddias>
whyrusleeping: you pinged?
<whyrusleeping>
daviddias: what is the status of th webui?
<daviddias>
krl had a version with iFrames, which AFAIK was never merged. The next step is to make it with WebComponents
inconshreveable has joined #ipfs
pfraze has joined #ipfs
<ipfsbot>
[node-ipfs-old] diasdavid pushed 1 new commit to master: http://git.io/vsd2M
<ipfsbot>
node-ipfs-old/master 34253cb David Dias: Update repo, deprecate old code, add roadmap
<ipfsbot>
[node-ipfs-old] diasdavid pushed 1 new commit to master: http://git.io/vsd2N
<ipfsbot>
node-ipfs-old/master 0928dc9 David Dias: Add right README
ianopolous has joined #ipfs
<daviddias>
jbenet: reverted the 'deprecation', added new README, removed code. can you call it again node-ipfs? (I can't access settings of that repo)
<daviddias>
whyrusleeping: are you on building IPLD stuff?
<daviddias>
(seeing that you are playing with CBOR)
kord has joined #ipfs
<whyrusleeping>
daviddias: lol, i'm looking for an implementation of cbor in go that isnt neglected
<daviddias>
got it :)
<whyrusleeping>
the ugorji one is a pile of crap, its not possible to modify it as the tools he used have changed since he used them
<whyrusleeping>
and its a giant mess of weird go generate templates and shell scripts
<daviddias>
no vendoring, ahn?
<whyrusleeping>
yeap, none
<whyrusleeping>
cespares cbor lib has a nice looking encoder i could work with
Encrypt has quit [Quit: Sleeping time!]
<kord>
jbenet ping
<whyrusleeping>
but no decoder
<whyrusleeping>
so thats out
<whyrusleeping>
i'm grabbing one from google code now
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<daviddias>
`JavaScript
<daviddias>
JavaScript implementations are becoming available both for in-browser use and for node.js`
<daviddias>
oh my :(
<whyrusleeping>
whoops, looks like he migrated it to bitbucket
<whyrusleeping>
my poor github activity feed...
<daviddias>
heresy!
<daviddias>
it seems that the Node.js one only runs in Node.js and it was developed in coffee script