Guest18193 has quit [Read error: Connection reset by peer]
}ls{ has quit [Ping timeout: 248 seconds]
}ls{ has joined #ipfs
chriscool has quit [Ping timeout: 248 seconds]
jungly has joined #ipfs
geir_ has quit [Quit: SafeBouncer - znc.safenetwork.org:4876]
museless has joined #ipfs
m0ns00n has quit [Quit: quit]
geir_ has joined #ipfs
IlanGodik has joined #ipfs
onikuma has quit [Read error: Connection reset by peer]
joocain2 has quit [Write error: Connection reset by peer]
neosloth has quit [Read error: Connection reset by peer]
neosloth has joined #ipfs
anthk_ has quit [Remote host closed the connection]
clemo has quit [Remote host closed the connection]
clemo has joined #ipfs
anthk_ has joined #ipfs
manveru has quit [Ping timeout: 255 seconds]
lemmi has quit [Ping timeout: 240 seconds]
robogoat has quit [Ping timeout: 240 seconds]
[0__0] has quit [Remote host closed the connection]
Xe has quit [Ping timeout: 240 seconds]
gorhgorh has quit [Quit: No Ping reply in 180 seconds.]
pepesza has quit [Ping timeout: 240 seconds]
[0__0] has joined #ipfs
Vaelatern has quit [Ping timeout: 240 seconds]
engdesart has quit [Ping timeout: 240 seconds]
shakalaka has quit [Ping timeout: 240 seconds]
joocain2 has joined #ipfs
gorhgorh has joined #ipfs
ion has quit [Ping timeout: 240 seconds]
fredmanglis has quit [Ping timeout: 255 seconds]
manveru has joined #ipfs
basilgohar has quit [Ping timeout: 260 seconds]
fredmanglis has joined #ipfs
basilgohar has joined #ipfs
cypher has quit [Ping timeout: 255 seconds]
ion has joined #ipfs
eerne has quit [Ping timeout: 248 seconds]
victorSN has quit [Ping timeout: 240 seconds]
pepesza has joined #ipfs
Xe has joined #ipfs
erictapen has joined #ipfs
erictapen has quit [Remote host closed the connection]
erictapen has joined #ipfs
shakalaka has joined #ipfs
lemmi has joined #ipfs
Vaelatern has joined #ipfs
cypher has joined #ipfs
eerne has joined #ipfs
maxlath has joined #ipfs
victorSN has joined #ipfs
igorline has joined #ipfs
Alpha64 has quit [Read error: Connection reset by peer]
eerne has quit [Ping timeout: 264 seconds]
robogoat has joined #ipfs
maxlath has quit [Ping timeout: 240 seconds]
testimonlkjd[m] has joined #ipfs
echoSMILE has quit [Ping timeout: 240 seconds]
xnbya has quit [Ping timeout: 240 seconds]
xnbya has joined #ipfs
erictapen has quit [Ping timeout: 240 seconds]
neosloth has quit [Read error: Connection reset by peer]
_whitelogger has joined #ipfs
erictapen has joined #ipfs
ZaZ has joined #ipfs
m0ns00n has joined #ipfs
eerne has joined #ipfs
eerne has quit [Ping timeout: 260 seconds]
yosafbridge has quit [Quit: Leaving]
erde74 has joined #ipfs
oed has quit [Quit: WeeChat 1.6]
joelto has joined #ipfs
joelto is now known as oed
yosafbridge has joined #ipfs
maxlath has joined #ipfs
maxlath has quit [Ping timeout: 240 seconds]
maxlath has joined #ipfs
<lgierth>
anybody here got pointers to implementing varints in hardware?
<lgierth>
i imagine it's quite tricky
<lgierth>
my thinking is the trickier it is the worse an idea it is to use varints at the lowest layer of a packet-based wire protocol
<lgierth>
Kubuxu: ^
<lgierth>
suggestions for search terms are appreciated too :P most of the stuff i'm finding is about variational integrators
maxlath has quit [Ping timeout: 260 seconds]
maxlath has joined #ipfs
reit has quit [Quit: Leaving]
daguiam has quit [Quit: Connection closed for inactivity]
ganlub has joined #ipfs
<ganlub>
Hi guys, I'm new with ipfs and I have a question, so If run my daemon/put some content in there and then I stop it, will this content ever expire?
<lgierth>
ganlub: if no other node has fetched it, it'll be unavailable as soon as you turn your node off
<lgierth>
there's no automatic storage by other nodes
<lgierth>
i.e. ipfs doesn't "upload" stuff to the network
<lgierth>
it's gonna be available as long as there's at least one node online which has it
<lgierth>
yeah the node behind ipfs.io has it fetched
<lgierth>
it'll keep it for a couple of days until it runs garbage collection
<lgierth>
ah that actually looks like ipfs's default init files. every node has these
<ganlub>
I see, okay thanks!
<ganlub>
lgierth: ah how is ipfs helpfull with censorship?
<ganlub>
let's say I put something online, point certain domain, and they want me to put it down, how is people able to create point other domains to it, etc
<lgierth>
data isn't addresses by a domain name, but by it's cryptographic hash, so tons of nodes can store the same data and all help keep it available
maxlath has quit [Ping timeout: 240 seconds]
<lgierth>
(you can use domain names as a pointer to a hash though)
<lgierth>
and if the domain goes down for whatever reason, you can still access the data by its hash, or point new domain names, etc.
<ganlub>
I understand
<ganlub>
but I should always keep the daemon on right?
<lgierth>
yeah - or find someone else who's willing to keep your content available
<lgierth>
e.g. filecoin will create a token-backed network to pay others to keep your content available
<ganlub>
I see, and there's nothing wrong on keeping it up, you aint violating any lay with that
<lgierth>
or you could get together with your friends and create an informal storage syndicate
<lgierth>
weeelll that certainly depends on the kind of data
coderdecoder01 has quit [Quit: Connection closed for inactivity]
<lgierth>
if you make illegal data available then that's still illegal of course
<ganlub>
I've know about ipfs with all it's going on in Catalonia, the gov hosted some relevant pages here so they can't be taken down
<ganlub>
yes sure
<lgierth>
yeah we heard of it, the pirate party put a mirror of the referendum website on ipfs
<lgierth>
and the .cat tld got their offices raided :/ fucked up
<ganlub>
so fucked up... let's see how it ends
<lgierth>
we actually have that referendum.pirata.cat data pinned on a few of our ipfs.io nodes -- so it'll stay available no matter if there's nodes in catalunia seeding it
<ganlub>
lgierth: thanks for the answers dude!
<lgierth>
and we'll be happy to pin more endangered .cat websites
<lgierth>
(pinning just means storing that data AND not having it evicted by regular garbage collection)
<lgierth>
you're welcome, keep it up
<ganlub>
so by pinning you have acces to the soruces?
<lgierth>
sources?
<ganlub>
in this cases, the soruces of the web
<DuClare>
Fwiw freenet allows uploads into the network, if you're after something like that. But I fear it's nowhere near as accessible as ipfs for the general public.
<lgierth>
the .html/.css/.js files? yeah that's what the dataset is made up of
<ganlub>
I'm saying this, bc I heard on the media, that all of this replicas, could be tempering the forms on the page, etcetera
<lgierth>
that'll just redirect to ipfs.io/ipfs/thehash
<ganlub>
Ah okay, so you can;t have many domains to point the same gateway
<lgierth>
you can, pirate party just did the most simple thing
<ganlub>
unles they are specified on the file inside your node
<lgierth>
you can point a CNAME to gateway.ipfs.io, and set the TXT dnslink record
<ganlub>
okay trying that!
<lgierth>
without the CNAME, it'll be accessible as ipfs.io/ipfs/example.com, and with the CNAME you can actually do http://example.com
<lgierth>
similar to github pages
<lgierth>
depending on your dns provider you might not be able to set CNAME on the root of your domain. do A then, with 147.135.130.181 and 217.182.195.23
<ganlub>
lets say CNAME referendum.example.com to gateway.ipfs.io
<ganlub>
right?
<ganlub>
and then the TXT referendum.example.com with value `dnslink=/ipfs/$actual_hash`
<Kubuxu>
lgierth: yes it is generally bad idea to use varints if you plan implementing it in hardware
<lgierth>
better make it _dnslink.referenedum.example.com
<lgierth>
Kubuxu: yeah :/ hrm.
<Kubuxu>
because you have to first parse the varint to be able to access what is behind it
<lgierth>
on the lowest layer we could always pad it to 4 bytes
<lgierth>
hrm
<ganlub>
okay!!!
<grawity>
as a side note about gateways, does ipfs.io/ipns/* cache the lookup results?
<grawity>
oh looks like it finally does
<lgierth>
it caches IPNS lookup results, but not dnslink results
<grawity>
somehow it was taking a long time to resolve the ipns
<lgierth>
note that ipfs.io is more than host so you might be hitting different hosts too
<lgierth>
yeah IPNS is fucking slow
<grawity>
just asking because it was noticeably slower than gateway.ipfs.io with --header "Host: …"
<lgierth>
i.e. IPNS itself -- IPNS with a domain name (= dnslink) should be pretty quick
<grawity>
could be just hitting different hosts...
<lgierth>
set -v on whatever tool you use there :) it's currently just DNS round-robin so it's actually visible to clients which host gets picked
<lgierth>
oh right. need a dot at the end of the CNAME value
<grawity>
you typo'd "referenedum" in the _dnslink
<lgierth>
domain names technically always end in a dot
<lgierth>
hehe yeah that too
<grawity>
the CNAME is fine, I think your DNS panel just assumed it's absolute name anyway
<ganlub>
I looks likt it's the typo... omg
<grawity>
but yeah, in the general case it's better to add the dot to be on the safe side.
<lgierth>
now wait 15 minutes ;)
jonnycrunch has joined #ipfs
<grawity>
hmm
<grawity>
is it recommended to point dnslinks to /ipfs or to /ipns?
<ganlub>
Oh I never added the do at the end for my other subdomains
<lgierth>
grawity: whatever your use case is. with /ipfs you'll have to update it everytime the site changes
<lgierth>
but /ipns is slow
<ganlub>
lgierth grawity thanks guys!
<lgierth>
it'll be faster sometime soon but right now it's slow
<lgierth>
ganlub: you're welcome!
<grawity>
hmm
<grawity>
yeah I guess /ipfs is a more direct approach, if one wants to rely on DNS anyway...
<lgierth>
grawity: note dnslinks resolve rescursively, so you can point one dnslink at another dnslink. that's nice when you want to give control over a dnslink to someone who shouldn't have control over the rest of the domain
jaboja has joined #ipfs
<lgierth>
e.g. protocol.ai points to website.protocol.ai, so the people being allowed to update the website don't have access to e.g. the email records
<lgierth>
grawity: and yeah word, dnslinks are not the last shot on naming ;)
<lgierth>
dnslinks are just a simple way of integrating with the largest and most well understood naming system that currently exists (DNS)
<lgierth>
there are plans e.g. for integration with namecoin, ethereum name system, gnunet name system, etc.
<lgierth>
there's also another somewhat similar approach called dns-over-https
<lgierth>
but that relies on TLS CAs, and running a server (in addition to the DNS system), so that while it's definitely more secure than dnslink, it relies on three centralized/hierarchical things while dnslink relies on only one (DNS)
<Magik6k>
Apparently there is a tool called ipfs on freebsd - 'ipfs - saves and restores information for NAT and state tables.'
<lgierth>
"IP Firewall State" :)
<lgierth>
yeah
<lgierth>
that's why the binary for freebsd is called ipfs-go i think
<Bat`O_>
hey, is there already some public relay node available to play with ?
<lgierth>
Bat`O_: not yet -- but very soon
<lgierth>
you'll have to try your own until then :)
<lgierth>
almost everybody is on vacation at the moment
<Bat`O_>
lgierth: thanks :)
<Bat`O_>
i'm not sure i got it right though ... my problem is that i have two node, each behind a NAT, that have to talk to each other (ipfs, ipns and pubsub) but at the moment it does works that well. Would having a relay configured on each node help in this situation ? does it have to be the same relay ?
fredthomsen has joined #ipfs
jaboja has quit [Ping timeout: 248 seconds]
bedeho has quit [Remote host closed the connection]
<lgierth>
Bat`O_: yes it'd have to be the same relay node
<lgierth>
there's support for chains of relays in the spec but i'm not sure that's already implemented
<Bat`O_>
does I have to manually connect the two nodes ?
jaboja has joined #ipfs
<lgierth>
es
<lgierth>
yes
<Bat`O_>
or put another way, does configuring a relay helps for the general connectivity of nodes behind nat ?
<lgierth>
yes to that too :)
<lgierth>
tricky NAT situations and generally situations where nodes can't properly reach each other, are the use case for relay
bedeho has joined #ipfs
<Bat`O_>
what can you do if you don't know the other node ID ?
<lgierth>
then you need to find it out somehow :)
<Bat`O_>
:-|
<Bat`O_>
well that doesn't help me much then ...
<ansuz>
you could try using cjdns
<lgierth>
yeah that can help with NAT too
<Bat`O_>
let me explain in more details, this is for my project (https://github.com/MichaelMure/Arbore). In this project each user publish his identity on ipfs/ipns and build a contact list by resolving ipns entries
bedeho has quit [Remote host closed the connection]
<Bat`O_>
then file sharing happen with ipfs and general communication over pubsub
<Bat`O_>
what i found is that often, ipns doesn't resolve at all because both nodes are behind a NAT and the connectivity is not good
<lgierth>
ah so you do have the PeerIDs of each
<Bat`O_>
not the peerID, but the public key of each
<lgierth>
and if all of them are connected to the same relay, you can do `ipfs swarm connect /ipfs/QmTheRelayNode/p2p-circuit/ipfs/QmTheRemoteNode^
<lgierth>
the PeerID is derived from the public key
<lgierth>
it's basically a sha256 hash of the raw rsa pub key
<Bat`O_>
it's a new key generated for each profile
<Bat`O_>
you can have multiple identities and move identities on another node
<Bat`O_>
lgierth: you are talking about the 'self' key ?
<lgierth>
ah yeah what i said applies to the self key -- you can't connect using other keys
<Bat`O_>
yes ...
<lgierth>
these other keys can currently only be used for IPNS, but not for the underlying connection
<lgierth>
anyhow, if they're all connected to the same relay node (or set of relay nodes), they could use pubsub to broadcast their identity+peerID
<lgierth>
and you can use the peerID to connect through the relay
<lgierth>
*or* maybe you don't even need direct connections? maybe pubsub is enough?
<lgierth>
ok gotta go, the beach is calling :):)
<Bat`O_>
... maybe i can query the profile over pubsub, but i lose the garantee that someone is not publishing fake profile ...
ZaZ has quit [Read error: Connection reset by peer]
fredthomsen has quit [Remote host closed the connection]
Bhootrk_ has joined #ipfs
Bhootrk_ has quit [Max SendQ exceeded]
fredthomsen has joined #ipfs
fredthomsen has quit [Remote host closed the connection]
<grawity>
lgierth: hmm how is that "recursive" _dnslink.protocol.ai -> _dnslink.website.protocol.ai redirect made?