stebalien changed the topic of #ipfs to: Heads Up: To talk, you need to register your nick! Announcements: go-ipfs 0.4.18 and js-ipfs 0.34 are out! Get them from dist.ipfs.io and npm respectively! | Also: #libp2p #ipfs-cluster #filecoin #ipfs-dev | IPFS, the InterPlanetary FileSystem: https://github.com/ipfs/ipfs | Logs: https://view.matrix.org/room/!yhqiEdqNjyPbxtUjzm:matrix.org/ | Forums: https://discuss.ipfs.io | Code of Con
WhizzWr has quit [Quit: Bye!]
WhizzWr has joined #ipfs
pecastro has quit [Read error: Connection reset by peer]
<Lostfile[m]1>
If only somebody were to make a Web browser that fully used or took advantage of ipfs I mean the beaker browser did at 1 point but eventually they dropped support for some reason from my understanding
<Lostfile[m]1>
And I mean like a solid Web browser
jesse22 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
woss_io has quit [Ping timeout: 240 seconds]
cyfex_ has joined #ipfs
cyfex has quit [Ping timeout: 250 seconds]
i1nfusion has quit [Remote host closed the connection]
i1nfusion has joined #ipfs
appa_ has joined #ipfs
dimitarvp has quit [Quit: Bye]
gniux has joined #ipfs
gniux has quit [Quit: WeeChat 1.9]
<postables[m]>
whats stopping you from doing just that 😮
cyfex has joined #ipfs
cyfex_ has quit [Ping timeout: 250 seconds]
Belkaar has quit [Read error: Connection reset by peer]
Belkaar has joined #ipfs
Belkaar has joined #ipfs
shizy has quit [Quit: WeeChat 2.3]
user_51 has quit [Ping timeout: 246 seconds]
}ls{ has quit [Ping timeout: 240 seconds]
xcm has quit [Remote host closed the connection]
user_51 has joined #ipfs
reit has quit [Ping timeout: 240 seconds]
xcm has joined #ipfs
}ls{ has joined #ipfs
i1nfusion has quit [Remote host closed the connection]
i1nfusion has joined #ipfs
cyfex_ has joined #ipfs
cyfex has quit [Ping timeout: 246 seconds]
xcm has quit [Remote host closed the connection]
xcm has joined #ipfs
i1nfusion has quit [Remote host closed the connection]
i1nfusion has joined #ipfs
clemo has joined #ipfs
clemo has quit [Client Quit]
}ls{ has quit [Quit: real life interrupt]
postables has joined #ipfs
Xenguy has quit [Ping timeout: 240 seconds]
lindeb has quit [Ping timeout: 240 seconds]
clemo has joined #ipfs
Xenguy has joined #ipfs
_whitelogger has joined #ipfs
kesenai has joined #ipfs
random_yanek has quit [Ping timeout: 245 seconds]
random_yanek has joined #ipfs
Xenguy has quit [Ping timeout: 250 seconds]
Xenguy has joined #ipfs
q6AA4FD has joined #ipfs
spinza has quit [Quit: Coyote finally caught up with me...]
<str1ngs>
I have used go-ipfs-api in the past, but unfortunately that still requires an online ipfs api instance.
<str1ngs>
I have also looked at github.com/multiformats/go-multihash. but in order to get a ipfs compatible hash I need to marshall the data first to a UnixFS and then to a Dag Node. It seems you can do this using JS libraries. But I can't figure out how to do this in go.
Fessus has joined #ipfs
seba-- has quit [Changing host]
seba-- has joined #ipfs
<seba-->
verify in what sense?
<str1ngs>
I need to reproduce a ipfs compatible multihash so it would be the same as if I did ipfs add ./file_foo --only-hash
<seba-->
why you don't use that? it doesn't need to be online afaik
<str1ngs>
because it would require the end user to have ipfs installed and it would also require me to exec a command from my code. I need to do this is in a programmable way and not require an online instance of ipfs
<seba-->
well you could use js-ipfs?
<str1ngs>
no my project is written in golang, javascript is not a dependency I would like to use there is no need to mix languages
<seba-->
i see, then just use the library in go
<seba-->
:)
<str1ngs>
I already have code that replicates ipfs add --hash-only by using core and coreunix. though that's not what I consider to be a "go library" :)
<seba-->
hm
<seba-->
what do you want then?
<seba-->
not sure
<str1ngs>
I'm looking for away to avoid pulling in ipfs core if possible by using a library that allows me to create a UnixFS then a ipfs Dag which I can then multihash
<seba-->
ah
<seba-->
good luck
<str1ngs>
ahh also ipfs uses gx so it then makes my code unable to go get
<str1ngs>
that was one of the biggest issues to using core and coreunix that I'm having
dasj19 has joined #ipfs
dimitarvp has joined #ipfs
zeden has joined #ipfs
zeden has quit [Client Quit]
Encrypt has quit [Quit: Quit]
<str1ngs>
another question I have does ipfs dedup blocks. ie if two files had the slam bock would it only store one version of the block?
<str1ngs>
s/slam/same/
riemann has joined #ipfs
<lordcirth>
str1ngs, if a object hashes the same as another, they are the same object
<lordcirth>
And therefore will only be stored once per node
<lordcirth>
However, if you are changing a file, odds are the change won't align with the 256kb block size as well as you'd like
<str1ngs>
ah so odds are blocks will not be shared between files? if I'm reading this right?
<lordcirth>
str1ngs, if the file is modified to be slightly longer or shorter, than the blocks after the change won't match
<lordcirth>
Note that the 256kb blocks are just a default; the whitepaper mentions that certain applications might benefit from content-aware block sizes
<str1ngs>
its more an interesting though I was having. I'm not truly grasping node and blocks quite yet.
tombusby has quit [Remote host closed the connection]
<lordcirth>
str1ngs, files larger than 256kb are split using root blocks that work much like directories
<str1ngs>
I guess that is what the concept of a node is?
<str1ngs>
root -> child etc
<lordcirth>
er, when I said node above, I was referring to 1 instance of an IPFS daemon
<str1ngs>
also I guess that is what the merkletree is based off ?
<lordcirth>
perhaps that was a poor choice of wording
hurikhan77 has joined #ipfs
<str1ngs>
ahh sorry I understand what you meant by node now
<str1ngs>
I should use term merkledag not tree
<lordcirth>
Technically it's a Merkle DAG, not tree, because each merkle node (talking about the structure now, not daemons) can store both data and links, yeah
<str1ngs>
it's interesting kinda like bitcoin meets GitterIntegratio
<str1ngs>
err git *
<lordcirth>
It's a merger of all sorts of things
recursive has quit [Ping timeout: 268 seconds]
<str1ngs>
currently I'm leveraging ipfs to distribute and verify compiled packages. in a systems package manager I've written. this allows me to avoid using so called http mirrors and central package repositories
recursive has joined #ipfs
<lordcirth>
str1ngs, ah, neat. Wrote your own pkg manager? Why not, eg, Nix?
<str1ngs>
nix is nice same with guix. but this package manager actually pre dates both of those
<str1ngs>
I think eventually I can add features that nix and guix provide but using ipfs instead
<str1ngs>
I lean towards guix over nix because most if what deal with is GNU related so kinda biased there
<lordcirth>
str1ngs, I don't know how far along your program is, but I think we need less duplication of effort
lidel` has joined #ipfs
<str1ngs>
maybe, but guix which I know the most about. is kinda loftly and slow for my taste. it also requires root privileges in order to setup and use and build. my package meta data is also machine readable and writable. making for some interesting use cases
<str1ngs>
don't get my wrong guix and nix are nice way more feature rich. but I think it kinda miss the point in some regards
lidel has quit [Ping timeout: 245 seconds]
lidel` is now known as lidel
saki has joined #ipfs
saki has quit [Excess Flood]
woss_io has joined #ipfs
saki has joined #ipfs
kiera[m] has left #ipfs ["User left"]
kaminishi has quit [Remote host closed the connection]
recursive has quit [Read error: Connection reset by peer]
<aschmahmann[m]>
anyone know how I can publish that an arbitrary multiaddr has a CID? For example, can I publish that a block is available at /http/example.com/exampleBlock?
woss_io has quit [Ping timeout: 250 seconds]
l2d has quit [Remote host closed the connection]
<lordcirth>
aschmahmann[m], not sure what you mean
<lordcirth>
aschmahmann[m], you could have a directory that you regularly update with all blocks, and use IPNS to version it
<aschmahmann[m]>
lordcirth: I may have misunderstood, but I thought that when you publish to the DHT with IPFS you publish the statement that is roughly "I have the content with CID X, ask me for it"
<lordcirth>
aschmahmann[m], ah, yes, that's essentially what happens
<lordcirth>
aschmahmann[m], so what do you want to do?
<aschmahmann[m]>
lordcirth: is the value corresponding to key = cid a peerID or multiaddr?
<lordcirth>
I believe it's a PeerID, let me check the whitepaper
<aschmahmann[m]>
lordcirth: I'd like to publish "the content with CID X is at /http/example.com/exampleBlock"
<lordcirth>
aschmahmann[m], over http?
<lordcirth>
You want IPFS lookups to redirect to outside IPFS?
<aschmahmann[m]>
lordcirth: ideally. I mean if it's a multiaddr instead of a peerID I should be able to do whatever.
<aschmahmann[m]>
well they do anyway, they currently use some libp2p transports, but data could be available anywhere. Some people could be hosting themseleves, but others might host on platforms that only support http/s
<lordcirth>
"For values larger, the DHT stores references, which are the NodeIds of peers who can serve the block."
<aschmahmann[m]>
hmmm... oh well I guess, thx for the help.
<lordcirth>
aschmahmann[m], if this is going to be viewed in a browser, you could upload an IPFS file which consists of <html> <body> <meta http-equiv="refresh" content="0; url=new"> </body> </html>
dsiypl4 has quit [Ping timeout: 246 seconds]
<lordcirth>
Where "new" is the link you want
<lordcirth>
Thing is, IPFS is designed to integrate with any content-addressable datastore - but HTTP isn't content-addressable. Allowing IPFS hashes to resolve to HTTP links would break a number of fundamental properties.
shizy has joined #ipfs
<aschmahmann[m]>
lordcirth: I feel like IPFS is the link layer that ties hash(content) to content location. Whether you go to a libp2p node or an http server should be essentially the same. You'll check on receipt of the block if the hashes match and if they don't you'll kill the connection.
<lordcirth>
aschmahmann[m], that's true, it ought to be possible to do securely. But it would still be centralized and have a single point of failure.
Xenguy has quit [Ping timeout: 246 seconds]
<aschmahmann[m]>
lordcirth: it wouldn't any more than it currently is. if currently only one libp2p node is hosting your content and it goes offline you have a single point of failure. what makes it not single point of failure is that the DHT can resolve hash(content) to multiple locations (e.g. 5 nodes, or 1 node and 1 http server). seems reasonable that one of them be allowed to use an http transport.
<aschmahmann[m]>
lordcirth: what I was thinking of was more like instead of worrying about running an ipfs-cluster on AWS I could potentially run a pretty small advertising daemon and store the blocks on S3.
<lordcirth>
aschmahmann[m], and this would be easier because you are already needing to store them on S3?
Xenguy has joined #ipfs
aeroplain[m] has joined #ipfs
<aschmahmann[m]>
lordcirth: yes and because there's less hassle involved in running a small node + S3 than managing a cluster. There are also some S3-like services that charge very little for data egress which feels optimal for the IPFS case, but don't have accompanying EC2-like services to run nodes on.
<lordcirth>
aschmahmann[m], how much data do you have? Just curious
<aschmahmann[m]>
lordcirth: I'm not sure. right now mostly focused on what will be easy + inexpensive to run.
<aschmahmann[m]>
probably not a ton of data, but enough I don't want a server in my house :)
<lordcirth>
aschmahmann[m], would the data be accessed through the browser?
<aschmahmann[m]>
lordcirth: not required
spinza has quit [Quit: Coyote finally caught up with me...]
<aschmahmann[m]>
I'd like it to be natively accessible via IPFS ideally so I can utilize other IPFS tools
<aschmahmann[m]>
lordcirth: my understanding is that helps, but still requires my node to proxy all the traffic.
<lordcirth>
Yes, I believe so
<lordcirth>
That's the only way to get what you want without patching IPFS, I think
<Mikaela>
Can I somehow disable IPv4 peers? I am behind two NATs and I think IPFS is needlessly overwhelming my router. I have disabled IPv4 swarm addresses and bootstrap peers (including /dnsaddr/), but most of my peers are still IPv4 and well above HighWater.
<aschmahmann[m]>
lordcirth: ya, that was my concern. thanks for your help, I'll think about it some more
<lordcirth>
I suspect there's a better way to solve your actual problem, though
<aschmahmann[m]>
lordcirth: perhaps, I'm just noticing that all of the ipfs pinning services I can find charge like 10x what S3 charges and I am guessing a lot of that is uncertainty from egress costs /- the costs of running a bunch of compute nodes. However, I don't know how many of them are actually using ipfs-cluster so that might be a problem in and of itself
dsiypl4 has joined #ipfs
alexgr has quit [Quit: Leaving]
<lordcirth>
Mikaela, good question. I took a look, it doesn't seem like there's an option.
<Mikaela>
Thanks, I guess I should open an issue?
chris613 has joined #ipfs
<lordcirth>
Mikaela, probably, yeah. In the meantime, you could firewall IPFS from making outbound ipv4 connections?
<Mikaela>
I don't know how would I do that and it may be too advanced for ufw
<lordcirth>
Not a fan of ufw, but it shouldn't be hard
spinza has joined #ipfs
<lordcirth>
Mikaela, btw, gufw is a GUI for it
<Mikaela>
I think ufw directly is simple enough for my purpouses and last time I checked the GUI was missing options such as allow/deny from. However I think ufw works more on ports/services and there isn't one well known IPFS port I could block, is there?
Encrypt has joined #ipfs
<lordcirth>
Mikaela, 4001
<lordcirth>
If you block all IPv4 to dest port 4001, it ought to work
<lordcirth>
People can run other ports, but I doubt many do
random_yanek has quit [Quit: random_yanek]
<Mikaela>
oh, I must have been thinking of some other service which randomizes port and tells in examples to pick a random port instead of using any example. Thanks
alexgr has joined #ipfs
random_yanek has joined #ipfs
random_yanek has quit [Max SendQ exceeded]
random_yanek has joined #ipfs
<Mikaela>
at first it worked, but then I attempted deliberately using it by going to ipfs.io and ipfs.pics and now I again have 60 peers mostly IPv4 :D
dsiypl4_ has joined #ipfs
saki has quit [Quit: saki]
dsiypl4 has quit [Ping timeout: 250 seconds]
MatMaul[m] has joined #ipfs
piti has quit [Ping timeout: 268 seconds]
i1nfusion has quit [Remote host closed the connection]
i1nfusion has quit [Remote host closed the connection]
i1nfusion has joined #ipfs
bailon has joined #ipfs
hurikhan77 has joined #ipfs
<SchrodingersScat>
did js-ipfs get really bad over the last couple days or is it just me? If it's a matter of setting up my own preload node thingy then is there instructions for that?
Encrypt has quit [Quit: Quit]
<shoku[m]>
has anyone solved making user feeds and whitelists (friends) in ipfs?
Encrypt has joined #ipfs
<SchrodingersScat>
pubsub?
<SchrodingersScat>
oh, is my solution really just preload: { enabled: true, addresses: [...] }
alexgr has quit [Remote host closed the connection]
hphs^ has joined #ipfs
<lordcirth>
shoku[m], could you be more specific?
<lordcirth>
shoku[m], do you mean allowing some list of specified people to send you stuff?
<shoku[m]>
I'm considering making epona more about citizen journalism
SunflowerSociety has joined #ipfs
<shoku[m]>
and one thing that'd help would be the ability of making a friends list that can see my content
<shoku[m]>
so I can whitelist people
<shoku[m]>
more like me sending to specific people
<lordcirth>
shoku[m], ah, ok. Well, you'd need a form of encryption on top of your IPFS content.
<shoku[m]>
but then how can I tell people there's new content
<shoku[m]>
eg can they subscribe to my stuff easily
<lordcirth>
shoku[m], IPNS
<shoku[m]>
so let's say I make a self generated site
<shoku[m]>
and I put new content there and it updates the page automatically
<shoku[m]>
and I auto update my page on ipns
<lordcirth>
shoku[m], you could encrypt each release of your data with a one-time AES-256 key. Update your IPNS link every time you have a new release. The IPNS resolves to a directory containing the data and a copy of the master key encrypted to each friend's key
<shoku[m]>
how would people know my page is updated, and how would I make it only some people would know that, or just give some people access to the content
Sunflowery has quit [Ping timeout: 244 seconds]
<lordcirth>
shoku[m], 2 ways. 1, they just keep resolving your IPNS link. 2, pubsub
<lordcirth>
You can't prevent the public from seeing that you've issued a new version of your page. But you can encrypt it
<shoku[m]>
but like I can't put a pw on the page I guess. It'd be on the content
<shoku[m]>
kinda like putting a password on a rar file
arthur has quit [Quit: leaving]
arthur has joined #ipfs
<shoku[m]>
I guess I'm curious if someone made a facebook type of feed where I could give people permissions to see my stuff
<senden9[m]>
<shoku[m] "I guess I'm curious if someone m"> I know many sites (like [Mastodon](https://joinmastodon.org/)) but no one with encryption 🤔
<shoku[m]>
Did you just make that? thanks
<lordcirth>
Now, this implies 0 dedup between versions of your blog, which is expensive. You could encrypt individual files, and only change keys on the ones that changed, but that does leak metadata
<senden9[m]>
<shoku[m] "Did you just make that? thanks"> No. I am just a user there
<lordcirth>
I think he's referring to my gist. Yes I did just think that up after you asked
<lordcirth>
In between respawns XD
<lordcirth>
In short, IPFS will do its job, getting your data from A to B, with authentication, and IPNS will let you version it easily, but you'll need to implement your own access control framework. Not that it's complicated.
<lordcirth>
Actually making a decent UX would probably take some work
<shoku[m]>
Please do have a look and let me know what you think!
<f0x2>
will pinning an ipns automatically update whenever it's block is changed?
<f0x2>
oh Epona is windows-only, so can't take a look
<shoku>
you on linux?
<f0x2>
yeah
<lordcirth__>
I would guess that most IPFS users are on Linux
<shoku[m]>
For now 😃
<shoku[m]>
I see that as a failure in UX tbh
<shoku[m]>
I mean less than 1% of the consumer market is on linux
<shoku[m]>
so that means we don't have solutions for 99% of people
<shoku[m]>
that, to me, means that we're only good at service a minuscule (but important) niche, which seems very much unlike the aspirations of ipfs
<shoku[m]>
hence why we started with windows
<lordcirth__>
shoku[m], getting people to use a robust, decentralized network while running a centrally-controlled OS seems of little use
<f0x2>
doing an ipfn pin add seems to take reaally long :/
<shoku[m]>
you gotta walk the horse to the water!
<f0x2>
also go-ipfs will probably compile fine on/for windows
<lordcirth__>
fair enough. I certainly don't think it's counter-productive.
woss_io has joined #ipfs
random_yanek has joined #ipfs
aeroplain[m] has left #ipfs ["User left"]
aeroplain[m] has joined #ipfs
q6AA4FD has joined #ipfs
sakalli_ has joined #ipfs
sakalli_ has quit [Client Quit]
user84962004[m] has left #ipfs ["User left"]
i1nfusion has quit [Remote host closed the connection]
<str1ngs>
or just dump a bucket of water on the horse
<str1ngs>
I know ipfs works with WSL
clemo has quit [Ping timeout: 245 seconds]
clemo has joined #ipfs
windowpane[m] has joined #ipfs
zerogue has quit [Quit: Goodbye Cruel World]
windowpane[m] has left #ipfs ["User left"]
random_yanek has quit [Ping timeout: 244 seconds]
daMaestro has joined #ipfs
zerogue has joined #ipfs
Ai9zO5AP has quit [Quit: WeeChat 2.3]
random_yanek has joined #ipfs
<f0x2>
seems I got my first site working :) both with a webproxy at testsite.foksipfs.cf and /ipns/testsite.foksipfs.cf
lindeb has joined #ipfs
charlton has joined #ipfs
jesse22 has joined #ipfs
plexigras has quit [Ping timeout: 250 seconds]
<postables[m]>
aschmahmann: my pinning service is still more expensive than S3 but it's cheaper than all known pinning services, and we use IPFS cluster with three nodes, 2 in one location 1 in an off-site location
<postables[m]>
Mikaela: in Swarm.Addresses (might be wrong config directive) remove the /ip4/.... part and you'll disable ipv4 if you remove the bootsrap peers you're not disabling ipv4