>What is IPFS?
It's basically BitTorrent on steroids.
>why would one use it
* Distributed, decentralized filesharing (for now - ipfs is merely the communication protocol, and additional applications can exist on top of it).
* You can have a mutable address (i.e. always points to the latest version of a site), or a static address (points to a specific file). Yes, you can host sites over IPFS.
* Peers are found fast for new downloads. You don't need to wait that much to start a download.
* You can watch your animu while it downloads, I watched few episodes that way and it didn't even buffer.
>how to upload a single file
$ ipfs add ./$file
Access it at localhost:8080/ipfs/$outputted-hash
>how to upload a dir
$ ipfs add -r ./$dir
Access it at localhost:8080/ipfs/$last-outputted-hash
>how to make the thing mutable
$ ipfs name publish ./$file-or-dir-hash
Access it at localhost:8080/ipns/$output-hash-aka-peerid (it's ipNs not ipFs)
To update, publish another hash and it will be available at the same IPNS address.
>gateways (how to access IPFS if you don't have it installed)
https://gateway.ipfs.io/
>most recent talk about it by the dev
https://www.youtube.com/watch?v=HUVmypx9HGI
>I2P and Tor support coming soon™. We need that thing anonymous so pls halp.
https://github.com/ipfs/go-ipfs/issues/1118
https://github.com/ipfs/notes/issues/37
https://github.com/jbenet/go-multiaddr-net/issues/7
Daily reminder to pin files that you care about.
Some links in next post.
Some 8chen anon's ipfs website, has some info, links and files (breaking up the links to avoid the filter):
/ipfs/QmP7LM9yH
gVivJoUs48oqe2bmMbaYccGU
cadhq8ptZFpcD/
Some links from that selection:
All atari games, over IPFS!
/ipfs/QmbKxNNCxBox7
Cmv3jiUZbiG3zpzmtnYz
VUuKHxfAjvpyH/
do/k/ument:
/ipfs/QmSTEUc53
GVQhWTCRc5wZpdhQjDzpgW
Hv56GaNBheqKBHG/
/g/entoomen library:
/ipfs/QmTmMhRv2nh
889JfYBWXdxSvNS6zWnh
4QFo4Q2knV7Ei2B/
First season of SAC:
/ipfs/QmVBEScm197eQ
iqgUpstf9baFAaEnhQC
gzHKiXnkCoED2c/
He's a big file:
/ipfs/QmfWQHVazH6so9
p27z27rr8TJSdBFGpH7hu
nDcaZ1EAQ2c/The.Dark.Knight.Rises.2012.720p.BluRay.x264.YIFY.mp4
>>51444075
pretty cool
>>51444075
Is it anonymous like GNUnet?
This won't work unless people have fast internet connections with no data caps. It might be good to go in 5-10 if burgerland can get it's shit together and give people unmetered internet connections.
Reposting this
http://blog.neocities.org/its-time-for-the-permanent-web.html
>>51444195
No. It expects you to use tor or i2p (for example) if you want anonymity on top of it.
>>51444075* Distributed, decentralized filesharing (for now - ipfs is merely the communication protocol, and additional applications can exist on top of it).
* You can have a mutable address (i.e. always points to the latest version of a site), or a static address (points to a specific file). Yes, you can host sites over IPFS.
* Peers are found fast for new downloads. You don't need to wait that much to start a download.
* You can watch your animu while it downloads, I watched few episodes that way and it didn't even buffer.
You can do all of that with bit torrent.
>>51444203
At minimum it requires exactly the same setup as in HTTP. The difference is that peers can pin content, becoming an explicit provider for it, and every other node will cache accessed content for a while. The load on average will be insignificant for consumers and nonexistent for providers of popular content.
>>51444250
Show me a website that runs on top of bittorrent. Protip: you can't.
>>51444300
Any website that hosts the HTML of their pages in torrents will do
>>51444325
Why can't you provide me an example?
>>51444357
Give me a few minutes to set one up and I will
>>51444203
It will work regardless, the worst case scenario is that this is HTTP with some benefits, which is still an improvement. Even with only 1 peer acting in the traditional client server model you still get permanent content links and better content resolution + traversal than just using domains with a DNS which is a pretty big selling point imo.
People always say this when it comes to P2P systems but they forget that modern P2P systems segment things so that swarming is possible, it's not like old P2P systems where it was literally 1 peer to another, it's several peers to 1, even if every peer contributes a small amount, combined they are fast, that concept alone is a big reason that bittorrent is so good. A benefit of distributed systems is that they're resilient too, so long as someone somewhere has a copy of the contnet the link will live forever, much like files in a torrent, even if the original host/uploader goes down the link to it does not.
Again though that's worst case scenario which is still better than what we have now, with additional peers it only gets better.
Anecdotally I am an American with a 50:50 (advertised, in use I actually get 60:40) with no bandwidth cap, this doesn't actually matter though because there are more countries on the internet than just America and content creators obviously don't have to host stuff out of their home just like is true today, they have the option to alone or in addition to a dedicated server to provide at least 1 source of redundancy just by themselves.
>>51444244
Nice, I don't like having to have yet another service running if I can just plug into one that's already running and setup.
>>51444250
>>51444300
I believe this is called Bittorrent DNA, I really wanted it to take off but it never took off the ground at all.
>>51444325
>>51444449
>have to rebuild an entire torrent file for every change
>no dedupliaction because lel swarm is unique to infohash, not single file/tree
>slow as shit DHT implementation, even emule is faster
>requires changes to work on anything other than IP transport
>2015 >sha1
Bittorrent was only designed to share large files fast, just because it's popular doesn't mean you have to use it for everything.
>>51444549
>building torrent files is hard
There's an overlay for gentoo but one has to google it and follow the instructions for it.
https://www.youtube.com/watch?v=HUVmypx9HGI
Around 50 minutes or so explains how dynamic data could work on IPFS.
Also, ethereum is going to use IPFS, so there's good synergies with future tech and good prospects for network size.
We should make a list of known sites and /g/ files, possibly add a search function to it too.
>>51445590
Sort of like a torrent index? Did they basically recreate the torrent protocol?
/g/ finally moving on from the concentrated shitfuck known as torrents and using things like IPFS and gnunet makes me insanely happy.
Torrents are cancer and must die.
>>51444549
I didn't mean to imply DNA was good at all, IPFS is leagues above it, it's a similar concept but done right.
>>51444300
project maelstorm, chrome client with bittorrent urls
Can someone explain how a retard like me will benefit from this? I don't really understand it, or what the purpose is.
So the goal is to have a single giant distributed filesystem? Sounds interesting. I might share it with a couple of professors I know who have an interest in distributed computing.
>>51445924
it allows fully distributed and decentralized data storage and communication. Hence, once content is up, it can only be taken down if every single user in the network who has the file decides to drop it and the cache of all users is purged. This profers censorship resistsance.
Moreover, the data is cached at nodes as it is accessed, so that downloading is extremely fast for popular content, and not particularly slower than with http for other files.
I like it.
>>51446186
For reference using the path with the filename works as well if you want it to have a real filename.
>>51446094
That sounds pretty awesome.
>>51446094
isn't a problem for people with shit upload speed or people with data caps?
>>51444594
>time consuming
ftfy
>>51446397
If you're the sole host and your upload speed is garbage, then you're no better off than with http if you're trying to upload massive files. If you're uploading many small files, though, people will quickly fill up their cache and alleviate the burden on you.
Datacaps typically only apply to download, not upload. You are downloading on-demand with ipfs.
>>51446449
yeah, but in the case you are only downloading something from a site, it would then also start to upload to another peer, pretty much like when you are sharing a torrent right?
Then what?
I know maybe this sounds like nitpicking but I had shit ISPs ebfore and I'm talking barely 1mbps up/down.
>>51446479
The only person who's inconvenienced by your shit upload rate is the peer trying to download a file off you. The peer actually doesn't care all that much, though, because they're not connected only to you, but also to another 20-100 peers.
The upload shouldn't count against your limit.
>hey guys
>this is my ip
>there is a direct link between that ip, the current time and my identity
>there are copyrighted files being hosted on that ip
>knock yourselves out
>>51446523
>copyrighted files
All files are copyrighted unless they're in the public domain
GNU/Linux is copyrighted but okay to share
Sorry just kind of rumbles me gears when people use "copyrighted" as a catch-all for "cannot make copies"
>>51446502
nice, I get it. In that case it would be even better for people with shit connections that have trouble getting files from someone hosting on a different country but it would improve as soon as someone near has the file or parts of it.
>>51446523
>what is tor
>>51446590
>P2P over tor
>>51446590
You can use anything to anonymize yourself, not just Tor. I hear a lot of talk about i2p being good for this.
>>51446663
lmao even with tor and i2p it's way too easy to fingerprint the shit out of your browser, sleep tight you pedo
>>51446686
What does that have to do with this? If you're that paranoid use one of those plugins that spoofs your browser information or don't use a browser to retrieve the files.
>pedo
wut, you're the one who wants to hide yourself, not me.
>404 page not found
>>51446686
>browser
Are you being retarded on purpose or were you born this way?
>>51446766
You have to append the ipfs link to the gateway url, noob.
>>51444075
how do we do this on chromebook?
>>51444213
Super interesting, great explanation on that page
>>51446856
Follow the instructions here: https://github.com/pr0d1r2/ipfs-gentoo-overlay
>>51444594
doing a one-off share with bittorrent is harder than it is with IPFS
with ipfs you just;
1. pin the content
2. give the other person a link
they don't even need to run the ipfs client, as there's a few ipfs-to-http services (http://ipfs.io/ipfs/<ipfs hash>)
>>51447004
how does it even work as quickly as it does?
just now i ran;ipfs add /tmp/a.png
added <ipfs hash> a.png
then immediately visited;
http://pastebin.com/BEJJbTtC
(ugh, spam filter)
and it loaded after about 5 seconds
it took far longer to post this post because of the damn spam filter
speaking of, why wasn't the spam filter deprecated when we started using captchas?
Thus far, I don't have too much to complain about the download speed, now that I've got shit working. I feel like there's a bit of bugs in some of the tools though. Documentation wants me to use paths like /ipfs/hash-name/filename, but if I do that, I get an invalid reference error. I have to use hash-name/filename. Also, I've been occasionally getting file lock errors, to which the only solution I can think of is to scrap the repository and try again.
>>51446523
>hey guys
>this is my seedbox's ip
>there is a direct link between an ip in another county, the current time and my identity
>there are copyrighted files being hosted on that ip
>knock yourselves out
>>51444250
probably the best advantage over bittorrent is that everything is in the same "torrent" with ipfs, and things are addressed by their content
so instead of multiple torrents on multiple trackers with multiple sets of peers that happen to contain the same file, you address the file itself, by its content (its hash), and you can download that file from anyone who has that file, regardless of where it is or how its named
if i make a torrent of say, a common tv show episode file, then you go to download it, i'll be the only seed, because i made a unique torrent which only i currently am seeding
this is in contrast to ipfs, where if i 'add' that episode, its hash is generated by its contents, which turns out the same every time, so i will be one of many seeders of that file immediately, i effectively add myself to the "swarm" of peers who have that file content
Does anyone know how to get ipfs to listen on all interfaces? I'm running it on a headless server in my basement and can't use browser to access localhost webui.
Anyone know how I can get it to accept connections from external sources? So I could connect from another computer by navigating to the servers ip.
>>51447424
presumably you'd change;"Gateway": "/ip4/127.0.0.1/tcp/8080"
in the config file to;"Gateway": "/ip4/0.0.0.0/tcp/8080"
>>51447452
That's probably it. Here's a pretty stupid question for ya. Where's the configuration file located by default? I can't seem to find it.
>>51447528
for me in arch it's at~/.ipfs/config
>>51444075
Sounds good.
I take it you can mount ipfs "shares" to e.g. "/mnt/shitshare" using FUSE or similar, correct?
>>51447537
Thanks m8. I changed the config. However I cant connect at all. Just get the "connection refused" error. Any idea whats goin on?
>>51447576
you can
<spam>
http://pastebin.com/sBGXVeRn
>>51447636
sure it's not firewalled?
>>51447650
Yeah. Opened all the ports necessary, and even disabled ufw for a test run.
>>51447638
>>51447576
and a folder, pic is the javascript atari emulator 'site'
https://ipfs.io/ipfs/QmeEDegBFx1gjHFDDLPhyinxdBXKrSbtgmcHn6tYEMhE77
>>51447342
So if I download a file on ipfs, am I automatically seeding it as soon as I have the file, or do I need to ipfs add it?
>>51447724
you need to pin it
you "add" new files (that is, from a file on your disk)
you "pin" ipfs hashes to keep a copy in your ipfs block store
https://www.youtube.com/watch?v=8CMxDNuuAiQ
introduction/demo
>>51447665
Well I changed the log level to debug and ran ipfs log tail, right after that I tried connecting to it throughhttp://192.168.1.101:5001/webui192.168.1.101 is the address of the server.
I found this in the mess of things, does anyone know what it means at all?{"addresses":["/ip4/127.0.0.1/tcp/4001","/ip4/192.168.1.101/tcp/4001","/ip6/::1/tcp/4001"],"event":"interfaceListenAddresses","system":"p2p/net/swarm/addr","time":"2015-11-21T05:50:16.207028629Z"}
>>51447733
that also applies to folders
and you don't need to worry about the paths/filenames of things you add
people could have on their disk for example;
~/Images/Ubuntu 5.04 32bit.iso
~/downloads/ubu5.0432.iso
~/media/ubuntu/5.4_32b.bin
etc, and as long as the /contents/ are identical, they all will have the same hash, and therefore the same ipfs address
>>51447770
Nevermind. Pretty sure this doesnt have anything to do with my connection. I cant seem to replicate it.
OK this is pretty cool. Just installed with homebrew and it works when I manually launch the daemon and leave it running in a terminal window. But how can I set ipfs to start at system launch in the background and keep running (since that seems the way it's meant to be run)
>>51447848
If I add something is that file automatically mirrored to the ipfs 'share' directory or is it shared in place?
>>51447848
Just addipfs daemonto your crontab. You can set stuff to run @reboot instead of at certain times.
is gateway.ipfs.io not working for anyone else?
Seems like a crazy number of peers.
>>51447807
>>51447770
I fixed this by restarting server.
But now the webui says I am connected to 0 peers. Anyone know anything?
>>51447678
Excellent. Distributed, written in Go (judging from the name), and Unixy as fuck in its abstractions.
>>51447863
It's mirrored, in place adding is on the todo list, I just can't find the issue on github. I want it too.
>>51448036
mhm
i pinned that atari site as well
and since things are addressed by content, even if the person who originally "added" that site stops hosting it for whatever reason (chooses to, goes offline, disk failure, etc, etc) the content will still be accessible from my machine /from the very same address/
imo this is the coolest thing about ipfs
What is this "pinning"? And what is its purpose?
>>51448121
Presumably bookmarking.
>>51448121
Think of it like bookmarks for files, when you pin a has you store a local copy of its contents, that way it will always be accessible to you. You can pin recursively too so you can mirror a whole directory with one command.
>>51448121
basically it marks cached blocks as "keep permanently (in the cache)"
things in the cache are accessible by others
for example, if i add a new file, then you pin it, i can turn my machine off and the file will still be accessible from at least your machine, since you've 'pinned' it in your cache
>>51448142
like bookmarks but actually permanent*
>a has
a hash
>>51448137
>>51448142
>>51448148
Thanks senpai. That makes sense.
>>51447863
>>51448060
Here it is:
https://github.com/ipfs/go-ipfs/issues/875
>>51448177
Holy cow. F a m changes to senpai. Didn't know that. Cool beans.
>>51448072
This is very cool. Something needs to be done to make it more user-friendly though, since it will only reach critical mass and become useful when millions of normal people install it.
Someone should build a PopcornTime like app around it (now that I think popcorntime is dead).
>>51448190
the aim is to replace http, this would mean web browsers would implement it
>>51444075
>be me
>TWC
>300mbps
>no data cap
>in TEXAS
Burger mode fucking DOMINATED
>>51448190
The end goal is browser adoption. There's already a web ui and api for controlling the daemon, you can add files via the webui.
If browsers had integrated support then the only thing people would need is a daemon/service that runs in the background on startup which is doable now. Outside of that the webui could be prettier and there's some core stuff that still needs to be finished/added.
This is handy for the time being
https://addons.mozilla.org/en-us/firefox/addon/ipfs-gateway-redirect/?src=cb-dl-recentlyadded
It redirects all links to the ipfs gateway to local host that way you can paste links everyone can use but not stress the gateway if you're running an instance. Seems a bit wonky on e10s though, could just be me.
>>51447884
Crontab isn't working for me. Are there any other ways to do it?
>>51448264
You're on OS X? You could make a launchd service for it.
>>51448281
Nah I'm on Ubuntu. Crontab works for all my other startup scripts, but the daemon doesn't start for whatever reason. No idea why.
>>51448214
I don't think that will happen unless there's a demand for it, and that won't happen unless it's already popular.
Actually, I can really easily envision how a piracy-focused app could be built that would be bigger than bit-torrent.
>Install PirateTime.app.exe.deb
>app scans for and shares all media files with ipfs (would have to implement share-in-place)
>reports the file's real names to dev's server
>dev's sever (PirateTime.by or whatever) hosts a search engine that lets people search by real file names, and they can download/view through their browser once they have the PirateTime daemon app running in the background
I suppose the problem is that the dev's server could be shut down. Off-hand I can't think of a way this would work (for the mainstream) without a search engine.
>>51448339
...on the upside however there'd be a huge opportunity for the developer of such an app to monetize it, since he could show ads on the search engine (even e.g. pre-roll ads before movies, since most people will be ignorant to the fact that all they need is the file's hash, not the whole web interface).
>>51448357
I'm on it.
How is this different from magnet uri?
>>51445924
Watch this
https://www.youtube.com/watch?v=skMTdSEaCtA
>>51448393
a magnet link in a hash of the metadata for a particular torrent
an ipfs hash references content directly
it's possible to put the same content in multiple torrents each with their own infohash (magnet link), where the same content with ipfs will /always/ have the same hash, which is a clear advantage
>>51448339
An alternative way to get the file hashes connected to file names would be to just scan torrent directories for the hashes (can you get the hashes for individual files, or just the entire torrent?)
>>51448339
There's an obviously leacher problem: someone could install the official ipfs and have access to all this, but without sharing.
Is it finally time for non-centralized internet?
This only works with static content, so it's pretty limited as a replacement for the modern web. How much on the internet is static? (I'd say just about none of it)
>>51447528
>linux problems
why isn't there a standard that forces all user config files and system config files to live a static location? Every fucking program wants to be special snowflake
This looks *really* cool, but needs adopters to take off...
>>51448187
It's the best thing asianmoot did
senpai posting was retarded
>>51448579
same can be said for any new protocol
http, gnutella, bittorrent...
>>51444549
Yeah, sha1 was probably a mistake, they should use something more recent
>>51448568
for basically everything it's at;
~/.program/, or
~/.config/program
windows is worse, with configs commonly at;
HKCU\SOFTWARE\company\program
HKCU\SOFTWARE\program
%appdata%\company\program
%appdata%\program
%userprofile%\documents\company\program
%userprofile%\documents\program
etc
How does hosting a website on this work? How are peers able to tell they have the newest version of the site without checking the original host? (Which would just completely get rid of the decentralized part of ipfs.)
>>51448579
It has received a lot of attention since it was announced and is constantly getting more coverage. I think this won't be a problem.
>>51448551
https://github.com/ipfs/faq/issues/28
https://github.com/ipfs/specs/issues/28
It isn't trivial but it can be done.
>>51448636
>IPNS
>IPFS hashes represent immutable data, which means they cannot be changed without the hash being different. This is a good thing because it encourages data persistence, but we still need a way to find the latest IPFS hash representing your site. IPFS accomplishes this using a special feature called IPNS.
>IPNS allows you to use a private key to sign a reference to the IPFS hash representing the latest version of your site using a public key hash (pubkeyhash for short). If you've used Bitcoin before, you're familiar with this - a Bitcoin address is also a pubkeyhash. With our Neocities IPFS node, I signed the image of Penelope (our site mascot) and you can load it using our IPNS pubkeyhash for that node: QmTodvhq9CUS9hH8rirt4YmihxJKZ5tYez8PtDmpWrVMKP.
>IPNS isn't done yet, so if that link doesn't work, don't fret. Just know that I will be able to change what that pubkeyhash points to, but the pubkeyhash will always remain the same. When it's done, it will solve the site updating problem.
>Now we just need to make the location of these sites human-readable, and we've got all the pieces we need.
>>51448551
Most of the internet compiles down to static files, the page this text is on right now is static, the only thing changing it is javascript. The same thing could be done with IPFS via ipns and pubsub.
Server generates html, generates hash of html, publishes hash to ipns, clients check it.
There was also some talk about live video streams being done in chunks this way, it's self archiving too.
https://github.com/ipfs/ipfs/issues/103
>>51448551
in terms of bandwidth share i'd say most of the web is static content
take youtube for example, the frontend is rather dynamic, but the video data is static, which is most of the actual data
>>51448655
They should probably make a TXT entry in the site dns or whatever pointing to the latest IPFS hash. Not really distributed, but would at least provide a link between domain names and IPFS.
>>51448655
Ah that makes sense. Thanks m8
>>51448673
https://github.com/ipfs/notes/issues/39 :^)
>>51448629
Windows user here, developers don't give a fuck about users or the file system standard locations. I sure love when config files are stored with the program binary, because that's where I think to look when backing up my shit.
*nix is literally 2(3) places for every config I've ever seen.
/etc/
(/usr/local/etc)
and ~
Even if they picked some random location in ~ it'd be better than Windows because you always backup you home, how often does anyone backup AppData\Local or Program Files\.
It's a god damn mess.
>>51448655
The example node given doesn't work. Is this not implimented yet?
>>51448733
not yet
I heard that i2p and tor routing is not implemented yet. Is there any way I can do it manually?
>>51448793
I'd Google it for you but I don't want to be put on a list.
>implying I'm not on one already
>>51448823
Kekd.
I'd Google it as well, but I'm on mobile and my browser is as slow as granny shirley playing fallout 4
>>51446856
You can get make and gcc on CrOS and use something like chromebrew, with that in mind I bet you could get pkg-src running on it. I need to actually try that this weekend now.
>>51444075
And now the pressing question
How good is encryption, and anonimity?
>>51449156
Not implemented yet. From what I understand there may or may not be a first party solution to this. The developers seem very focused on interoperability with other things, for instance if you want anonyimity you use it with Tor, I2P, or whatever other means you trust instead of relying on a first party solution that will not be as vetted as something already in use.
IPFS is just a replacement for http, DNS, and optionally filenames, if you want anonyimity and security you can get it in the same way you get with http (use a system around it like https+I2P).
>>51449156
encryption and anonymity defeats some of the advantages to ipfs, so it won't be integrates functionality, though as mentioned several times you can run through tor/similar if you need it
>>51446094
>here, let me store the web on my computer
Yeah, no.
>>51449302
I bet you posted that via a web browser that uses caching.
I'm trying this out;
ipfs/QmVnoWMgu5tfqZzwe ob2Qgp7xc4Q53KE88Zdms5tpE2eLT
can someone give me a screenshot of this? it should be an album cover of a game soundtrack.
curious to see how it works with larger files, later
>>51449367
>Ofu !!LLxDM1YR6nJ 11/21/15(Sat)17:59:47 No.514
>>51449455
you forgot to remove the space he added
>>51449367
not loading here, make sure the ipfs daemon is running and that the file is pinned
>>51449455
...actually nevermind there's a space in the hash you posted
>>51449455
He put a space in the key to avoid the spam filter dingus.
>>51449367
>>51444203
r u saying u dont like burgers
Can't post hashes directly it seems gets tagged as spam. Try this guys (just 'added'): http://pastebin.com/9d8dYtnd
>>51449455
did you remove the space?
>>51449460
alright, still figuring this out...
>>51449476
well, apparently it's already pinned.
this is really nice.
>>51449138
>Teffen
WHO? I'm a regular on #4chan and I've never heard of you before.
>>51448715
On Windows, the default install location for all of ipfs' shit is ~/.ipfs
>>51449367$ ipfs ls QmVnoWMgu5tfqZzweob2Qgp7xc4Q53KE88Zdms5tpE2eLT
Error: merkledag: not found
>>51449367
works, cool
>>51449502
I don't believe you, everyone on IRC knows that IRC is for idling only.
Any way to find out how many times a file has been uploaded?
>>51449503
there is a space in the hash, remove it
Poo in the loo
How does one write a webpage to host on ipfs? Do you include images locations as their hashes?
>>51449575ipfs cat Qmf45HsWvzQiGXTMu7PtLa9RNN73rQQCCWufSrUsqUvYSv
>>51449601
just use relative paths
>>51449624
It's a bit annoying since you have to 'upload' all the resources beforehand to find out what their hash will be
>>51449661
No you don't.
Create a directory, put files in it, use relative links to those files, add the directory recursivley, share the directory hash or the directoryhash/filename.html
i.e. QmP7LM9yHgVivJoUs
48oqe2bmMbaYccGUcadhq8ptZFpcD
/ipfs-site-how-to/index.html
broken up for spam reasons.
>>51449661
no, it's literally the same as a normal website
don't need to do any silly hash shit or change anything
https://blog.neocities.org/its-time-for-the-permanent-web.html
>>51449661
use relative paths
for example, look in;
http://localhost:8080/ipfs/QmbKxNNCxBox7C mv3jiUZbiG3zpzm tnYzVUuKHxfAjvpyH/
you'll see paths like;<script src="zepto.min.js"></script>
<script src="javatari/javatari.js"></script>
<script src="library.js"></script>
>>51449661
>>51449698
Related to this, the directory hash contains the filenames and IPFS hashes for everything so you don't lose the permanence despite using file names.
Using the hash in that post combind with the deepest directory before that index.htmlipfs resolve *dirhash*/ipfs-site-how-to/
Gives us an IPFS hash of that directory, then we doipfs object get *the hash from the previous command*
And we see a list of hashses and their friendly name.
>>51449698
But do the files that aren't directly hashed then benefit from peer mirroring?
>>51444395
Dude your reply just made me inspired to do programming and become a good programmer. I hope you are not a neck beard. What do you do in real life?
>>51449746
Yes, all files are treated in the same way, there's no special things going on. The only thing I can think of is that when you pin a directory recursively the children of it depend on it, so when you unpin the parent the children get unpinned as well (unless you pinned them directly before or afterwards). That's the only difference between a direct and indirect pin.
Can you limit your upload? I don't have a very fast upload and I have a bandwidth cap, but this looks really cool. Does it only upload things you pin? Can you move files that you've pinned, or do they need to remain in the same directory?
Maybe I am asking a stupid question now,
If I am the BBC, or a newspaper website eg. how can I know that the users will get text I want them to get. For example if I correct a mistake in my text, will users still get a uncorrected version from a different node?
>>51449750
This isn't the place for us to get to know each other but I wish you luck on your endeavors.
>>51449884
It depends, if you link out an IPFS hash then they will get that file every time, what you'd want is IPNS which is essentially a hash that points to another hash and the hash it points to can be changed. So like.
User goes to myIPNS.biz
myIPNS.biz points to (->) *filehash-old*
you make your corrections and generate a file/directory hash for it, then publish that hash to your IPNS so that
myIPNS.biz -> *filehash-new*
>>51444244
>downloading over Tor
You are scum.
>>51449981
Are you some kind of retard?
>>51449868
>Can you limit your upload?
I don't know if you can yet
>Does it only upload things you pin?
No, it uploads everything you've accessed through IPFS but what's not pinned gets garbage collected after a while.
so is this like freenet? If not, what are the differences?
>>51449868
>>51449997
It's on the todo list
https://github.com/ipfs/go-ipfs/issues/1482
I'm assuming they're going to finish the core stuff before this.
>>51450008
Think of freenet but opt in data redundancy instead of forced. Also read the thread, come on my mans.
>>51446094
what are the legal implications of someone posting CP in this model? woudn't posting cp to a website send the cp to everyone's cache?
>>51450214
I guess you have to access the CP and pin it in order for it to be kept on your computer?
>>51450214
People should use common sense and delete the files when they see them. If they aren't detected they will still end up being garbage collected after a while anyway, as long as they aren't pinned.
I guess that if the network gets really popular we'll also start getting lists of bad nodes that should be ignored.
>>51450290
Delicious chicken
>>51450308
ah true, that is a good point.
>>51450290
no, bombing it won't help
>>51450290
Good to know.
>>51450308
Blacklisting functionality is being worked on for the gateway and optionally your daemon if you so choose. The idea being that you subscribe to a blacklist and your node will never get those hashses, in the case where people upload to you and you yourself generate the hash then the data should be discarded, the hash will not be pinned if it's on the blacklist.
Personally I think this is a rather scary thing, having a list of all banned content coupled with a system that makes it very easy to retrieve content just seems like trouble, there's nothing anybody could do to take down any forbidden content outside of hunting down each node.
>>51450495
>Personally I think this is a rather scary thing, having a list of all banned content coupled with a system that makes it very easy to retrieve content just seems like trouble
What if that list of banned contentent contained the hashes but hashed one more time? This way it would be impossible to get to those files by downloading that list.
>>51450534
I didn't think of that, sounds like it would work.
>>51450534
oh yea, like hashed passwords
Im amazed at how this actually seems pretty decent, nice job at finding this, OP
Could this be used to create a PopCorn Time alike program? since this shit seems to be more secure than torrents while loading much more faster.
>>51447213
I don't know what you're doing, but /ipfs/hash/file should work just fine. You're using ipfs-cat or ipfs-get, right? That or using your browser and going to localhost:8080/ipfs/...
>>51447724
You will seed it because it's in your cache, but not indefinitely. If you pin it, it will be seeded indefinitely.
>>51448214
They say they're in talks with a major browser vendor (I'm guessing google) to support ipfs natively already.
>>51447988
It always says that, the log and peers info pages don't work in this release. Use ipfs swarm peers to see which peers you're connected to.
>>51447896
What's not working for you in particular?
>>51450214
Only to the cache of people accessing it. All communications are encrypted, so that's good, but if e.g. a chan is run, you visit the chan, and someone posts CP, you'd better hope you're using tor (same as over HTTP but you're now hosting it until your cache fills up), or delete it manually.
It's likely that you can make a case that it should be treated the same as browser caches, however, but I wouldn't risk that if I were you.
>>51452064
Popcorntime simply parsed outputs from yts and eztv and stored the results in a DB, obtaining torrents to serve the user. It is entirely possible to mirror eztv in ipfs, and it is entirely possible to download the videos via torrent and add them to ipfs, and then serve ipfs links instead of torrents. The latter, however, is obviously more trouble (double upload), but ipfs is already fast enough for this.
Can somone explain to me how this is different from Share, Perfect Dark, Winny, etc...?
This caching stuff sounds like a unity folder used in the aforementioned Japanese p2p systems.
>>51452516
It's pretty much like a browser cache/caching proxy, when something is fetched from ipfs it's stored in the datastore, which can be cleaned removing everything minus the pinned objects.
All that japanese stuff is based on pushing encrypted blocks to other peers, while ipfs is based on pulling from others and then storing.
Started working on a links archive (I'll probably augment it with a search function eventually):
/ipns/QmaGks9KKzu2WykHQjJFJkcUAN4ZoF7ok9h2hXj1WQn47U/
Current version is
/ipfs/QmXHnrxgXfFtLZZ2oyksAHbjfYFsGofSLzrxaiqKQRUvow
>>51452850
*So, if you're publishing anything, post itt and I'll add a link on this site.
>this thing
>>51452249
Can ipfs 'swarm' downloads. I.e. download parts of a file from more than one peer at a time?
>>51452889
Yes, that's how it works.
>>51452850
Oops, slight mistake; corrected. Same ipns address, current version is /ipfs/QmeLhR7VpP6ENGjXcHQ7SFZwYQSLNGpJWFZ77J3h3WTjVf
>>51444075
>>51452850
>>51452913
How about using https://ipfs.io/ipfs/... links so that plebs can directly see how it works and others can use a redirection addon to redirect to a local gateway (which might not be on localhost)
https://addons.mozilla.org/en-US/firefox/addon/ipfs-gateway-redirect/
https://chrome.google.com/webstore/detail/ipfs-gateway-redirect/gifgeigleclkondjnmijdajabbhmoepo
https://github.com/loadletter/ipfs-redirect-userscript
>>51452990
The site itself uses relative links so it works whether you're using ipfs.io or if you're running locally (or whatever other gateway).
It shouldn't be too hard to add ipfs.io before the urls posted itt now come on.
>>51453030
>want to get more users and make it easier
>have to copy-paste all over the place
>>51453093
Fine you little faggot.
ipfs.io/ipfs/QmY17p25SpwQeQ
hs9DvYNhNT39uerhZ8
sCqLcT6fESe9tK
Latest version of the content index. Install Gentoo!
ipfs.io/ipns/QmaGks9KKzu2
WykHQjJFJkcUAN4ZoF7ok
9h2hXj1WQn47U/
"Permalink", always refers to the latest version.
>Path Resolve error: context deadline exceeded
what do?
>>51453561
Nothing, it just werks.
>>51446218
I can't get this to work, what is the right format?
I'm confused. If everything is distributed by hash, then what of files that must be updated?
Sure, it could be used as a CDN to serve media or backup discussion, definitely. But how do you link to it WITHIN IPFS?
Like, you can't create a blog if that blog could never change its hash. You'd have to link to an updated hash constantly, which means you'd need a stable address elsewhere.
Am I missing something?
>>51454405
you use an IPNS hash
>>51454405
That's what ipns is for. That's why the ipns address at >>51453159
never changes, but you can see individual versions of the site at >>51453159
>>51452913
and >>51452850
via the ipfs link.
>>51454423
>>51454460
>>51454469
Oh. That's pretty damn neat.
>>51446094
>Moreover, the data is cached at nodes as it is accessed, so that downloading is extremely fast for popular content, and not particularly slower than with http for other files.
Also, the information is stored at center points as it is accessed, so that downloading is significantly quick for well known substance, and not particularly slower than with hyper text data exchange protocol for distinctive documents.
>>51454423
/ipfs/QmdjnJcXmAFTrSGtKKYg2MZUCCxGaL2CAAbG2mT8kjoMQH
So you can use this as a bittorent program? Using current bittorent sharing websites?
>>51447342
I realize I'm late to this party, but wouldn't hash collisions devastate IPFS?
>>51455023
no, it's uses some bittorrent technologies/ideas, but is not bittorrent compatible
>>51455296
Hash collision rate is something like one in 10^45, multihash are used, AND it's designed to be fully upgradable.
>>51449226
>>51449248
Well if data transfer goes over hops then I can't really see a reason not to use onion encryption similar to Tor.
>>51449868
as a workaround you can run the daemon through trickle
QmcP27Xx3Pkbw5UZBWxdd4F cp537FzxwNS7wdANEc7cJbk
MP4 v2 NSFW
Works?
>>51455296
this is brought up with almost anything that uses hashes to a large extent
the tldr is that the likelyhood of a collision is way, way lower than you'd think, considering the apparent small size of the hash vs. the sheer number of potential files and their sizes on something like IPFS, should it get big
>>51455352
Downloading
>>51455009% ffprobe <(ipfs cat /ipfs/QmdjnJcXmAFTrSGtKKYg2MZUCCxGaL2CAAbG2mT8kjoMQH)
...
title : BWV 0532 Praeludium et Fuga in D - Praeludium
album : J.S. Bach - Organ Works CD 05
artist : Ton Koopman
genre : Classical
track : 11/12
encoder : Audiograbber 1.83.01, LAME dll 3.98, 320 Kbit/s, Joint Stereo, High quality
>>51455376
Once people have downloaded it, I can delete it and it will still be available at the network? Just like bittorrent?
>>51455393
so long as other people pin it (analogous to seeding)
>>51455393
Only if they rehost it or as long as it remains in people's cache.
>>51455408
ps, you pin it like this;ipfs pin add /ipfs/<hash>
does what you think, just 'pins' it to your cache, so it doesn't get cleared out automatically
>>51455393
Boner achieved! Kek at the metadata.
Does it work on Windows?
>>51455386
It works then!
>>51444244
That kinda makes it less, cool. If I one needs ipfs to access the sites, one could've built encryption/anonymity into it. Or build ipfs on top of i2p or something.
>>51449559
I did.
>>51455415
Is there a command for downloading and pinning so we don't have to do it manually?
>>51455512
What metadata?
>>51455526
Yes. Works on anything from plan9 to solaris to windows to GNU to OSX. It probably even works on templeos.
>>51455532
You can use encryption on top of it. You can't mask the IP addresses, however. That said, the creator is working for i2p and tor integration, it seems.
>>51455563
Does it run in the background like a service after I init it or do I have to do everything manually?
I might just wait until it hits beta
>>51455534
well if you pin something directly, it will be downloaded (since you actually need to data to have it pinned)
>>51455567
>You can't mask the IP addresses, however.
You mean without i2p/tor? Because they do exactly that.
>>51455532
Why reinvent the wheel when there are already secure protocols, IPFS doesn't make any assumptions about the underlying network transport so it's easier to use i2p/tor instead of TCP/IP.
>>51455352
>QmcP27Xx3Pkbw
5UZBWxdd4F cp537FzxwNS7wdANEc7cJbk
2nd part of the same NFSW:
QmP5XAx1jVDaR2jupqc 1CxKWr3rWVGBXDCd
QmegpLtqmqi
>>51455583
You can run it as a daemon withipfs daemonor use the initscript at https://github.com/dylanPowers/ipfs-linux-service
>>51455567
Is there an ipfss analogous to https? I hate certs and all but we need some bidirectional encryption that's validated by authorities/network/whatever.
>>51455646
on arch it's as simple as;systemctl --user enable ipfs
Is there a way to see how many peers are downloading/seeing a file of mine?
>>51455652
>that's validated by authorities
No.
>>51455652
MITM is impossible as everything is content-addressed
since you directly request using a hash of the content, getting anything other than what you asked for would result in content with the wrong hash
>>51455652
ipns keys are based on the public key of the client
>>51455693
It would still be vulnerable to surveillance without effective anonymizing methods.
>>51455563
Would IPFS be considered an extension of Plan 9's design goals?
>>51455732
yea, but it's not intended to be anonymous
one of its' advantages relies on not being anonymous
you can run it atop an anonymizing protocol if you wish
>>51455745
from what i've read about plan9, it does seems pretty plan9-ey
>>51455693
Does that mean one can only deliver static content? How is this supposed to replace http then?
>>51455793
ipns, another kind of hash which is a signed reference to the latest version of some content
>>51446686
>lmao even with tor and i2p it's way too easy to fingerprint the shit out of your browser, sleep tight you pedo
Knock yourself out, loser: https://github.com/CrisBRM/user.js/
>>51455408
Does it mean that if I justipfw geta file I won't be sharing it unless I pin it?
>>51455863
i believe it'll be share while it's in your cache (might be wrong on this), but it won't stay in the cache permanently unless it's pinned (it'd be silly if it caches everything permanently by default, you'd eventually run out of space)
What's the advantage of this over the current internet model?
AFAIK, it's a whole lot of complexity for little to no gain. It can neither beat bittorrent when it comes to filesharing nor beat http when it comes to website hosting.
>>51455863
You will share it until your cache gets garbage collected, if you pin it then you will share it indefinitely until you unpin it.
>>51455526
Everything except adding files in my experience.
>>51455895
Read the thread.
>>51455693
But the content sent to you can be sniffed.
>>51455891
ps. this would work fine for commonly-accessed files
as long as some people occationally access it, it'll be in some peoples' caches, and so there'll be people seeding it
>>51455895
literally better in every way, see some of the talks on youtube
>>51455920
I am, and honestly? It doesn't do shit. This is onion routing without the onion.
