kevina has quit [Remote host closed the connection]
ZaZ has joined #ipfs
j12t has joined #ipfs
mgue has joined #ipfs
byteflame has joined #ipfs
byteflame has quit [Client Quit]
byteflame has joined #ipfs
tmg has joined #ipfs
r04r is now known as zz_r04r
byteflame has quit [Ping timeout: 244 seconds]
pfraze has quit [Remote host closed the connection]
jedahan has joined #ipfs
<disgusting_wall>
Hi, I have a question. Does js-ipfs-api have a version that works to .add and .cat files while embedded in the browser, and if so is there a particular method to add it to the page, and which browsers has it been tried on?
<disgusting_wall>
I've tried several different versions with chrome, and none of them could add a file, although an older version returned "undefined" when a file already existed instead of returning an error. I also haven't tried browserify, just the npmcdn method
ZaZ has quit [Read error: Connection reset by peer]
reit has joined #ipfs
jedahan has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<deltab>
what error did you get?
<disgusting_wall>
i'll try again with the file i set up and give you specifics, but it returned a promise object and probably said something about file streams
<disgusting_wall>
Error: Can not add paths in node
<disgusting_wall>
in response to "window.IpfsApi().add(["test"])"
pfraze has joined #ipfs
<disgusting_wall>
i may just not know how to use it, but i've tried all different kinds of parameters to the add function, and earlier versions were clearly changing behavior in response to my node's status as i said above
<disgusting_wall>
that, or what is the best way to create a compatible buffer or stream?
<brendyn>
How can I start learning to understand go-ipfs? Should I do some go tutorials first or?
pguth has quit [Remote host closed the connection]
pguth has joined #ipfs
chrisg_ has quit [Ping timeout: 244 seconds]
<JesseW>
brendyn: Do you feel like you understand the IPFS protocol? If not, that would be a good place to start.
<brendyn>
I understand DHT informally. But I don't understand anything on a technical level. Basically a programming noob, but I've used GNU/Linux for years so I have a little head start.
sol56 has quit [Ping timeout: 260 seconds]
<JesseW>
IPFS is more than just DHT
<brendyn>
I want to figure out how to do things like how many "seeders" there are for a hash, or how big a file is given it's hash, before downloading it
<JesseW>
If you want to read the go-ipfs source code, yeah, doing a Go language tutorial is probably a good idea, though.
<JesseW>
Ah, I'm interested in those things, too.
<JesseW>
I'd suggest reading the documentation on the ipfs CLI interface as a first step.
sol56 has joined #ipfs
<brendyn>
I was thinking of creating a service where people could pay money/cryptocoins to pin X GiB for Y upload quota, so I'd need those numbers to set prices. I mean, I'd do it for free but I don't have much money for servers :p
pfraze has quit [Remote host closed the connection]
<JesseW>
Regarding how many seeders -- you can make a request on the DHT level for "providers" (i.e. seeders), with `ipfs dht findprov`
<brendyn>
We could make a GUI reminiscent of a torrent client where people could for example click Gutenberg, 20GiB space, 200GiB upload limit, and then they would dynamically join a cluster to help seed public domain books.
<deltab>
might even be possible to add that to a torrent client
<brendyn>
Probably bittorrent has had more years to deal with malicious clients so there may be solutions in that domain.
jsrockss has joined #ipfs
tmg has quit [Ping timeout: 276 seconds]
qgnox[m] has joined #ipfs
jsrockss has quit [Ping timeout: 244 seconds]
rgrinberg has joined #ipfs
jaboja has quit [Ping timeout: 260 seconds]
<brendyn>
Is it possible to make IPFS store whole files instead of blocks?
<JesseW>
brendyn: what do you mean?
<brendyn>
When I pin a file it gets duplicated in $IPFS_PATH/blocks/. If it was simply the original file, like bittorrent, I could use my BTRFS filesystem to deduplicate it with the original copy
<brendyn>
I don't want to have two copies of everything on the same drive.
<brendyn>
The alternative is I somehow mount ipfs and use symlinks but that seems quite tedious
<JesseW>
Ah, that makes sense. I remember hearing something about that... hopefully someone else can pipe up?
<brendyn>
Libgen seeds everything with torrets, but they use MD5 instead SHA. I'd like to seed on both torrent and IPFS without duplicating everything
rgrinberg has quit [Ping timeout: 240 seconds]
PrinceOfPeeves has quit [Quit: Leaving]
matoro has joined #ipfs
matoro has quit [Max SendQ exceeded]
pguth has quit [Remote host closed the connection]
pfraze has quit [Remote host closed the connection]
<ipfsbot>
[go-ipfs] JesseWeinstein opened pull request #3018: Put all Taglines & most Option descriptions in active, present tense (master...adjust_tense) https://git.io/v6eqH
jedahan has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
meadman has quit [Read error: Connection reset by peer]
<Codebird>
Someone else compiled a list of IPFS-hosted sites.
<jarboot>
woah thanks
<Codebird>
Some of dubious legality. But NOT MINE.
<jarboot>
Any idea who to talk to to add something to it?
<jarboot>
It needs some Serial Experiments Lain in the anime
<Codebird>
You don't. That's an IPFS link, not IPNS. Immutable.
<jarboot>
ah
<Codebird>
There may be a newer version, but I don't have it.
<jarboot>
shit should have known the difference I just watched that stanford talk he did
<cow_2001>
git checkout v0.4.3-rc1 && make install :D
<Codebird>
The index does show that IPFS is already reflecting the diversity of the conventional web though: Political ramblings, piracy and pornography. The bread and butter of the internet.
<Codebird>
Just needs a library of cat photos now.
<Codebird>
Nope. You create any communications technology, it will be used for things the creator did not intend.
<cow_2001>
oh rhgt
<cow_2001>
sorry
<jarboot>
looks like it's time to read through the code of conduct :)
<Codebird>
I have a list of my own somewhere - a smaller index of sites I found, without any copyright infringement or pornography. Give me a minute to find it.
<cow_2001>
who in their right mind host stuff on ipfs without anonymization?
<cow_2001>
i mean, that kind of stuff
<cow_2001>
you should add some "ipfs is not torified" warning so that people in teheran or china won't get burned
<Codebird>
People going below the radar. It's a new network - the various 'authorities' both public and private sector probably aren't very aware of it yet. And tracking someone on IPFS is no easier than finding them on plain http, anyway.
<Codebird>
No harder, either.
<cow_2001>
hmm
<Codebird>
Plus, while it may no harder to trace, it's a lot harder to take things down.
<cow_2001>
i mean, people might be confused and think it's torified-ish
<Codebird>
Plus it'll scare some of the undesireables back to Freenet from whence they came.
ric2b has joined #ipfs
IlanGodik has joined #ipfs
erde74 has quit [Ping timeout: 276 seconds]
zz_r04r is now known as r04r
pguth has quit [Remote host closed the connection]
pguth has joined #ipfs
pguth has quit [Remote host closed the connection]
pguth has joined #ipfs
cyberwolf has joined #ipfs
cr[m] has joined #ipfs
cyberwolf has quit [Quit: Konversation terminated!]
Encrypt has quit [Quit: Lunch time!]
xenic has joined #ipfs
Oatmeal has quit [Ping timeout: 250 seconds]
cketti has joined #ipfs
Oatmeal has joined #ipfs
pguth has quit [Remote host closed the connection]
pguth has joined #ipfs
rgrinberg has joined #ipfs
zorglub27 has quit [Ping timeout: 276 seconds]
computerfreak has joined #ipfs
sametsisartenep has joined #ipfs
zorba has joined #ipfs
zorba has quit [Quit: Lost terminal]
rgrinberg has quit [Ping timeout: 252 seconds]
Foxcool has joined #ipfs
zorglub27 has joined #ipfs
herzmeister has quit [Quit: Leaving]
herzmeister has joined #ipfs
reit has quit [Ping timeout: 258 seconds]
frankyr has joined #ipfs
pfraze has joined #ipfs
G-Ray has joined #ipfs
<Kubuxu>
Codebird: please don't use gateway.ipfs.io , we don't want it to be used for links that are visible, only when you do for example CNAME redirects.
<Kubuxu>
or fs:/ipfs/ for displaying them with https://ip... as href
<Kubuxu>
^^ that if you want to show the hash to user
<Kubuxu>
and you but you can direct click to https://...
<Kubuxu>
we want users to get accustomed with fs:/ipfs/... so then know what to do with theme if their browser or program does not.
<brendyn>
So you don't use ipfs:/Qm.... ?
herzmeister has quit [Ping timeout: 250 seconds]
<Kubuxu>
no, we would love to use just cannoncial path `/ipfs/...`, `/ipns/...`, `/ipld/..` but this won't be supported for some time
<Kubuxu>
by external software
<Kubuxu>
that is why we decided to use `fs:$canoncial_path`
<cjd>
ipfs://abcdx would be cool since it looks like http
<brendyn>
Except the second / is pointless :p
<cjd>
well so is the first one, but it is familiar for users of http/https
<demize>
To use UNIX like paths it would be more proper to use the file: protocol rather than a custom one, really.
<Kubuxu>
demize: we can't use file: as it is already handled by browsers
<Kubuxu>
and policies on this path are just screwed up
Stebalien has joined #ipfs
<Kubuxu>
cjd: we don't use `ipfs:/` because then we would need `ipns:` `ipld:` in future `iprs:` and so on.
<demize>
Wouldn't be a problem if you're okay with just requiring ipfs to be running locally.
<cjd>
I would
<cjd>
because it's memorable, all IPFS branding has been around the term "ipfs"
<demize>
And handling fs: in any way that doesn't work universally for all filesystems would be misleading.
<Kubuxu>
not really, file: protocol has already established for example: origin policy (which is origin per file) which doesn't work on IPFS.
<Kubuxu>
also there is no support for FUSE on Windows so far.
<Kubuxu>
there is project that can fix that but it is far from complete
<Kubuxu>
also current fuse handling doesn't work as great as we would like and we don't have enough manpower right now to fix it.
matoro has joined #ipfs
<Codebird>
I'll update the script that makes that index page to use https://ipfs.io/ipfs/
<Kubuxu>
great thank you
<brendyn>
I imagine we could transparently access ipfs like a file system if we were all on GNU Hurd
kholdstayre has joined #ipfs
<Kubuxu>
the problem we have with fuse (and Linux fs in general) is that many operations that could be optimized in IPFS can't be optimized using FUSE interface
<Kubuxu>
or Linux in general
<Kubuxu>
example: IPFS supports 0 cost copy, there is no way for us to support it from FUSE
m3s has quit [Read error: Connection reset by peer]
taaem has joined #ipfs
jaboja has joined #ipfs
<brendyn>
Is there a solution on the horizon?
palesz[m] has joined #ipfs
<palesz[m]>
did anyone do any performance analysis with ipfs recently?
<palesz[m]>
let's say: I add a file locally
G-Ray has quit [Remote host closed the connection]
<palesz[m]>
how much is the expected time to be able to get it through the ipfs.io/ipfs/hash url (not locally)?
<palesz[m]>
based on my experience, ipfs.io is incapable of resolving the urls/hashes (I've waited 10+ minutes, still nothing)
jaboja has quit [Ping timeout: 276 seconds]
<brendyn>
How big were the files?
<palesz[m]>
1 picture, 2-3 MB
<brendyn>
Try it with a few bytes
<palesz[m]>
I see 70+ connected peers locally
<brendyn>
Plus, this is just testing the gateway?
<brendyn>
For small files It takes a few seconds for me
<Codebird>
I get that sometimes.
<Codebird>
I just just exiting the local ipfs daemon and relaunching usually fixes it.
<IlanGodik>
There seems to be a problem with the automatic migration on Windows in 0.4.3-rc1
<IlanGodik>
Was it reported already?
jaboja has quit [Ping timeout: 250 seconds]
zorglub27 has joined #ipfs
jedahan has joined #ipfs
pguth has quit [Remote host closed the connection]
pguth has joined #ipfs
taw00 has joined #ipfs
taw00 has quit [Remote host closed the connection]
kants has quit [Quit: sleep.]
jedahan has quit [Client Quit]
rardiol has quit [Ping timeout: 240 seconds]
zorglub27 has quit [Ping timeout: 244 seconds]
frankyr has quit [Quit: Connection closed for inactivity]
<Kubuxu>
IlanGodik: I've seen a mention.
<IlanGodik>
I've filed an issue now
<Kubuxu>
Thanks
<Kubuxu>
we should fix it before release of 0.4.3, best in 0.4.3-rc2
<Kubuxu>
if you want to use 0.4.3-rc1 you can run migrations manually
<IlanGodik>
I already did, thanks
Stebalien has quit [Ping timeout: 240 seconds]
TheWhisper has quit [Read error: Connection reset by peer]
TheWhisper has joined #ipfs
jedahan has joined #ipfs
zorglub27 has joined #ipfs
jedahan has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<demize>
Any clue when to expect 0.4.3 to be released?
Looking has joined #ipfs
<Kubuxu>
We planned to release rc2 this week (didn't happen), so rc2 next week and if nothing is broken 0.4.3 week after that.
<Looking>
I don't have any POC but I think adding stuff to IPFS make my wifi crash, Ik it's wierd but time correlation seem clear to me .
<demize>
cool.
<Kubuxu>
Looking: it is possible, if the router handles conntracker overflow incorrectly (and has too low conntracker size).
<Looking>
oh, it make sense then
<Looking>
it's possible indeed, it's a cheap wallmart router.
<Kubuxu>
we were trying it fix that with 0.4.1 and 0.4.2 but had to reverse it completely in 0.4.3 as utp library was too unstable.
JesseW has joined #ipfs
<Looking>
I had a question on IPNS, I first do: 'ipfs add' then 'ipfs name publish' after this. If I want to change the content i just 'ipfs add' and 'ipfs name publish' agan MAGICAL! But how does it know it's me and that Im legitimate to change the hash ? and Can I share this auth with others nodes ?
<Looking>
usecase is to have a public folder where everyone can dump files.
<demize>
The private key of your identity is used to authenticate you.
<demize>
The ID that whatever you add is published at is your peer id
<Looking>
ok I see. Can I share it with a remote server to give him write access ? or it would cause collision.
xenic has joined #ipfs
<JesseW>
Looking: it is planned to allow publishing to other keys, but that isn't implemented yet
<Looking>
ok. ill stick with having a single owner for now. Thx for your time.
<apiarian>
is there something tricky with IPNS and IPTB? i'm making calls to Publish from a go-ipfs-api client and not getting an error, but then when I try to call Resolve, i'm not seeing the new hash. was there a minimum number of nodes below which IPNS is unreliable?
<apiarian>
oddly, i see the correct hash if i ask another node in the testbed to resolve things. it is only the original node that is failing to resolve itself
jedahan has joined #ipfs
ianopolous has joined #ipfs
G-Ray has joined #ipfs
mildred has quit [Quit: Leaving.]
mildred has joined #ipfs
jarboot has quit [Ping timeout: 264 seconds]
<apiarian>
ah, has something to do with caching. if i send a nocache = true with the resolve request, It works correctly