<tsp>
Now my daemon is just sitting there, I don't think it's doing anything. There's no size increase on the directory I sent it to. I'll try a restart.
<tsp>
I couldn't interrupt, had to send a quit.
<tsp>
To the foreground ipfs get, that is.
<tsp>
There we go.
<achin>
if you add up all the sizes in the output of "ipfs ls" you can get teh total size
<achin>
(in this case, it's 383K)
<tsp>
Yeah, ipfs get told me after it was done. 415.50 KB 0
voxelot has quit [Ping timeout: 240 seconds]
reit has quit [Quit: Bye]
<tsp>
There's a typo in the bootstrap docs. ipfs add vs ipfs bootstrap add
<whyrusleeping>
thats how youre supposed to spell segway??
<whyrusleeping>
whoa
<deltab>
segue is the usual word, segway is the vehicle
<ansuz>
ansuz is spelling bee champ 1993-1996
<deltab>
.w segue
<multivac>
[WIKIPEDIA] Segue | "A segue /ˈsɛɡweɪ/ (Italian pronunciation: [ˈseːɡwe]) is a smooth transition from one topic or section to the next.The term is derived from Italian segue, "follows"...." | https://en.wikipedia.org/wiki/Segue
<livegnik>
Isn't Alexandria building on-top of IPFS? That would basically be the IPFS-Youtube, would it not?
<livegnik>
With a micro-payments model, that is.
<davidar>
livegnik (IRC): possibly, have to admit I haven't looked at it in too much detail yet, probably should
<shyamsk>
davidar: Is there a way to limit the amount of data stored on a node? Say maybe my daemon only allows caching upto 1 GB or something of that sort. (I had forgot to ask it along with my last set of questions)
<shyamsk>
also ipfs.pics code, source? I really would like to know how the backend works
<davidar>
it's not open source as far as I'm aware
<davidar>
I'm not even that sure who built it tbh
<spikebike>
Do not let your CDN betray you: Use Subresource Integrity - @fmarier & mozfreddyb explain. In #Firefox Dev Ed 43 now http://mzl.la/1Lbz9LT
<spikebike>
(tweet from Brian Warner)
r04r is now known as zz_r04r
<livegnik>
davidar: It's hard to keep up with everything. I know the feeling.
akhavr has quit [Ping timeout: 268 seconds]
voxelot has quit [Ping timeout: 264 seconds]
<shyamsk>
are there any applications (open sourced) with some backend processing that I could go through. I saw a few examples at https://github.com/ipfs/examples/tree/master/webapps but they all seem front-end driven. Anything that has a little more of a backend?
captain_morgan has joined #ipfs
Eudaimonstro has quit [Read error: Connection reset by peer]
Eudaimonstro has joined #ipfs
akhavr has joined #ipfs
<r1k0>
hi, noob here, just discovered ipfs, ran ipfs init yesterday, could read basic files but now I lost/forgot the actual hash from the url. How can i find it back without re running ipfs init -f? thx
akhavr has quit [Ping timeout: 264 seconds]
<r1k0>
is it ok to run ipfs init -f? noticed the hash url stays the same but the peer identity changes
Stard0g101 has joined #ipfs
<gendale_>
alexandria is built on top of ipfs
<gendale_>
and yeah, hosts video, audio, etc
<gendale_>
also stores hash links in a blockchain, so the indexing is p2p as well
<gendale_>
r1k0: init file is stored at .ipfs/config
dignifiedquire has joined #ipfs
Judge_Dredd has joined #ipfs
Judge_Dredd has left #ipfs [#ipfs]
jfred has quit [Remote host closed the connection]
jfred has joined #ipfs
jhulten_ has quit [Ping timeout: 240 seconds]
<_fil_znc>
YT has a conversion pipeline which was its real strength
<_fil_znc>
not only hosting
ygrek has quit [Ping timeout: 268 seconds]
<davidar>
livegnik (IRC): the authors are here sometimes, but from my perspective it kind came out of the blue
reit has quit [Quit: Leaving]
zz_r04r is now known as r04r
<davidar>
zignig, pipeline! ^
reit has joined #ipfs
<davidar>
s/d/da
<multivac>
davidar meant to say: livegnik (IRC): the authors are here sometimes, but from my perspective it kinda came out of the blue
<gendale_>
_fil_znc: that's a good point
mafintosh has quit [Ping timeout: 240 seconds]
ffmad has quit [Ping timeout: 240 seconds]
<r1k0>
gendale_, thank you for your reply, unfortunately I cannot find the hash url pasted by ipfs init inside .ipfs/config
mappum has quit [Ping timeout: 240 seconds]
henriquev has quit [Ping timeout: 240 seconds]
ffmad has joined #ipfs
mafintosh has joined #ipfs
<davidar>
r1k0 (IRC): yeah, I'm not sure where that hash comes from
<r1k0>
how do you guys remember your url hash?
henriquev has joined #ipfs
mappum has joined #ipfs
<r1k0>
I'm tempted to rerun 'ipfs init -f' to find my hash url again but my peer identity changes and I guess it's not good, so I was looking for a way ( a command or a file) that will tell me my hash url
<davidar>
There's also an old hash in the docs somewhere
ilyaigpetrov has quit [Remote host closed the connection]
<r1k0>
I'm sorry if I'm not clear. This is what I mean: http://pastebin.com/wuQg470n if I've run 'ipfs init' a while ago and I forgot it, how do I find again the /ipfs/<hash>/readme url?
<r1k0>
or more simply: how do you guys find the /ipfs/<hash> again so you can cat a file and see its content? Do you store it aside on your own? do you type a command to figure it out or do you learn it by heart? ^^
<r1k0>
does my question even make sense? Maybe what I'm asking is totally non sense, if so, please let me know, that'll me ask myself better questions
<r1k0>
make me ask better questions
<r1k0>
thx for reading
<dignifiedquire>
r1k0: did you try looking at the files list in the web ui?
<r1k0>
dignifiedquire, I'm aware there is a webui (read from the docs) but I'm not there yet, still playing/discovering around how the ipfs command work, I will try to have a look at it, maybe that will help me answer my questions, thx for your suggestion
hellertime has quit [Quit: Leaving.]
<dignifiedquire>
r1k0: as soon as the ipfs daemon is running you can go to http://127.0.0.1:5001/webui and there is section called “Files” which lists all the files, you can see a screenshot here: http://grab.by/KWLk
<r1k0>
dignifiedquire, that's help indeed, thank you for pointing that out
<dignifiedquire>
r1k0: :)
pfraze has joined #ipfs
<r1k0>
in any case I'll be hanging around here, because I've been dreaming for years (litterally) on software like ipfs seems capable and spent quite some time trying to bend tribler in such a way but it's not there yet, ipfs design is purely mind blowing and it's an understatement
Syntist has joined #ipfs
pfraze has quit [Ping timeout: 255 seconds]
<davidar>
r1k0 (IRC): welcome aboard :)
<Syntist>
go get -u github.com/ipfs/go-ipfs/cmd/ipfs package github.com/ipfs/go-ipfs/cmd/ipfs: cannot download, $GOPATH not set. For more details see: go help gopath
<Syntist>
can online games use ipfs? like if there are people from same country playing with you will lesser the lag?
<rendar>
Syntist: how online games could benefit from ipfs? if you mean for static resources like images, textures etc, sure..but sometimes the lag is due other kind of dynamic data
<Syntist>
i mean if you are playing on the same net. The only one user need to download the data and it can share you that folder. you just need to upload the data
<rendar>
static data like textures, or game savings, yes, why not?
<Syntist>
lol sorry its data instead of folder
<rendar>
but consider that big games already have such optimizations
<Syntist>
k
<Syntist>
can i torrent video in ifps xD and it will stream it online haha
ei-slackbot-ipfs has quit [Remote host closed the connection]
ei-slackbot-ipfs has joined #ipfs
<DavidBurela>
I saw https://github.com/noffle/ipget and got excited that there was stand alone code to hook into IPFS to retrieve files that I could appropiate and convert into a .net app. Then i looked deeper and realised that it was just importing the go-ipfs library v_v;;
<locusf>
could someone cat this QmdEdeSHi9n7RgHPTDcgqGLqKHYNBxqcW3QdEZsajpnqeg to pxz and see if it outputs base64 data?
<davidar>
DavidBurela (IRC): does .net not have an ffi?
<locusf>
or xz
<DavidBurela>
o_0
hellertime has joined #ipfs
<davidar>
DavidBurela (IRC): I don't understand what your issue with go-ipfs is?
<DavidBurela>
my issue is, I'm a .Net developer. I'm trying to find a way to embedd IPFS within standalone .Net applications. My goal is to try and get it working with Window 10 store applications (the new deployment model. These apps run on desktop, tablets, Windows Phone, Hololens, etc). You can't just "run a go-ipfs daemon on the tablet".
<davidar>
DavidBurela (IRC): hence my question, does .net not have a foreign function interface?
<DavidBurela>
won't work within the new deployment model. Win 10 UWP (Universal Windows Platform).
<DavidBurela>
in .Net classic you can p/invoke .dlls and COM objects sure. But not in the new world.
<DavidBurela>
You can take raw C++ code and compile it in Visual Studio and that is a linkable assembly.
<davidar>
Didn't android already try that, and eventually had to relent? Restricting a platform to a handful of languages is problematic :/
<DavidBurela>
languages aren't a problem. Can write Win10 apps in JS, C#, F#, C++, ruby etc. etc. But it just needs to be compiled with the Win10 libraries.
<davidar>
DavidBurela (IRC): so, why not do that with go-ipfs?
<DavidBurela>
because they are go libraries... not .net libraries.
<Syntist>
can we create youtube type website in ipfs. i ma really new to this stuff :P
<DavidBurela>
Syntist, I think that is the most common question ;-) Yeah the videos themselves is easy, there are already cool demos of how IPFS can stream video. And you could create some static HTML pages with links to the videos... the problem is the dynamicness of it. How to update the page with new videos dynamically (dynamic content is the tricky bit)
<Syntist>
k
<DavidBurela>
davidar, My goal is to try and get IPFS implemented as a Win10 "App Service" https://msdn.microsoft.com/en-us/library/windows/apps/xaml/mt187314.aspx If we get it written once, then you have a base service that any app on Win10 can call into (either classic .Net or Win10 apps). Anyone will be able to Install the IPFS service from the store, and all store apps will be able to just depened on this store package, instead
<DavidBurela>
of needing to reimplementing it in every individual app.
<davidar>
DavidBurela (IRC): so, I guess that means you'll have to port go-ipfs (or libp2p) to whatever language .net libraries use?
<davidar>
DavidBurela (IRC): actually, it probably would make sense to port libp2p to more languages once it stabilises
<DavidBurela>
davidar, I think the best bet may be to find a Go -> C# converter, to continuously convert the go-ipfs code as it is checked in. And then reimplement the go libraries, but just as small wrappers that translate to the .Net constructs (networking stack, files).
xelra has quit [Remote host closed the connection]
locusf has quit [Ping timeout: 264 seconds]
<davidar>
DavidBurela (IRC): there used to be a CLR backend for LLVM a while ago, but I suspect it's long dead by now
ilyaigpetrov has quit [Remote host closed the connection]
Spinnaker has joined #ipfs
Guest18443 is now known as sasha
Syntist has quit [Quit: Page closed]
slothbag has quit [Quit: Leaving.]
locusf has joined #ipfs
akhavr has joined #ipfs
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 240 seconds]
akhavr1 is now known as akhavr
jhulten_ has joined #ipfs
akhavr1 has joined #ipfs
jhulten_ has quit [Ping timeout: 240 seconds]
akhavr has quit [Ping timeout: 264 seconds]
akhavr has joined #ipfs
azariah has joined #ipfs
akhavr1 has quit [Ping timeout: 272 seconds]
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 256 seconds]
akhavr has joined #ipfs
akhavr1 has quit [Ping timeout: 255 seconds]
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 264 seconds]
akhavr has joined #ipfs
<ion>
It might be useful to endeavor to create a portable C implementation for FFI purposes at some point.
akhavr1 has quit [Ping timeout: 268 seconds]
<davidar>
ion (IRC): yeah
<davidar>
And then a kernel module :)
* davidar
wonders if Linus would rage about that
<davidar>
Although there's already a 9p kernel module iirc
<davidar>
ion (IRC): I kind of feel like a c implementation + emscripten would actually make more sense than the js port, but whatever
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 256 seconds]
akhavr1 is now known as akhavr
<davidar>
.ask jbenet if there are plans for a c implementation of ipfs/libp2p at some point for maximal portability
<multivac>
davidar: I'll pass that on when jbenet is around.
ekaron has quit [Ping timeout: 244 seconds]
akhavr1 has joined #ipfs
<achin>
ion: it should be written in rust with a C API :)
<daviddias>
davidar: one of the goals of the JS implementation is to have it modular and have custom builds for different browser apps set ups
<davidar>
achin (IRC): yeah but you can't write kernel module in rust ;)
<achin>
why not?
akhavr has quit [Ping timeout: 255 seconds]
akhavr1 is now known as akhavr
<davidar>
daviddias (IRC): wouldn't it make sense for the core to be automatically translated from a common codebase though?
<davidar>
achin (IRC): same reason you can't use c++ I guess
Guest73396 has joined #ipfs
ekaron has joined #ipfs
<achin>
you just need to expose the right interface, right?
<daviddias>
we could do it for some parts, but libp2p multi protocol approach wouldn't make it a easy task to handle a mixed implementation
<achin>
someone wrote a simple rust kernel module using an ancient version of rust, using a small shim written in c. i admit i'm not familiar enough with kernel modules or stdlib-less rust to know why this is necessary
<achin>
in any case, a small shim doesn't seem like a bad price to pay for getting to use a nice language
<davidar>
daviddias (IRC): fair enough
<richardlitt>
Sprint people! We need your spints for this week. Please add them to https://etherpad-mozilla.org/sprint-36. That includes @mappum @whyrusleeping @jbenet @daviddias @davidar @lgierth @amstocker @noffle and @rht
voxelot has joined #ipfs
<davidar>
achin (IRC): shims are generally frowned upon, cf nvidia drivers
<richardlitt>
spints >> sprints
<achin>
sorry, i didn't understand that. what about the nvidia drivers?
<DavidBurela>
daviddias, but wouldn't a ASM/WASM implementation make more sense than JS? Compiled JS, native like performance. New browsers are all supporting it
captain_morgan has quit [Ping timeout: 240 seconds]
akhavr has quit [Ping timeout: 246 seconds]
akhavr has joined #ipfs
Erkan_Yilmaz has quit [Quit: 194 people are helping our research, how about you? ( info in channel #musicology )]
akhavr has quit [Ping timeout: 240 seconds]
haadc has joined #ipfs
akhavr has joined #ipfs
haadc has quit [Client Quit]
haadcode has joined #ipfs
akhavr1 has joined #ipfs
sseagull has joined #ipfs
akhavr has quit [Ping timeout: 250 seconds]
akhavr1 is now known as akhavr
<vijayee_>
whyrusleeping: so..... what is a peerstore and...what is its purpose?
<vijayee_>
I assume this is not the same as a blockstore and is not shared over the network
<achin>
maybe is stores information about your peers?
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 246 seconds]
<gamemanj>
this is weird... the "ipfs object patch set-data" example shows using stdin, but that doesn't seem to be possible?
akhavr has joined #ipfs
<daviddias>
DavidBurela: those are some good valid points and were definitely taken into account. Initially we even looked at gopherJS as a quick way to have a IPFS.js running, but the outcomes weren't as trivial
<achin>
i've used it successfully
akhavr1 has quit [Ping timeout: 240 seconds]
<gamemanj>
Oh, apparently it only complains about args if I try to get it to use the console...
<daviddias>
although producing a code base in C once and porting it everywhere would mean that lesser lines of code would be developed and we would have feature parity at all times, in reality, what happens is that our dev cycle gets slower, C is known to be a less productive language than Go or JS, we would increase greatly the friction for new contributors
<gamemanj>
achin: Ah, thanks. Though why doesn't it like using the console directly...?
<achin>
sorry, i'm not sure what you mean by the console
<daviddias>
and these are some of the aspects we have to take into account when developing IPFS
<gamemanj>
Well, if I do: cat | ipfs object patch QmdfTbBqBPQ7VNxZEYEj14VmRuZBkqFbiwReogJgS1zR1n set-data
<achin>
gamemanj: that seems to work fine for me
<gamemanj>
then I can type in the desired data and then do Ctrl-D to EOF.
<daviddias>
nevertheless, both IPFS and libp2p are protocols, not language libraries, everyone can push for a new impl in another language
<gamemanj>
But if I try doing it directly, simply by having no parameter,
<gamemanj>
without the "cat |"
<gamemanj>
then it doesn't work and it wants an arg.
akhavr1 has joined #ipfs
<gamemanj>
It's a bit confusing.
<achin>
oh, you mean by just running "ipfs object patch QmdfTbBqBPQ7VNxZEYEj14VmRuZBkqFbiwReogJgS1zR1n set-data" ?
<gamemanj>
Yes.
<achin>
with no redirection or anything?
<gamemanj>
Yes.
<daviddias>
for example, I'm very interested personally in making it in Rust, but right now I'm prioritising the interop between go and JS
<achin>
hmm, i'm tempted to say that is expected? since i think your shell will have started ipfs with no stdin handle?
akhavr has quit [Ping timeout: 246 seconds]
akhavr1 is now known as akhavr
<dignifiedquire>
daviddias: I just pushed a fix, so now file uploading is working again, just no UI notifications yet :)
<achin>
but i do see why that might be confusing
<daviddias>
dignifiedquire: cool! :)
voxelot has joined #ipfs
voxelot has joined #ipfs
bsm1175321 has joined #ipfs
<daviddias>
what about the beep? I liked the beep :)
<dignifiedquire>
I didn’t touch the beep, it should still be there
<daviddias>
nice!
<dignifiedquire>
but we need to figure out a better way for the clipboard, cause if you upload multiple files you only get the url to the last file
<gamemanj>
achin: The shell I'm using is ordinary bash, so the stdin is available and set to the terminal...
<achin>
daviddias: if you do end up doing anything with rust, please loop me in?
<daviddias>
achin: noted! :)
<daviddias>
wanna help me push things on JS and Go sides so we can jump into Rust land? :D
<achin>
i'd need to learn go first!
<achin>
gamemanj: yeah, i dunno then. a bug, maybe!
<dignifiedquire>
dignifiedquire: :D sorry that is just an the jellyfish in the image being transparently overlayed with the menubar
<blame>
SWingedSeraph: I just commented on your github comment and I am here to argue if you want
<blame>
*issue
<gamemanj>
achin: It seems it's not the only command which supports stdin if and only if it's not a terminal, though. It might be defined behavior.
<achin>
(if anyone else in here is interested in rust and ipfs, i've just shared my scratchpad of IPFS-related code at https://github.com/eminence/ipfsrs )
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 240 seconds]
akhavr1 is now known as akhavr
<achin>
gamemanj: since most IPFS data is binary, maybe the ipfs team thought it rare for someone to manually type in data via a terminal
<daviddias>
whyrusleeping: around? :)
<daviddias>
dignifiedquire: ahaha got it. the jelly fish idea was top!
<gamemanj>
achin: It seems to be consistent behavior for the last arg to be set via stdin if, and only if, a terminal is being used...
<dignifiedquire>
thanks :) whyrusleeping seems to be very fond of it as well if I’m reading his comment right
akhavr1 has joined #ipfs
<gamemanj>
s/is being used/is not being used/
<multivac>
gamemanj meant to say: achin: It seems to be consistent behavior for the last arg to be set via stdin if, and only if, a terminal is not being used...
<dignifiedquire>
daviddias: yes they are awesome, but I was luck to find that image in high quality plus super permissive licensing
<dignifiedquire>
*lucky
<gamemanj>
Oh, wait, ipfs with no args does help by default, there goes that test
<daviddias>
dignifiedquire: he is a fan of jellyfishes (but I intent to convert him to the cuttlefish church when he visits me in Portugal :P)
<dignifiedquire>
:D
akhavr has quit [Ping timeout: 240 seconds]
akhavr1 is now known as akhavr
<whyrusleeping>
dignifiedquire: <3 jellyfish
<whyrusleeping>
daviddias: i'm aroundish
<whyrusleeping>
attempting to flash my phone to android m
<gamemanj>
This is quite confusing... I thought that there was some consistency in how the magical stdin arg-adding behavior works...
<daviddias>
dignifiedquire: haven't seen that one by doing ctrl+c, but it seems that the IPFS node is being killed first (which should be ok) and the electron app is still trying to contact it
<dignifiedquire>
it’s a bit annoying because there about 10 of these dialogs :( and I have to click ok for each of them
<dignifiedquire>
but I’m not sure where to look to handle it
akhavr has quit [Ping timeout: 246 seconds]
akhavr has joined #ipfs
Not_ has quit [Remote host closed the connection]
Not_ has joined #ipfs
akhavr1 has joined #ipfs
Tv` has joined #ipfs
akhavr has quit [Ping timeout: 264 seconds]
<ipfsbot>
[go-ipfs] whyrusleeping force-pushed fix/too-many-fd from 82084fe to 2787ba0: http://git.io/vcFSw
<daviddias>
nice, I'm curious how are you holding the tray open ??
<whyrusleeping>
dignifiedquire: youre blowing my mind
<dignifiedquire>
atm I have enabled ‘always-on-top’: true in the config
<dignifiedquire>
that way it only closes if you click the icon a second time
pfraze has joined #ipfs
<daviddias>
cool, didn't know of that setting! :)
<gorhDroid>
hi hi daviddias dignifiedquire :)
<daviddias>
hi gorhDroid :)
<dignifiedquire>
hi gorhDroid
<dignifiedquire>
whyrusleeping: always happy to blow minds :)
<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to real-trailers: http://git.io/vcFxY
<ipfsbot>
go-ipfs/real-trailers b548449 Jeromy: update go version requirement in readme...
__uguu__ has joined #ipfs
akhavr1 has joined #ipfs
adrusi has joined #ipfs
akhavr has quit [Ping timeout: 252 seconds]
akhavr1 is now known as akhavr
Not_ has quit [Remote host closed the connection]
akhavr1 has joined #ipfs
jimbeam has quit [Ping timeout: 244 seconds]
akhavr has quit [Ping timeout: 240 seconds]
akhavr1 is now known as akhavr
vijayee_ has joined #ipfs
Not_ has joined #ipfs
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 240 seconds]
akhavr1 is now known as akhavr
vijayee_ has quit [Client Quit]
nicolagreco has quit [Quit: nicolagreco]
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 268 seconds]
akhavr1 is now known as akhavr
NeoTeo has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
Henry_ has joined #ipfs
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 268 seconds]
gorhDroid is now known as gorhgorh
akhavr has joined #ipfs
akhavr1 has quit [Ping timeout: 244 seconds]
akhavr1 has joined #ipfs
NeoTeo has joined #ipfs
akhavr has quit [Ping timeout: 260 seconds]
akhavr1 is now known as akhavr
Spinnaker has quit [Ping timeout: 256 seconds]
Spinnaker has joined #ipfs
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 240 seconds]
ygrek has joined #ipfs
<daviddias>
dignifiedquire: I personally can't wait for that PR to be merged so that I start sending the install url to all the people ! :D
akhavr has joined #ipfs
<dignifiedquire>
yeah I can’t wait to finally tell people to use it either :)
<dignifiedquire>
and then get onto the webui
<dignifiedquire>
I have great plans for it ;)
<whyrusleeping>
dignifiedquire: i cant wait for the webui to be udpated
<whyrusleeping>
its been sooooo long
akhavr1 has quit [Ping timeout: 250 seconds]
<giodamelio>
Who's in charge of the webui? I was thinking that might be a good place for me to help out. My js is way better then my go
sonatagreen has joined #ipfs
<whyrusleeping>
giodamelio: daviddias and dignifiedquire are the people i would say are in charge right now
<daviddias>
whyrusleeping: for the webui specifically I believe that would be krl
<daviddias>
but I can shine some light there too
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 240 seconds]
akhavr1 is now known as akhavr
<giodamelio>
Is there any easy things I could get started on to get used to the codebase?
jimbeam has joined #ipfs
Henry_ has quit [Remote host closed the connection]
akhavr1 has joined #ipfs
<locusf>
hmm pulsar has been open sourced
akhavr has quit [Ping timeout: 250 seconds]
akhavr1 is now known as akhavr
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 246 seconds]
jimbeam has quit [Ping timeout: 272 seconds]
<dignifiedquire>
giodamelio: I haven’t touch the code yet, but I do plan to draw up a list of things to do next week or so when I have some initial mockups
nikor has joined #ipfs
<dignifiedquire>
giodamelio: also if you want to prepare, knowledge about es2015 and react are needed
akhavr has joined #ipfs
akhavr1 has quit [Ping timeout: 240 seconds]
<giodamelio>
I am pretty strong with both ES2015 and React. Most of my latest projects have been webpack + babel + react.
akhavr has quit [Ping timeout: 240 seconds]
<dignifiedquire>
giodamelio: that’s good to hear
<giodamelio>
Are you thinking of switching to ES2015?
zero-ghost has quit [Read error: Connection reset by peer]
<dignifiedquire>
babel stage0 + react + webpack
<dignifiedquire>
webui is already using babel and react, but the setup needs some improvements to get all the goodness of hot reloading
akhavr1 has joined #ipfs
<giodamelio>
Sounds good. How would you feel about me putting togather a pull request converting the build system to webpack? Just a basic setup to make way for hot reloading and stuff.
<dignifiedquire>
sounds good to me, any objections daviddias, jbenet ?
akhavr has quit [Ping timeout: 240 seconds]
akhavr1 is now known as akhavr
zero-ghost has joined #ipfs
<locusf>
I'd love to make the babelian link backend
<jbenet>
IIRC we spent time going to webpack and then back to browserify. Can look in irc logs for details. (I would Really love to avoid stupid tool wars -- and there's people for both sides of the webpack/browserify debate)
<multivac>
jbenet: 2015-10-06 - 12:52:46 <davidar> ask jbenet if there are plans for a c implementation of ipfs/libp2p at some point for maximal portability
<jbenet>
multivac tell davidar yes, and go itself will be able to meanwhile
<multivac>
jbenet: I'll pass that on when davidar is around.
<jbenet>
Ok, gotta jet.
* jbenet
puts on sunglasses (deal with it) and takes off on a jet plane.
<dignifiedquire>
jbenet: no toolwar from my side, just that the tooling for react + webpack is better in terms of plugins and dev friendliness than with browserify in my experience, which is why I opted for webpack in electron-app. I think it would be good to have the same/similar setup for the ui apps as much as possible, so my thinking was, to do a similar setup as I did for electron-app for the webui
<jbenet>
daviddias see you soon, I get into airport ~13:00
Jos_ has joined #ipfs
<jbenet>
dignifiedquire: sgtm, I really don't care about this one. Both work well for me
<jbenet>
(Just brace for eventual messages from others asking to move it back or whatever)
<dignifiedquire>
(bracing for playing a hard defense )
<jbenet>
Btw dignifiedquire would be awesome to have upload progress bars on the dropdown menu (gj on getting that working!)
<dignifiedquire>
giodamelio: you heard it ;) if you could try to keep as close to the setup I did in https://github.com/ipfs/electron-app/pull/43 that would be great, maybe we can even extract common parts into a seperate repo at some point
<dignifiedquire>
jbenet: file upload is pretty much all working in the latest version, even with system notifications
<dignifiedquire>
just no progressbar atm I know, but not sure how to do one that is actually usefull as there are no progress events coming from the api as far as I can tell atm
<locusf>
so is that screenshot of yours of a desktop tool for ipfs?
<giodamelio>
dignifiedquire: Ok, I'll get too it. Maybe at somepoint there should be a organized place to keep track of build systems, style guides, etx across all the js repos?
<dignifiedquire>
locusf that is the live running version from this pr, you can download it and use it
<dignifiedquire>
giodamelio: yes we talked about sth like this, jbenet can you remind me of the repo where stuff like this was supposed to be collected?
<alhof>
whyrusleeping thanks, i'll look into that. it seems like a cool project.
akhavr1 has quit [Ping timeout: 256 seconds]
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 240 seconds]
akhavr1 is now known as akhavr
NeoTeo has joined #ipfs
<dignifiedquire>
locusf: did you solve the issue you linked to?
<locusf>
dignifiedquire: nope
<dignifiedquire>
do you have ipfs installed/running in another way?
<locusf>
nope
<dignifiedquire>
and are you running on my fork or on the master branch of the offical version?
<dignifiedquire>
and what node version are you running on?
<locusf>
nodejs-4.1.2-1 npm-3.3.5-1
<locusf>
and its master
<dignifiedquire>
hmm, try my fork, I just tested it and it worked fine, though you’ll need node 0.12 atm for that
jimbeam has joined #ipfs
akhavr1 has joined #ipfs
pfraze_ is now known as pfraze
akhavr has quit [Ping timeout: 260 seconds]
<locusf>
dignifiedquire: ok, will install nodejs-git package then
akhavr has joined #ipfs
<nagydani>
Still a bit confused. Is the directory object simply a flat list of links as defined in merkledag.proto?
akhavr1 has quit [Ping timeout: 264 seconds]
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 250 seconds]
akhavr has joined #ipfs
akhavr1 has quit [Ping timeout: 240 seconds]
<achin>
not quite
<achin>
it's that, plus a unixfs object in the PBNode.Data field. the unixfs object has Type=Directory
rendar has quit [Ping timeout: 250 seconds]
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 244 seconds]
akhavr1 is now known as akhavr
Eudaimonstro has quit [Remote host closed the connection]
<locusf>
are you guys going to give talks at FOSDEM?
<achin>
(a merkledag with a links could also be a big file that's been chunked. in this case, the unixfs object will have type=File and a "blocksizes" entry for each link)
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 250 seconds]
<achin>
((oops, instead of merkeldag i should have said PBNode))
<nagydani>
achin, how about one directory with lots (e.g. tens of thousands) of files?
rendar has joined #ipfs
<achin>
as far as i know, that has to be 1 single [large] PBNode
akhavr has joined #ipfs
<nagydani>
Are files listed in lexicographic order of their names?
<achin>
but it's possible that you might be able to "chunk" the links by linking to other unixfs objects, but ensuring that the links don't have names. i've not tried this, though
<achin>
i'm not sure, actually (about the ordering)
akhavr1 has quit [Ping timeout: 246 seconds]
<achin>
it depends on how ipfs constructs the PBNode object, and i've not looked to see if it does any ordering
<achin>
s/ordering/sorting/
<multivac>
achin meant to say: it depends on how ipfs constructs the PBNode object, and i've not looked to see if it does any sorting
<achin>
thank you robot, but please shutup
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 250 seconds]
akhavr1 is now known as akhavr
akhavr has quit [Ping timeout: 255 seconds]
akhavr has joined #ipfs
<nagydani>
So, does this mean that
<nagydani>
a) lookups in a long directory take a linear amount of time
<nagydani>
b) changing, adding or deleting one file in a long directory requires the re-hashing of the whole thing?
<achin>
yes, running "ipfs refs" on a big directory will require downloading an object that has a size proportional to the number of objects within the directory
<achin>
and yes, if you want to change a directory and then make the changed version available on IPFS, you have to rehash it to add the directory
<nagydani>
That's unfortunate.
akhavr1 has joined #ipfs
<nagydani>
Thank you!
<achin>
but for child objects (files/folders within that directory) that are unchanged, they will keep their same hash (so the total size of objects will not increase by that much)
akhavr has quit [Ping timeout: 264 seconds]
akhavr1 is now known as akhavr
alhof has quit [Remote host closed the connection]
jhulten_ has quit [Ping timeout: 252 seconds]
alhof has joined #ipfs
<achin>
"ipfs object patch rm-link" might also work in this case
akhavr1 has joined #ipfs
<achin>
and would be a more efficient way to update a directory, if you know exactly the file that was removed
Wolf481pl has joined #ipfs
akhavr has quit [Ping timeout: 268 seconds]
akhavr1 is now known as akhavr
<Wolf481pl>
Hello. Is there a way to optimize the splitting algorithm for thousands of small files that differ from each other by 1 or 2 characters?
<Wolf481pl>
(ok, maybe not 1 or 2 characters, but 1 or 2 lines)
<daviddias>
giodamelio: apologies for leaving you hanging, I was reviewing and taking notes of what could be the good places to start and then got into a f2f conversation
<daviddias>
another big thing we really need, is putting tests on WebUI so that we can bring more features with confidence https://github.com/ipfs/webui/issues/38
<achin>
Wolf480pl: yes, i think there could be lots of clever ways to chunk files
<achin>
that would be a neat research project
<Wolf480pl>
if so, do you think it would allow storing loose git objects as they are stored on disk with git packs?
<Wolf480pl>
*if so, do you think it would allow storing loose git objects *as efficiently* as they are stored on disk with git packs?
<achin>
i admit i don't know a lot of the technical details of git's packfiles, but i suspect the theory could be applied to ipfs object files
<Wolf480pl>
well, in git, a packfile is a bunch of objects stored together delta-encoded and compressed
<Wolf480pl>
but then you need an index for each pack to figure out in which pack the object is
<achin>
git likes to have its objects ordered by date (i think it assumes some stuff about how source files change ove time), but i'm not sure in ipfs we'd have that
<giodamelio>
daviddias: I just started converting it to webpack. Should I contiue with that? I dont want to mess up any work done in https://github.com/ipfs/webui/pull/73
akhavr1 has joined #ipfs
<achin>
Wolf480pl: are you only talking about how ipfs stores blocks on disk?
<Wolf480pl>
no, I'm not concerned with the on-disk storage
<achin>
ok, so you also care about how data it sent to/from nodes
<Wolf480pl>
I mainly care how it's send to/from nodes
akhavr has quit [Ping timeout: 268 seconds]
akhavr1 is now known as akhavr
nicolagreco has joined #ipfs
<dignifiedquire>
daviddias: why webcomponets and not react components if it’s written in react?
<Wolf480pl>
also, would it be possible to implement something like git smart transport over ipfs ?
<achin>
i don't know the IPFS protocol on the wire, but this sounds like something that is totally doable
<Wolf480pl>
(I haven't read much of the docs, nor the whitepaper)
pfraze_ has joined #ipfs
<achin>
this might actually relate to that wantlist comment ion linked a few moments ago
alhof has quit [Quit: Be back later ...]
devbug has quit [Ping timeout: 240 seconds]
<achin>
(under the general category of: "make ipfs smarter about moving bits around")
<daviddias>
dignifiedquire we picked WebComponents for the encapsulation because can make IPFS served WebComponents and since they are a Web Standard, every browser will know how to use them without requiring any other library, independently if they use React inside or anything else
voxelot has quit [Ping timeout: 255 seconds]
hellertime has quit [Quit: Leaving.]
<Wolf480pl>
achin, git smart transport is basically "I want sha1s A, B, and C, but I already have D", "ok, I'll pack them for you and delta-encode them against what you have, and send"
pfraze has quit [Ping timeout: 255 seconds]
<dignifiedquire>
daviddias: I see
jimbeam has quit [Ping timeout: 246 seconds]
<achin>
Wolf480pl: i wonder what to do about the fact that ipfs has many many more roots (and smaller trees) when compared to git repos
<daviddias>
giodamelio could you open a issue and make a case of webpack vs browserify scripts? It would be the best place to get the whole group on board with the idea
akhavr1 has joined #ipfs
nagydani has quit [Ping timeout: 240 seconds]
akhavr has quit [Ping timeout: 246 seconds]
akhavr1 is now known as akhavr
<daviddias>
instead of just seeing it came out of the blue. Probably it will even be orders of magnitude better and it's great if we can all understand it first
<giodamelio>
daviddias: Sure
<Wolf480pl>
achin, what's wrong with it?
jhulten_ has joined #ipfs
<ion>
Wolf480pl: Delta encoding by a smart sender is less useful when all objects are chunked according to a rolling hash algorithm. go-ipfs will do that soon AFAIU.
<daviddias>
dignifiedquire: also moved with electron-app to use WebPack, we could move the whole team to start using across all the projects :)
<dignifiedquire>
daviddias: yes, I was trying to setup something we can use everywhere (maybe even a shared code base that pulls in defaults)
<Wolf480pl>
ion, so if we have a rolling hash algo optimized for small, very similar objects, we can store loose git objects in IPFS and not worry about anything?
nagydani has joined #ipfs
nagydani has left #ipfs [#ipfs]
nagydani has joined #ipfs
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 250 seconds]
akhavr1 is now known as akhavr
<achin>
Wolf480pl: "i have SHA D" might mean "here is my latest maste branch", so there could be a whole lot of history (aka objects) behind SHA D. but with ipfs, a lot of the time you'll be asking to download a root, in which case it's not clear to me what hashes you should claim to have
akhavr1 has joined #ipfs
<ion>
Wolf480pl: I suppose, assuming we chunk the git objects when importing.
<Wolf480pl>
achin, but I'm gonna keep a git DAG in IPFS, not a .git directory
Soft has quit [Read error: Connection reset by peer]
akhavr1 has quit [Ping timeout: 240 seconds]
voxelot has joined #ipfs
voxelot has joined #ipfs
Soft has joined #ipfs
zml_ is now known as zml
<achin>
i would love for ipfs to be well suited for git without the need for special cases (but im' not sure if that is possible)
akhavr has quit [Ping timeout: 268 seconds]
akhavr has joined #ipfs
legobanana has joined #ipfs
pfraze has joined #ipfs
akhavr1 has joined #ipfs
akhavr has quit [Ping timeout: 268 seconds]
akhavr1 is now known as akhavr
pfraze_ has quit [Ping timeout: 252 seconds]
<ipfsbot>
[go-ipfs] ChristianKniep opened pull request #1805: [1804] change Dockerfile order to create VOLUME under user ipfs (master...1804_ipfs_volume) http://git.io/vcN09
akhavr1 has joined #ipfs
gamemanj has joined #ipfs
akhavr has quit [Ping timeout: 250 seconds]
akhavr1 is now known as akhavr
<ipfsbot>
[go-ipfs] ChristianKniep closed pull request #1805: [1804] change Dockerfile order to create VOLUME under user ipfs (master...1804_ipfs_volume) http://git.io/vcN09
<ipfsbot>
[go-ipfs] ChristianKniep opened pull request #1806: change Dockerfile order to create VOLUME under user ipfs (master...1804_ipfs_volume) http://git.io/vcNup