<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to master: http://git.io/vCpGP
<ipfsbot>
go-ipfs/master 2afe4a4 Jeromy: ipfs version 0.3.8 changelog and version bump...
<whyrusleeping>
chooo choooo!
<ipfsbot>
[go-ipfs] whyrusleeping tagged v0.3.8 at master: http://git.io/vCpGy
Guest73396 has quit [Ping timeout: 252 seconds]
<ipfsbot>
[go-ipfs] whyrusleeping deleted feat/sysdiag at 1cef170: http://git.io/vCpZI
Spinnaker has quit [Ping timeout: 256 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Spinnaker has joined #ipfs
<achin>
is that the sound of the progress train?
Guest73396 has joined #ipfs
<whyrusleeping>
achin: that it is!
<whyrusleeping>
0.3.8 is shipped!
<achin>
\o/
<achin>
do you have one of those dancing robots in here?
<whyrusleeping>
lol, no
<achin>
something to be rectified!
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
dts has joined #ipfs
dts has quit [Ping timeout: 272 seconds]
pfraze has joined #ipfs
<gendale_>
woahhh congrats!
Guest73396_m has joined #ipfs
Guest73396 has quit [Ping timeout: 252 seconds]
<ipfsbot>
[go-ipfs] whyrusleeping merged master into release: http://git.io/vCp4Y
notduncansmith has joined #ipfs
apophis has quit [Quit: This computer has gone to sleep]
notduncansmith has quit [Read error: Connection reset by peer]
<alu>
whose been playing with docker ?
<whyrusleeping>
me
od1n has joined #ipfs
<whyrusleeping>
why?
<alu>
I'm dockerizing the janus multiserver and running into problems where the node application keeps exiting, can you take alook at the dockerfile and see?
<whyrusleeping>
alu: try ENTRYPOINT instead of CMD
<whyrusleeping>
also, note that the 'RUN' directives are run at build time
<whyrusleeping>
not at container start
<alu>
did I butcher it
<alu>
do I get a D- ?
<alu>
as long as I pass...
<whyrusleeping>
lol
domanic_ has joined #ipfs
<whyrusleeping>
i would do the 'sh generate_key' part after the container starts
<whyrusleeping>
unless you want them all to have the same key
<whyrusleeping>
i generally make a shell script to be my 'entrypoint'
<alu>
then just ADD ?
<whyrusleeping>
yeah, i ADD the shell script
<whyrusleeping>
and then set my entrypoint to it
<alu>
add to like /tmp/
<alu>
delete after its done?
<whyrusleeping>
no, that doesnt matter
<whyrusleeping>
you just dont want your keys hard coded into the image
<alu>
i thought about that a lot...
<alu>
im highly security conscious
<alu>
thanks for your feedback ^^
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<alu>
whyrusleeping have you setup the mysql container before?
<whyrusleeping>
nope
ketohexose has quit [Remote host closed the connection]
PsuzeannePsy has joined #ipfs
Suzeanne has joined #ipfs
Bat`O has quit [Ping timeout: 252 seconds]
Bat`O has joined #ipfs
<alu>
whyrusleeping I built it and when I ran "docker run -it --rm --name avalon-n0de alusion/avalon-node" it exited again
SuzieQueue has quit [Ping timeout: 268 seconds]
PsuzeannePsy has quit [Ping timeout: 260 seconds]
<alu>
whyrusleeping: after this step I plan to implement kubernetes in a system for potentially scaling to thousands or a million simultaneous clients in teh same virtual reality shared space ^^;
<whyrusleeping>
alu: why are you running it with -it?
<alu>
¯\_(ツ)_/¯
<alu>
sec
<alu>
i didnt even specify open port
starver has joined #ipfs
<whyrusleeping>
does it not print logs anywhere?
<alu>
thatdocker run -p 5566:5566 -p 8080:8080 alusion/avalon-node:latest
<alu>
Actually I just ran docker run -p 5566:5566 alusion/avalon-node:latest
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<alu>
docker ps shows nothing but ill look at docker ps -a
<alu>
and inspect
<alu>
what was the thing to see container logs
<alu>
docker logs containerid doesn't say much but
<alu>
boo not even a hostip
<whyrusleeping>
alu: lol, it shouldnt be that hard
<whyrusleeping>
is your image pushed to docker hub?
<alu>
n-no >.>
<alu>
its all on gitlab
<whyrusleeping>
lol, okay
<whyrusleeping>
no worries
<whyrusleeping>
(pushing to dockerhub is NOT required)
* alu
crashed through docker with engines still burning
M-erikj has quit [Ping timeout: 240 seconds]
M-erikj has joined #ipfs
<alu>
right now at this coffeeshop i feel like the embodiment of a rabbit hole
<whyrusleeping>
haha
<alu>
if someone had enough interest to ask me what it is i do
<whyrusleeping>
i went pretty far down the docker rabbit hole last week
<alu>
im prepared to change their life forever in a minutes
<whyrusleeping>
ended up filing a PR against their networking lib
<alu>
i got my gearvr in my backpack lol
<whyrusleeping>
haha, nice
<stoopkid>
alu: lol
<alu>
the server.js runs but its just exits afterthat
<whyrusleeping>
does server.js normally exit? or does it run until you kill it?
<alu>
yes it normally just runs in background like a daemon
<whyrusleeping>
oh, thats why
<alu>
then you just like tail the server log
<whyrusleeping>
docker exits when your entry command does
<alu>
tail -f server.log
<alu>
oh I see, can I fit tail -f server.log then at the end?
<whyrusleeping>
sure
<whyrusleeping>
you'd have to make it a script
<alu>
last time it freaked when i gave it the flag
<whyrusleeping>
and run the script
<alu>
I ought to convert some these procedures into scripts
M-erikj has quit [Ping timeout: 260 seconds]
M-erikj has joined #ipfs
<whyrusleeping>
scripts are generally easier to test and verify :)
M-rschulman1 has quit [Ping timeout: 240 seconds]
M-rschulman1 has joined #ipfs
<alu>
hmm I'm thinkking about how it generates that key, perhaps I can fill the parameters with variables in the boot script?
<whyrusleeping>
nah, just put it in the same script you use to run the server and tail the logs
<whyrusleeping>
kinda like how ipfs's docker image runs init when you start the docker container
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
od1n has quit [Quit: This computer has gone to sleep]
* zignig
is going to have to upgrade , like..like , four computers.
<whyrusleeping>
zignig: i had a file in there that would cause the builds to fail and print out a message saying you needed go1.5
* zignig
calls the waaaambulance.
<whyrusleeping>
but it broke gobuilder, because gobuilder wants to format the code before building it
<zignig>
all good ;) , is the older client still going to work with 0.3.8 ?
<whyrusleeping>
yeap
<whyrusleeping>
they interop just fine
<zignig>
my understanding is that 0.4 is going to BORK all the things.
<whyrusleeping>
yeah, 0.4.0 will break all the things
<whyrusleeping>
which is why its taking so long
<whyrusleeping>
we want to break as many things as we can all at once
<whyrusleeping>
instead of incrementally breaking things
<zignig>
utp + new wire protocol + awesome ?
<tinybike>
alright so i've got my noob hat on here. I'm familiar with DAG in the context of git and bitcoin (where the DAG is for commits and block hashes, respectively) -- this is the DAG of a file?
<whyrusleeping>
utp will land before 0.4.0 actually
<tinybike>
how does that work? or should I just read the code and find out
<whyrusleeping>
tinybike: yeap! ipfs files are represented as DAGs
<tinybike>
not changes in files, though, the file itself? how is a file represented as a DAG?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<zignig>
tinybike: yes , it's just breaking things up into blocks and constructing the DAG to get you the files bak.
<whyrusleeping>
its data is broken up into chunks, and each chunk is put in the data of a dag node
<tinybike>
ohhh I get it
<whyrusleeping>
and then a 'tree-like' structure (hey! a dag!) is made from them
<tinybike>
and the hash I'm seeing is actually the root hash of the dag
<zignig>
whyrusleeping: there seems to be a regression on ipfs ls.
<zignig>
with a big dag it can take ages to resolve.
<zignig>
where ipfs object stats , goes really fast.
<tinybike>
is the hash used (by default) sha256?
<whyrusleeping>
zignig: 'ipfs ls' fetches each child node
<whyrusleeping>
tinybike: yeap
<zignig>
whyrusleeping: recursive ?
<zignig>
and why ?
<whyrusleeping>
zignig: not recursive, just one layer for type checking
<tinybike>
ok so I have a directory hash here: QmaUJ4XspR3XhQ4fsjmqHSkkTHYiTJigKZSPa8i4xgVuAt. "Qm" is just a special IPFS prefix, right?
<whyrusleeping>
tinybike: Qm means sha256
<tinybike>
Got it
<zignig>
ah, still takes ages to resolve, even with a fairly small number of top level files.
<whyrusleeping>
zignig: hrmmm
<tinybike>
and it's base58 encoded?
<whyrusleeping>
tinybike: yep!
<tinybike>
hmm...when I decode it (without the Qm prefix) I get 33 bytes
<tinybike>
extra byte i shake my fist at you
<whyrusleeping>
yeah, theres a little extra data thats part of the sha256 tag
<tinybike>
ah ok -- i don't know base58 encoding too well, is it like base64 where each character is actually 6 bits?
<whyrusleeping>
tinybike: yeah... the number of bits to characters isnt an even number
domanic_ has quit [Ping timeout: 240 seconds]
<zignig>
whyrusleeping: also make sure that the utp interface is replaceable with a pure go impl in the fullness of time.
<whyrusleeping>
the utp code is pure go
<tinybike>
ok... dumb question: if I want to extract the raw 32 byte array from the "Qm..." hash...how do I do that?
<zignig>
I thought it was cgo , or did that change
M-rschulman has quit [Ping timeout: 240 seconds]
<zignig>
?
<whyrusleeping>
tinybike: base58 decode the entire thing
<whyrusleeping>
then take the last 32 bytes
<tinybike>
oh
<tinybike>
lol
<whyrusleeping>
zignig: that was the UDT lib
M-rschulman has joined #ipfs
<tinybike>
ah I see...the leading byte is just a 1
<tinybike>
:D
<tinybike>
tyvm
<gendale_>
does 0.38 enable publishing to multiple ipns hashes?
<tinybike>
another (unrelated) question: is ipns publish/resolve really slow for everyone else?
stoopkid has quit [Quit: leaving]
<whyrusleeping>
gendale_: nope, that wont happen until after 0.4.0
<whyrusleeping>
tinybike: its slow right now, its needs some attention. but it now provides better guarantees than it used to
<whyrusleeping>
your record will stay on the record (barring really bad churn) until you take your node offline for a while
<whyrusleeping>
that second record should be network
<gendale_>
whyrusleeping: gotcha. the documentation used to say:
<gendale_>
" Publish an <ipfs-path> to another public key: (not supported yet)"
<gendale_>
or something like that, but it doesn't seem to say that anymore
<whyrusleeping>
gendale_: oh, i think i just removed that altogether so people would stop getting the 'error: keychains not yet implemented' thing
* zignig
like removing doc to stop complaints.
<whyrusleeping>
its still going to happen, just not quite yet
<zignig>
whyrusleeping: that will be cool , keychainy things make the mutable records much niftier
<whyrusleeping>
yeap! very much agreed
domanic_ has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
hjoest has quit [Ping timeout: 260 seconds]
hjoest has joined #ipfs
Suzeanne has quit [Quit: Leaving]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<alu>
whyrusleeping I'm kinda confused about where to put tail -f server.log so that I can still build the image. The shell script does npm install, generate_key, and starts the node.js service. It builds successfully, but tailing the server.log will keep it open
<whyrusleeping>
ENTRYPOINT or CMD runs the script
<whyrusleeping>
those get run when you run docker run
notduncansmith has quit [Read error: Connection reset by peer]
M-mistake has joined #ipfs
<dignifiedquire>
VictorBjelkholm: I know how you could solve it with karma, you can tell karma to serve arbitrary files and not include them, so that you can load them via simple http request, see http://karma-runner.github.io/0.13/config/files.html for more details on how to configure that. Then you could write a helper method, `getFile` which detects if it's run in the browser and then uses an http request to fetch the file or if it's run under
<dignifiedquire>
and fetch the file va fs.readFile
<VictorBjelkholm>
dignifiedquire, yeah, that's why I went with testem to start with, it requires no setup from my side, simply point it to the file and create the test page and everything
<VictorBjelkholm>
ooh
<VictorBjelkholm>
no, wait. I misunderstood you, you're talking about the fs.readFile and such...
thelinuxkid has joined #ipfs
<VictorBjelkholm>
The problem is that the browsers file api looks different from node, and node-ipfs-api allows adding files with paths and stuff, which is not supported in the browser. Take a look at the test/test.js file and you'll see :)
<dignifiedquire>
right but how is this supposed to work in the browser in the first place?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
voxelot has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<davidar>
VictorBjelkholm (IRC): yeah, I don't think node-ipfs-api is intended to run in the browser, but node-ipfs is eventually
rand_me has quit [Ping timeout: 240 seconds]
<davidar>
*AFAIK
<VictorBjelkholm>
davidar, I'm pretty sure node-ipfs-api is supposed to be able to run in the browser, at least according to the readme in the repo
<VictorBjelkholm>
+ daviddias have been talking about getting the tests to run in the browsers for a while
<VictorBjelkholm>
dignifiedquire, some things are not gonna be able to work in the browser, unless we provide shims or something like that
<VictorBjelkholm>
let's get the tests running properly and then it'll be easy to think about/fix
<VictorBjelkholm>
daviddias, dignifiedquire created a todo in the pr (about the tests, https://github.com/ipfs/node-ipfs-api/pull/77) about what's left to do in the PR. Anything else you would like to add?
<ipfsbot>
[node-ipfs-api] Dignifiedquire opened pull request #78: Start implementing browser tests using karma [do not merge] (master...karma) http://git.io/vChwc
<dignifiedquire>
solves the fs.read issues for most problems
<davidar>
VictorBjelkholm (IRC): but that link doesn't say anything about running in the browser?
<VictorBjelkholm>
davidar, oh, yeah. Is my assumption because it's a api client library in js, should run on both node and browser but you're right, it does not mention running in the browser. My thinking was that you confused node-ipfs-api and node-ipfs
<VictorBjelkholm>
dignifiedquire, cool! Great to see a comparison between karma and testem. Could you try add the running of the daemon before the tests? Looks simpler than testem at this point
<davidar>
VictorBjelkholm (IRC): no, what I meant was my understanding is that node-ipfs is going to be the recommended way for running ipfs in the browser
<davidar>
In which case node-ipfs-api seems redundant
hjoest has quit [Ping timeout: 260 seconds]
<dignifiedquire>
VictorBjelkholm: sure will add the daemon spawner
<dignifiedquire>
<VictorBjelkholm>
davidar, ah, I see. Good point. Might be worth discussion in some issue on github. if nods-ipfs is a gonna be able to run in the browser without running any daemon at all, node-ipfs-api seems redundant indeed
hjoest has joined #ipfs
nagydani1 has joined #ipfs
Zuardi has quit [Remote host closed the connection]
Zuardi has joined #ipfs
ei-slackbot-ipfs has quit [Remote host closed the connection]
ei-slackbot-ipfs has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
nagydani1 has left #ipfs [#ipfs]
barnacs has joined #ipfs
edsilv has quit [Ping timeout: 244 seconds]
Guest73396_u has quit [Ping timeout: 240 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
cemerick has quit [Ping timeout: 256 seconds]
<dignifiedquire>
VictorBjelkholm: done
mildred has joined #ipfs
elima_ has quit [Ping timeout: 244 seconds]
rand_me has joined #ipfs
cemerick has joined #ipfs
hellertime has joined #ipfs
hjoest has quit [Ping timeout: 260 seconds]
hjoest has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
elima_ has joined #ipfs
Suzeanne has joined #ipfs
edsilv has joined #ipfs
<VictorBjelkholm>
dignifiedquire, cool! Looks definitely simpler than testem
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
MOO5E has quit [Ping timeout: 264 seconds]
okket_ has quit [Remote host closed the connection]
okket has joined #ipfs
hjoest has quit [Ping timeout: 260 seconds]
ianopolous3 has joined #ipfs
hjoest has joined #ipfs
ianopolous2 has quit [Ping timeout: 252 seconds]
ianopolous2 has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
MOO5E has joined #ipfs
<tinybike>
i wrote a little helper script to install go-ipfs, poke port 5001 open in iptables, & create an upstart script -- not sure if it's generic enough to be useful to anyone else, but thought i'd share just in case: https://gist.github.com/tinybike/77357ca69cd852655546
ianopolous3 has quit [Ping timeout: 272 seconds]
rand_me has quit [Ping timeout: 240 seconds]
Nitori has quit [Ping timeout: 268 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<drathir>
tinybike: take on mind the ports in future probably will be changed...
<tinybike>
yeah it's a bit hackish :/
<tinybike>
could add port number as an argument i guess
infinity0 has joined #ipfs
<cryptix>
tinybike: sure you want to open 5001 (json api) and not 4001 (p2p swarm)?
<tinybike>
oh good point -- it should open both i guess
rand_me has joined #ipfs
mildred has quit [Ping timeout: 260 seconds]
slothbag has joined #ipfs
<slothbag>
hey guys, i'm trying to pin a manually created block and it says Error: pin: incorrectly formatted merkledag node: Unmarshal failed. unexpected EOF
<slothbag>
so is it not possible to pin custom created blocks
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
MOO5E has quit [Ping timeout: 252 seconds]
<slothbag>
ipfs block get Qm12345 | ipfs block put
<slothbag>
seems to do the same thing I guess
slothbag has quit [Remote host closed the connection]
mildred has joined #ipfs
pod has quit [K-Lined]
bren2010 has quit [K-Lined]
bren2010 has joined #ipfs
hjoest2 has joined #ipfs
pod has joined #ipfs
arpu has joined #ipfs
pod has quit [K-Lined]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
hjoest has quit [Ping timeout: 260 seconds]
MOO5E has joined #ipfs
hjoest has joined #ipfs
hjoest2 has quit [Ping timeout: 260 seconds]
hjoest has quit [Ping timeout: 260 seconds]
hjoest has joined #ipfs
bedeho has quit [Ping timeout: 272 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Guest73396 has joined #ipfs
rubiojr has quit [Quit: Connection closed for inactivity]
arpu has quit [Remote host closed the connection]
bedeho has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Eriqa has joined #ipfs
Spinnaker has quit [Ping timeout: 256 seconds]
<Eriqa>
hi I am a newbie and wanted to know how I can install ipfs on ubuntu. Which of the prebuilt packages should I install ? I didn't see one for debian linux. Thanks in advance.
Spinnaker has joined #ipfs
<VictorBjelkholm>
Eriqa, Linux x86_64 or Linux x86
<M-rschulman1>
Eriqa: Hi and welcome! There isn't, as far as I know, an apt-get package.
<VictorBjelkholm>
"go get -u github.com/ipfs/go-ipfs/cmd/ipfs" also works great if you already have go installed
<M-rschulman1>
But you can download one of those .zip files and just move the executable into your path.
<Eriqa>
I noticed that I could get "go" from apt-get install gccgo-go but I wasn't sure if it was the same version
rehelmin has joined #ipfs
<richardlitt>
Morning team.
mildred has quit [Ping timeout: 240 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<VictorBjelkholm>
Eriqa, I think to install go you'll have to install their pre-built binaries that is not in any debian repository. Otherwise you'll get an old version
<VictorBjelkholm>
richardlitt, "I write a JS API that other programs can require, and then also bundle in a cli that consumes that api to expose the cli ux as a way to help validate the first use of that API." < You mind elaborating on that? Especially the "help validate the first use of that API"
<richardlitt>
Sure.
<richardlitt>
Will respond in the issue, one second.
<richardlitt>
Quick question: anyone know why git might tell me I am up to date with origin/master, but git pull brings down more commits, and then it tells me I am still up to date? Example: https://gist.github.com/RichardLitt/7effcf8d71f8c982ed33
<VictorBjelkholm>
richardlitt, maybe it's not up to date locally. If you're doing git fetch, it updates the status without pulling. Then git will see changes
<Eriqa>
Thanks guys for helping me so fast, it worked! I now have ipfs version 0.3.8. Now it's off to integrating Ethereum contracts with ipfs.
yaraju has joined #ipfs
stoopkid has joined #ipfs
mildred has joined #ipfs
<VictorBjelkholm>
richardlitt, now we're on the same page :) There are two PRs currently, that implements both browser and nodejs tests. Take a look at them if it interests you
<richardlitt>
VictorBjelkholm: Sweet. Added it to the list of things to do on IRC.
<richardlitt>
Actually, maybe my Trello board should be public. Interesting.
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
rehelmin has quit [Quit: Leaving.]
pod has joined #ipfs
Encrypt_ has joined #ipfs
pod has quit [K-Lined]
Encrypt_ is now known as Encrypt
<richardlitt>
NeoTeo: Hey! Are you planning on having a meeting on October 28th, too?
<NeoTeo>
Hi richardlitt . No, the first one is in Nov. as it's the first one and I need for ppl to discover it.
notduncansmith has joined #ipfs
<richardlitt>
NeoTeo: awesome. I've added it as a one-off - I'll keep my eye on it if it starts recurring
notduncansmith has quit [Read error: Connection reset by peer]
Eriqa has left #ipfs [#ipfs]
<richardlitt>
Also, I've just won a free month's coworking at Republikken in Copenhagen, so... might be there within a year or so.
<NeoTeo>
Oh wow :) That's a nice prize. I've been there a fair few times to work, it's a cool place.
<NeoTeo>
So hopefully you'll find time to come to a few meetups ;)
sseagull has joined #ipfs
<richardlitt>
Hope so!
<richardlitt>
I won a month in 14 different places around the globe, so... gonna have to plan this pretty good. :)
zPlus has joined #ipfs
<NeoTeo>
That's a lot! How are you going to find time to do actual work with all that traveling?
<richardlitt>
Won two in Portugal, going to have to visit daviddias haha.
<richardlitt>
haha. Well, I like working on the road, and coworking spaces is traditionally where I spend my time anyway. We'll see! :)
<richardlitt>
Should cut down on coffee costs.
<NeoTeo>
Republikken has good coffee so...
<zPlus>
hi all, I'm trying to learn about ipfs. The video I'm watching https://www.youtube.com/watch?v=8CMxDNuuAiQ says that ipfs connects to some bootstraps nodes to find peers. Who are there nods? Moreover, I'd like to know if files are downloaded from multiple peers at once or one only
<NeoTeo>
It's a really cool way to get to know a lot of devs.
<ipfsbot>
[go-ipfs] RichardLitt created feature/go-contrib-guidelines-link (+1 new commit): http://git.io/vWenC
<ipfsbot>
go-ipfs/feature/go-contrib-guidelines-link dcc2e91 Richard Littauer: Pared down contribute to link to new go guidelines...
___thomasreggi__ has quit [Ping timeout: 244 seconds]
<ipfsbot>
[go-ipfs] RichardLitt force-pushed feature/go-contrib-guidelines-link from dcc2e91 to d90f2fd: http://git.io/vWecY
<ipfsbot>
go-ipfs/feature/go-contrib-guidelines-link d90f2fd Richard Littauer: Pared down contribute to link to new go guidelines...
<richardlitt>
NeoTeo: yep!
Pharyngeal has joined #ipfs
Guest73396 has quit [Ping timeout: 244 seconds]
afdudley has joined #ipfs
<afdudley>
has anyone made a content resistant twitter using IPFS?
<afdudley>
ugh
<afdudley>
censorship resistant.
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<richardlitt>
Hey team: the weekly sync will be starting in roughly two hours. Please be ready with your weekly checklists before hand. daviddias kyledrake krl richardlitt whyrusleeping dignifiedquire jbenet mappum noffle ike amstocker
<kpcyrd>
afdudley: that could actually work by following /ipns/ which links to json
<VictorBjelkholm>
kpcyrd, that's almost what I did! I have ipns resolve to a directly, where I keep id, name and messages
<afdudley>
Yeah, I was thinking of using JWS messages.
<VictorBjelkholm>
messages = json
<afdudley>
VictorBjelkholm: do you have some source I can look at?
<VictorBjelkholm>
the rest of the code in the repository is cleaning up that mess
Tv` has joined #ipfs
<richardlitt>
zPlus: I'm not sure! Anyone else know?
<afdudley>
VictorBjelkholm: what does an account require a node?
<VictorBjelkholm>
afdudley, one ipfs identity = one node = one account in EverythingStays
<VictorBjelkholm>
but since everything is running inside containers, it's easy to start multiple containers with different ipfs identies, hence, having many accounts
<afdudley>
hmm.
voxelot has joined #ipfs
voxelot has quit [Changing host]
voxelot has joined #ipfs
<afdudley>
Are you familiar with JOSE, JWS in particular?
<ipfsbot>
[ipfs] RichardLitt opened pull request #110: Added awesome-ipfs to project directory (master...feature/add-awesome-ipfs-to-pd) http://git.io/vWeu9
edsilv has quit [Read error: Connection reset by peer]
<richardlitt>
zPlus: Yep!
<richardlitt>
Open an issue there if you don't see one answered here, we'll get to it.
<richardlitt>
AFAIK, we multistream file downloads.
<richardlitt>
I think there are some nodes that IPFS maintains to make sure that there are live nodes available for any given file, but I don't know if those are the same as the bootstrapping nodes
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<zPlus>
thanks richardlitt I still have one question more; I'm curious about how the DHT works, let's say I join the network and start with 10 peers. Can my files be downloaded from *any* other node, or just by other nodes that have me as their peer?
<cryptix>
zPlus: content is not relayed - the dht is only used to get who is where and has what
<cryptix>
and content is not transfered to other nodes until they request it activly
mildred has quit [Ping timeout: 240 seconds]
dignifiedquire has quit [Ping timeout: 268 seconds]
<zPlus>
cryptix: let's say I join ipfs and have 10 random peers (from the bootstrap nodes). Will you be able to download my content immediately? Or do I have to have you as a peer?
<VictorBjelkholm>
zPlus, it's handles it automatically so if you give me content you just did ipfs add with and are running the daemon, and I ipfs get it, I will get it
<cryptix>
if i have peers that can talk to yours, i can ask them for your ip4 or 6 addresses and we can try to connect directly (nats!..) to initiate bitswaping the content i want to get from you
<cryptix>
also VictorBjelkholm said - unless you are on different 'swarms' but there is only one public one afaik
<cryptix>
what*
<kpcyrd>
so an ipv4-only node can't pull from an ipv6-only node?
<zPlus>
VictorBjelkholm: cryptix: cool! Thank a lot! I'm intrigued by this though, can you guys please explain briefly how the DHT is stored? I mean, after I "ipfs add" a file, is my local list of files being distributed to all my direct peers? And what happens when I input a file (hash) to retrieve? All nodes ask their peers if they know the [hash, ip]?
<VictorBjelkholm>
zPlus, no, it sits in your own local daemon until someone requests it
<VictorBjelkholm>
how it's implemented I don't know
hjoest2 has joined #ipfs
hjoest3 has joined #ipfs
hjoest has quit [Ping timeout: 260 seconds]
notduncansmith has joined #ipfs
hjoest2 has quit [Ping timeout: 260 seconds]
notduncansmith has quit [Read error: Connection reset by peer]
bedeho has quit [Ping timeout: 272 seconds]
captain_morgan has quit [Ping timeout: 246 seconds]
<ion>
zPlus: Look up distributed hash tables on YouTube, the introduction videos should give a good idea of how they work.
bedeho has joined #ipfs
pfraze has joined #ipfs
SuzieQueue has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<daviddias>
richardlitt: that sounds awesome! :D
<daviddias>
when is that month happening?
arpu has joined #ipfs
<richardlitt>
Not sure. I gotta plan out next year.
<zPlus>
is there any search functionality in ipfs, or do I have to know a file hash to download it?
Suzeanne has quit [Ping timeout: 268 seconds]
r04r is now known as zz_r04r
<pjz>
zPlus: you have to know
<zPlus>
thanks pjz
<zPlus>
here "/ipfs/<hash>/<path>" <hash> is the hash of a file, right? And <path> instead?
<pjz>
...ish, yeah, mostly
<pjz>
<hash> in that case is likely a directory not a file
<zPlus>
pjz: looks like "hash" here is the "peer identity"
Nitori has quit [Quit: "I hate eggs with no yolk. My friend Herensuge was beaten to death by one."]
<nomoremoney>
dignifiedquire, does the binary exists there?
domanic_ has quit [Ping timeout: 240 seconds]
<dignifiedquire>
nomoremoney: it does when I run it locally
nomoremoney is now known as VictorBjelkholm
notduncansmith has joined #ipfs
<VictorBjelkholm>
Huh, that's true...
notduncansmith has quit [Read error: Connection reset by peer]
<VictorBjelkholm>
ah
<VictorBjelkholm>
it's using a weird path, no?
<dignifiedquire>
what do you mean?
jmaguire has joined #ipfs
domanic_ has joined #ipfs
<VictorBjelkholm>
"./node_modules/.bin/concurrent --kill-others "node_modules/.bin/ipfs daemon"" should be "./node_modules/.bin/concurrent --kill-others "./node_modules/.bin/ipfs daemon""
<VictorBjelkholm>
no?
<VictorBjelkholm>
maybe it won't change anything
Nitori has joined #ipfs
<dignifiedquire>
ahr right
<dignifiedquire>
of course
<dignifiedquire>
hopefully that fixes it
juul has joined #ipfs
<dignifiedquire>
VictorBjelkholm: same error :()
atrapado has joined #ipfs
<VictorBjelkholm>
doh
<VictorBjelkholm>
what environment travis runs in?
PsuzeannePsy has joined #ipfs
<VictorBjelkholm>
dignifiedquire, hm, it's trying to run a command called "ipfs daemon" that doesn't exists... The command is "ipfs" and the argument is "daemon"
<ion>
ianopolous: IPFS doesn’t use BitTorrent, it’s just a source of inspiration.
<ion>
ianopolous: The data structure being the same as in Git might be misunderstood by people. That is true on a specific level, the Merkle DAG, but the data structures stored on top of it are not identical. Even the Merkle DAG isn’t generated in an identical way, IPFS uses multihash for object IDs and Git hardcodes SHA-1 AFAIU.
domanic_ has quit [Ping timeout: 240 seconds]
od1n has quit [Quit: This computer has gone to sleep]
<VictorBjelkholm>
ion, I think it's ok for a simplified view. Most people know and recognize how git works, that every change is built on top of another change. Also, the ipfs homepage says "In some ways, IPFS is similar to the Web, but IPFS could be seen as a single BitTorrent swarm, exchanging objects within one Git repository."
domanic_ has joined #ipfs
elima_ has quit [Remote host closed the connection]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<ion>
VictorBjelkholm: Sure. What i’m after is that it would be better to say its parts are “like BitTorrent” and “like Git” (or “could be seen as”) than “it uses BitTorrent” and “the data structure is the same as the one in Git”.
<VictorBjelkholm>
ah, yeah. That I agree on for sure
<whyrusleeping>
Goooood morning everybody!
<whyrusleeping>
I think its time for some role call and the start of a sprint
<daviddias>
- [x] registry-static working with ipfs-blob-store
<daviddias>
- Get all of npm inside IPFS (_blocked_, need to have good Internet for 24h++, planning on doing that 1st thing when I get back to Lisbon on Sunday)
<jbenet>
daviddias: re npm, why not do it from a remote machine?
<jbenet>
daviddias: put the registry in pollux or one of the other nodes
<whyrusleeping>
jbenet: there isnt space on pollux, is there?
<jbenet>
not sure. we can get another node, too.
<daviddias>
do we have a machine with that many gb? from the convo I had with whyrusleeping, it didn't sound like a possibility
<daviddias>
but yeah, if there is a remote machine that I can use, all in for that
<jbenet>
sirius has 120GB free atm.
<whyrusleeping>
thats not enough
<jbenet>
pollux is maxed out.
<jbenet>
how much do you need daviddias?
<kyledrake>
Is the idea behind sprinting that you try to finish everything on your list, or that you create at least a week's worth of stuff? Sorry, I'm new to this.
<daviddias>
last time bengl tried , it was more than half a TB
<whyrusleeping>
kyledrake: the goal is that you finish things on your list
<daviddias>
and that was sometime ago
<daviddias>
should be around ~800 Gb now (or more)
<jbenet>
kyledrake: ideally finish everything on your list. technically, should probably be able to distinguish "committed to finish" vs "hopefully finish" beforehand, but we need a pm to really make all of that work well.
<whyrusleeping>
but my thing is to try and come up with the main goals i *can* accomplish this week, and then a couple stretch goals as well
<daviddias>
kyledrake: we should always try what was on our list because that list is compiled from the sprint discussions and needs/plans of the team. If you happen to get blocked or manage to do other stuff (because you had the time or it became more important), you should report as well so that everyone is in sync
<jbenet>
daviddias: ok we should get another pollux or something. can ask lgierth how to go about it. it has 2.7TB i think
<kyledrake>
ok
<lgierth>
yep let's talk storage in infra hangout
<lgierth>
davidar wants more++ too
<daviddias>
jbenet: that sounds good :)
<lgierth>
(me too)
<whyrusleeping>
i was thinking about building a 2u server and colocating it in seattle
<whyrusleeping>
looking at around 50TB of usable space
<richardlitt>
(back)
<daviddias>
I can also mount a server at the office I have in Lisbon, we have solid 200mb fiber and unlimited power
devbug has joined #ipfs
<whyrusleeping>
daviddias: that works too :)
<daviddias>
whyrusleeping: wanna give me some advices on which hw I should get? This sounds like a fun and good thing to do :)
domanic_ has joined #ipfs
<jbenet>
ok
<whyrusleeping>
daviddias: yes :)
<whyrusleeping>
i can go next
<daviddias>
awesome :D thank you
<whyrusleeping>
- [x] ship 0.3.8
<whyrusleeping>
- [x] fix goprocess panic in swarm/dialer
<whyrusleeping>
- [x] fix goprocess panic in bitswap
<jbenet>
everyone, would be awesome if we add screenshots of what we're making to https://github.com/ipfs/pm/issues/39 so can see in one page a bunch of things coming together
hjoest has quit [Ping timeout: 260 seconds]
<ion>
richardlitt: Doesn’t go-ipfs already have a HTTP API?
<richardlitt>
*docs
<kyledrake>
What does [~] imply
<ion>
It would be very nice if the HTTP API was RESTful. Work with HTTP, not just on top of it.
<richardlitt>
working on, not finished.
<richardlitt>
ion: I mean documentation for the HTTP api, not on the API itself.
<ion>
richardlitt: I got it, thanks
<richardlitt>
ion: cool.
mildred has joined #ipfs
gritzko_ has quit [Ping timeout: 252 seconds]
hjoest has joined #ipfs
<jbenet>
richardlitt: getting the api docs done is important
<whyrusleeping>
^ +1
<richardlitt>
jbenet: I know. Will make that my only priority this week.
<jbenet>
edrex has quit [Ping timeout: 240 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
edrex has joined #ipfs
zz_r04r is now known as r04r
<kyledrake>
I'm ready whenever.
<ion>
daviddias: You have unlimited power? Amazing, I didn’t know that’s physically possible.
<richardlitt>
kyledrake: I'm done.
<richardlitt>
kyledrake: go ahead
dts has joined #ipfs
<daviddias>
ion: ahaha I had to train a lot
<kyledrake>
First, I want to congratulate daviddias on getting a very important talk into the first node.js foundation backed conference in Portland in a few months!
<daviddias>
thank you kyledrake :D
<kyledrake>
Had a great IPFS meetup in Portland with daviddias and brixen this week (thanks guys!). I put together an announcement blog post for libp2p (love this idea!), and then had to spend time finishing a talk for BSides (this took a while, had to double my talk length). Gave the opsec world a good intro to IPFS.
* lgierth
deploying 0.3.8 to the gateways
<kyledrake>
No progress on DMCA. Wasn't a good week to coordinate with jbenet and lgierth schedule-wise.
<kyledrake>
This week: More DMCA stuff, starting with a contact page for ipfs.io. Improvements to libp2p web site, electron app blog post, 0.4.0 announcement post. RE blog posts: If you have something you think is blog-worthy, ping me so I can write it up!
<revolve>
link to the electron app?
<whyrusleeping>
revolve: github.com/ipfs/station
<dignifiedquire>
kyledrake: I'll ping you as soon as I have the first packes for station
<kyledrake>
If I get time by end of week, I want to try putting together an IPFS talk that's specifically for IPFS. I've got some ideas for making a more participatory talk that is more accessible to less technical users and lets them get a feel for what IPFS can do.
<dignifiedquire>
*packages
<revolve>
sound, whyrusleeping
<kyledrake>
Past:
<kyledrake>
- [x] Finish draft of libp2p announcement
<revolve>
whyrusleeping: I got my distributed collaborative editor serving resources from remote peers this morning!
<kyledrake>
- [x] Finish BSides Talk
<kyledrake>
- [x] Work on getting a process for more frequent blog posts.
<kyledrake>
- [ ] Work to improve the libp2p web site
<lgierth>
i have to run -- jbenet whyrusleeping daviddias infrastructure hangout in 1:15h?
<daviddias>
sounds good :)
<whyrusleeping>
lgierth: sure!
<lgierth>
ah no, 45m
<kyledrake>
jbenet you should put your slides on slideshare.net or something similar, or send a few of them to me so I can use them as a reference
<lgierth>
according to schedule
<lgierth>
ok then i really gotta run, but i'll be there at 7pm UTC :)
<ipfsbot>
[go-ipfs] RichardLitt deleted feature/go-contrib-guidelines-link at d90f2fd: http://git.io/vWviP
<Tv`>
whyrusleeping: that thing could use an actual agenda ;-)
<whyrusleeping>
daviddias: agenda pls
<Tv`>
Is attending.io a meetup.com competitor?
<daviddias>
ok , I've written one and then took it off, but I can add again
<jbenet>
kyldrake: i've a few on speakerdeck.com/jbenet but i've more on ipfs.
<ipfsbot>
[go-ipfs] RichardLitt created feature/go-contrib-guidelines-link (+1 new commit): http://git.io/vWvPP
<ipfsbot>
go-ipfs/feature/go-contrib-guidelines-link 7dcf92e Richard Littauer: Fix broken link...
<dignifiedquire>
jbenet: we should talk about slide design at some point ;)
<ipfsbot>
[go-ipfs] RichardLitt opened pull request #1862: Fix broken link (master...feature/go-contrib-guidelines-link) http://git.io/vWvP5
<jbenet>
Tv` yeah we're using attending because meetup.com is a pain for non-regular geo-specific meetups. and their support is awful.
<daviddias>
added an agenda
<kyledrake>
whyrusleeping can I get a quick update on IPNS and what's going on in that camp?
<jbenet>
dignifiedquire: am open to suggestions :) -- wish there was something as easy as keynote but open + with a source code.
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
kyledrake: sure thing, i'll make a post on the tracking issue for it
<kyledrake>
great thx
<kyledrake>
chomping at the bit to get to start doing some mad science with it ;)
<jbenet>
kyledrake: it should work better in 0.3.8
<jbenet>
though it probably is slow atm -- whyrusleeping we should look into why the dht query is not terminating quickly. dialing more should not be as much of a blocker as it is, since it should crawl to the end of the dht and terminate.
<dignifiedquire>
jbenet: not sure I find keynote for code heavy things quite annoying tbh but haven't found anything better on the other hand
<whyrusleeping>
jbenet: its because we bumped the record count to 20
<whyrusleeping>
that basically means it has to exhaust the entire network before it gives u
Myagui has quit [Max SendQ exceeded]
pod_ has joined #ipfs
pod_ is now known as pod
Myagui has joined #ipfs
<kyledrake>
I've got some crazy ideas for using deterministic keys or certificates for allowing an IPFS node to sign for thousands or hundreds of thousands of IPNS records instead of only one, but that's probably best pushed to the future for now.
<jbenet>
kyledrake: an ipfs node will be able to generate any number of keys
<jbenet>
it's waiting on keystore to be implemented (the way to store and keep keys safely)
<jbenet>
whyrusleeping: no it does not need to exhaust the whole network, it only needs to exhaust in the direction of the query
<jbenet>
if the dht is exhausting _the entire network_ then it's implemented incorrectly. it should move in the direction of lower XOR distance only, hence log N
<kyledrake>
jbenet: cool. I'm not sure if it's worth it, but give BIP39 a read if you get a chance this week. It's a way to create billions of keys from a single one. If you dump that single key into mnemonic string generator, it becomes very easy to back up. That one key can then be used to recover the other keys, all you need to know is the index you're on.
<kyledrake>
I could theoretically just make something external to IPFS that does this too if I can insert the keys through an API
<dignifiedquire>
jbenet daviddias and anyone else interested ready for apps hangout?
pfraze has quit [Remote host closed the connection]
captain_morgan has quit [Ping timeout: 240 seconds]
od1n has quit [Quit: This computer has gone to sleep]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<ipfsbot>
[node-ipfs-api] VictorBjelkholm closed pull request #77: Initial work on #57 with cross-browser + node testing [do not merge] (master...cross-browser-testing) http://git.io/vCxde
od1n has joined #ipfs
devbug has quit [Quit: No Ping reply in 180 seconds.]
wopi has quit [Read error: Connection reset by peer]
wopi has joined #ipfs
<VictorBjelkholm>
dignifiedquire, you saw what I wrote before about the error in travis? About bash -c
devbug has joined #ipfs
<dignifiedquire>
VictorBjelkholm: yeah but I couldn't get it to work :(
<dignifiedquire>
concurrently somehow didn't understand the quotes and failed
<VictorBjelkholm>
dignifiedquire, ok, I just pulled your branch, I'll give it a shot
<dignifiedquire>
thanks
od1n has quit [Quit: This computer has gone to sleep]
<VictorBjelkholm>
dignifiedquire, quick question, can you run the node tests directly in karma?
<dignifiedquire>
it's the same file that gets run, the difference is that karma takes the file, runs it through browserify and then executes it in the browser
bsm1175321 has quit [Remote host closed the connection]
<dignifiedquire>
I added a couple of conditionals by checking if we are in the browser or not though by checking if global.window is defined
<multivac>
edsilv: 2015-10-18 - 19:54:39 <ion> tell edsilv The second “exec” line overwrites the first. You'll want a “post-start exec" to run things after starting the main process.
<multivac>
edsilv: 2015-10-18 - 19:57:37 <ion> tell edsilv That config would start the daemon as root. You *definitely* don't want that. See the config I suggested to you earlier.
<ipfsbot>
[go-ipfs] lgierth force-pushed bootstrap-pluto from 2e3feed to a872199: http://git.io/vc1BD
<ipfsbot>
go-ipfs/bootstrap-pluto e912612 Lars Gierth: config: update bootstrap list hostname...
<ipfsbot>
go-ipfs/bootstrap-pluto a872199 Lars Gierth: config: update pluto's peerID...
<edsilv>
this is probably a silly question, but if you view source of that page, you can see that I'm using /ipfs/QmdJ3XgTdWoKwWW8H9UJrW482kJWSZogtYbDfEW1ty6NKq as the root
<edsilv>
if I were to change index.html to use /ipfs/QmU5cU5ixhaHtEYMZVfp6DuHzah3L3fvVajD4StXH7r7MQ/ as the root, this would change the hash of the containing directory
<kpcyrd>
lgierth: I kind of know what changed and what didn't, so I'll probably look into building my own tree objects
<lgierth>
kpcyrd: ipfs-shell is useful for that
<edsilv>
Do I need to use the newest containing directory hash, or am I missing the point entirely here? :-)
___thomasreggi__ has quit [Remote host closed the connection]
<kpcyrd>
lgierth: 1) try to find hash for path 2a) found hash, use it 2b) not found, read, hash, add to db 3) add hash to tree
* lgierth
to the supermarket real quick
od1n has quit [Quit: This computer has gone to sleep]
NeoTeo has joined #ipfs
<VictorBjelkholm>
dignifiedquire, tried pushing the same thing to circleci and it works perfectly. Seems to be a travis issue
hjoest2 has joined #ipfs
Encrypt has quit [Quit: Eating time!]
___thomasreggi__ has joined #ipfs
hjoest has quit [Ping timeout: 260 seconds]
fleeky has joined #ipfs
od1n has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
od1n has quit [Quit: This computer has gone to sleep]
od1n has joined #ipfs
devbug has quit [Quit: No Ping reply in 240 seconds.]
<lgierth>
daviddias: jbenet still on the way? we can do infra later
<whyrusleeping>
nobody knows where jbenet is, i think he was kidnapped by robots
<lgierth>
noes
<whyrusleeping>
judging by the sound of his hangouts
<jbenet>
01011111001010101010121010
<lgierth>
wee there he is
<whyrusleeping>
i think thats binary for 'i am definitely not a ro2bot'
<dignifiedquire>
VictorBjelkholm: strange
<ion>
Packet loss artifacts in audio compression formats are neat.
<lgierth>
i'd be semi-off for a while then if you wanna do infra later
<codehero>
am i correct in my assumption that accessing a hash from ipfs.io/ipfs/ will cache the file on the gateway just like doing ipfs get would cache it on my pc?
<lgierth>
codehero: yep but only on that one host you hit
<codehero>
okay
<codehero>
cool
<lgierth>
codehero: and until `ipfs repo gc` is triggered the next time
<dignifiedquire>
daviddias: are you doing node-ipfs hangout now?
<lgierth>
it's eventually inconsistent, so to say
<dignifiedquire>
or is that delayed?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
PsuzeannePsy has quit [Quit: Leaving]
<daviddias>
I would prefer to wait some minutes for jbenet to arrive at the room, have somethings that I would like to have his feedback too
<jbenet>
Am almost There
<dignifiedquire>
daviddias, jbenet, whyrusleeping where do talks about the webui belong?
<whyrusleeping>
probably belonged in the ipfs-apps talk
<whyrusleeping>
honestly
wopi has quit [Read error: Connection reset by peer]
<dignifiedquire>
asking for next week, don't have much to talk about yet but hope to have more ready in terms of drafts so it would make sense to have some discussion feedback around it
wopi has joined #ipfs
<daviddias>
jbenet: arrived :)
devbug has quit [Quit: No Ping reply in 210 seconds.]
<jbenet>
(we can also do go-ipfs if lgierth prefers to talk later)
<jbenet>
gamemanj: what do you mean? more clear example maybe?
<gamemanj>
jbenet: I used git-format-patch to split a git repository into nicely serialized (Hopefully, I have no idea if they'll be reconstructed with the same hashes) commits, and then put each one into a merkledag node, each commit linking to the one it's based on.
<jbenet>
guessing lgierth isn't around -- whyrusleeping maybe do go-ipfs first in case he comes back?
<kpcyrd>
can you have a look at #1863 during that? that's a blocker for me currently
<kpcyrd>
AUR build is also broken because they got this wrong
<kpcyrd>
it's really hard to find build instructions besides `go get -u github.com/...`
hellertime has quit [Quit: Leaving.]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
chriscool has quit [Quit: Leaving.]
chriscool has joined #ipfs
<jbenet>
ok go-ipfs done. let's give up to 20min for lgierth to appear again, and if not let's just do infra.
<jbenet>
any other hangouts people want to do today?
<jbenet>
kyledrake: may be good to sync up re blog posts and so on later.
___thomasreggi__ has quit [Remote host closed the connection]
<whyrusleeping>
kpcyrd: responded to issue.
<whyrusleeping>
go code relies on proper directory structuring to be built.
<kpcyrd>
whyrusleeping: that doesn't fix the Makefile, does it?
<kpcyrd>
unless you're executing make inside $GOPATH/src/github.com/ipfs/go-ipfs/cmd/ipfs/
od1n has quit [Quit: This computer has gone to sleep]
<whyrusleeping>
replied again, re: building for an AUR package
<ion>
jbenet: If you have idle time, i’d like to explain the algorithm whose visualization i linked to you. But it’s not pressing at all, it’s about livestreaming content onto IPFS and efficient seeking.
edrex has quit [Remote host closed the connection]
<jbenet>
ion: ah yeah, it looked cool. i couldn't tell from the visualization where the links where, i think if you draw those a bit more clearly, i'll get it right away
<jbenet>
were*
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<ion>
jbenet: Each row shows one of the objects in red and its links with a black outline.
<ion>
If you stream about 1 GB of data onto IPFS in 250 kB blocks, there will be at most 12-ish links along with offsets in each 250 kB object and seeking to any object will require at most 12-ish steps.
wopi has quit [Read error: Connection reset by peer]
<kyledrake>
jbenet sounds good. I've got a meeting at 15:00 Pacific and should be available after 17:00
<whyrusleeping>
kpcyrd: you have to build from inside GOPATH, but you can change GOPATH to do so if you need
<ion>
jbenet: cap.png is not loading for me.
wopi has joined #ipfs
<jbenet>
oh right, im at stanford guest network -- they block all the ports
<codehero>
whyrusleeping: but can it, say, add files to your local node or pin files?
<codehero>
sort of like the webui does
<whyrusleeping>
no
ianopolous2 has quit [Quit: Bye]
<codehero>
well. that's what i meant
edsilv has quit [Ping timeout: 244 seconds]
<codehero>
i personally don't mind adding files by terminal since i can even nicely automate that. but some people may prefer a webui for a code pasting service for example
qgnox has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
typical bought off solutions. i'll speak with the organizers about streaming the video with something open source.
<codehero>
:)
jvalleroy has joined #ipfs
NeoTeo has quit [Quit: ZZZzzz…]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<dignifiedquire>
codehero: I could open the asx file fine with vlc
<codehero>
oh. really? didn't know vlc could do that
od1n has joined #ipfs
<dignifiedquire>
I didn't either, but luckily vlc linked itself to the file on my machine :)
<dignifiedquire>
as it would be tricky to get wmp on running on a mac ;)
___thomasreggi__ has joined #ipfs
<kpcyrd>
whyrusleeping: btw, #1863 also has a git related error in it that only happens when build from .tar.gz
<ion>
Are you actually getting something from there? Because i’m not. Perhaps it will work when the stream is up, perhaps it won’t. I was going to try with mpv.
<whyrusleeping>
kpcyrd: oH! i totally missed that bit.
<whyrusleeping>
hmm..... jbenet o/
<whyrusleeping>
the git commit sha in the makefile thing that merged breaks things...
<dignifiedquire>
ion: no it's just blank, but that's always the case with these files they only work while the stream is live
od1n has quit [Ping timeout: 255 seconds]
od1n has joined #ipfs
devbug has joined #ipfs
domanic_ has quit [Ping timeout: 264 seconds]
r04r is now known as zz_r04r
fabrice has left #ipfs [#ipfs]
ygrek has joined #ipfs
<ion>
whyrusleeping: I wonder if the network test thing could simulate bufferbloat? Latency resulting from that would be a really useful metric.
<ion>
Latency and packet loss
<whyrusleeping>
hmm... maybe!
<whyrusleeping>
my next steps once i finish some code i'm working on, is to route all the traffic for the test nodes through a separate container that acts as the 'router'
notduncansmith has joined #ipfs
<whyrusleeping>
i bet once i do that, we can simulate just about anything
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
that change to add the git commit sha into ipfs
<jbenet>
yes?
<whyrusleeping>
breaks when you get the source through a tarball
<whyrusleeping>
which is how most package managers operate
<jbenet>
"breaks" meaning what?
<whyrusleeping>
like, the makefile fails
<jbenet>
ah because it does not have git-- right. i missed that.
<ion>
Would this be useful? git-get-tar-commit-id (1) - Extract commit ID from an archive created using git-archive
<jbenet>
it should check, then.
<whyrusleeping>
go build -ldflags="-X "github.com/ipfs/go-ipfs/repo/config".CurrentCommit=`git rev-parse --short HEAD`"
<whyrusleeping>
fatal: Not a git repository (or any of the parent directories): .git
<jbenet>
cant count on the archive created using git
<ion>
Or the archive even being anywhere when the build is being done, now that i think of it.
<jbenet>
whyrusleeping: it should maybe set it to `unknown` by default or something.
<whyrusleeping>
yeah, might be able to also try that script ion posted
<whyrusleeping>
kpcyrd: sorry about that, you can replace the make command with simply 'go build'
Encrypt has joined #ipfs
<jbenet>
whyrusleeping: want to PR up a fix?
<whyrusleeping>
jbenet: yeap, on it
<jbenet>
thanks
domanic_ has joined #ipfs
zPlus_ has quit [Quit: Konversation terminated!]
bsm1175321 has quit [Ping timeout: 268 seconds]
<ipfsbot>
[go-ipfs] whyrusleeping created fix/makefile-sha (+1 new commit): http://git.io/vWJto
<ipfsbot>
go-ipfs/fix/makefile-sha 6976ca7 Jeromy: fix makefile failing when not in a git repo...
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #1865: fix makefile failing when not in a git repo (master...fix/makefile-sha) http://git.io/vWJtF
od1n has quit [Ping timeout: 268 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Encrypt has quit [Quit: Sleeping time!]
od1n has joined #ipfs
Spinnaker has quit [Ping timeout: 252 seconds]
<ipfsbot>
[go-ipfs] lgierth closed pull request #1543: core/corehttp: add support for posting directories to gateway (master...post_dir_gateway) http://git.io/vOvd5
Spinnaker has joined #ipfs
<ion>
whyrusleeping: The Makefile only uses the commit variable once so it’s not a problem, but in general it’s best to define variables with := instead of = if you want the value to be evaluated at the definition and not every call site.
<kpcyrd>
the .service file needs to be inside go-ipfs/ for packaging
<codehero>
oh. cool. a systemd service
<whyrusleeping>
kpcyrd: maybe add a comment in the service file referring to that readme then?
<codehero>
is ipfs secure enough to be running as a service all the time?
<kpcyrd>
codehero: do you understand what a service is?
<codehero>
i do
Guest73396 has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<kpcyrd>
there isn't much difference from `ipfs daemon` and `systemctl --user start ipfs`
<codehero>
i'm aware of that
<codehero>
and i don't have the daemon constantly running
<codehero>
so that's my question
<codehero>
if it's secure enough to be running all the time
tinybike has quit [Ping timeout: 272 seconds]
Guest73396_d has joined #ipfs
<achin>
i think it's dangerous to think "i don't trust this to be running for too long, but it's ok if i run it for a short while"
<codehero>
heh. well, true
Guest73396 has quit [Ping timeout: 240 seconds]
<achin>
to put it another way: if you don't trust IPFS to be running 24/7, why would you trust it to run for 5 minutes?
<whyrusleeping>
i run ipfs all the time on most of my computers. i'm fairly confident in it
<kpcyrd>
whyrusleeping: do I need to sign off again if I add more commits?
<whyrusleeping>
kpcyrd: each commit needs the sign-off. gitcop links to a script that does it automatically
<codehero>
achin: well. it's not like there's a progress bar for hacking, but running insecure software for short periods of time reduces the chances of being targeted
<achin>
yes, that is certainly true
<whyrusleeping>
codehero: but a hacker can run it on their own machine
<whyrusleeping>
find the exploit
<codehero>
although that's less true when ip addresses are publicly visible as with ipfs
<whyrusleeping>
wait for your to come online
<whyrusleeping>
and then... yeah
<codehero>
yeah
Guest73396 has joined #ipfs
<whyrusleeping>
but, ipfs runs with user level permissions, is written in go (so no buffer attacks) and doesnt exec anything.
<whyrusleeping>
worst case, your disk gets filled up
<whyrusleeping>
which is actually going to be addressed fairly soon
Guest73396_t has joined #ipfs
<achin>
are buffer overflows truely impossible in go?
<whyrusleeping>
as long as you arent using the 'unsafe' package
<whyrusleeping>
yes
<achin>
does ipfs use any unsafe code?
<whyrusleeping>
the protobuf marshaler does it under certain optimization modes