notduncansmith has quit [Read error: Connection reset by peer]
www has joined #ipfs
www has quit [Ping timeout: 244 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
domanic has quit [Ping timeout: 240 seconds]
domanic has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
humanbsd has quit [Ping timeout: 246 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<rschulman>
hi all
humanbsd has joined #ipfs
pfraze has joined #ipfs
rtlong has joined #ipfs
rtlong has quit [Ping timeout: 260 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
g'mornin!
<alu>
morning
<rschulman>
hey whyrusleeping
<rschulman>
how goes?
<whyrusleeping>
its going, i hate mornings, but thats nothing new
<whyrusleeping>
the mfs api is pretty much built out
<whyrusleeping>
i just need to finish up a few more calls, but it should be pretty good
domanic has quit [Ping timeout: 256 seconds]
domanic has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<Luzifer>
ohai whyrusleeping
<whyrusleeping>
Luzifer: hey there!
<rschulman>
whyrusleeping: That’s awesome
<rschulman>
can’t wait to test
MatrixBridge has quit [Remote host closed the connection]
hellertime has quit [Read error: Connection reset by peer]
hellertime1 has joined #ipfs
domanic has quit [Ping timeout: 246 seconds]
<pguth2>
Let's say I give a presentation and have a big file, that all participants should download in *streaming mode*. All participants are using IPFS and I present a shortened link to the IPFS ressource that is currently hosted only by me. All the participants open the link at roughly the same time. Will the IPFS instances in the room coordinate somehow or will everyone basically just download from me clogging up my bandwidth? (ping ThomasWaldmann)
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
M-prosodyContext has joined #ipfs
M-hash has joined #ipfs
M-staplemac has joined #ipfs
M-prosodyContext has quit [Remote host closed the connection]
M-hash has quit [Remote host closed the connection]
M-staplemac has quit [Remote host closed the connection]
MatrixBridge has joined #ipfs
M-mistake has joined #ipfs
M-prosodyContext has joined #ipfs
M-hash has joined #ipfs
M-staplemac has joined #ipfs
M-mistake has quit [Remote host closed the connection]
M-prosodyContext has quit [Remote host closed the connection]
M-staplemac has quit [Remote host closed the connection]
M-hash has quit [Remote host closed the connection]
MatrixBridge has quit [Remote host closed the connection]
<cryptix>
but there are to much native javascript speakers here that i dont see a point in advertising it
<voxelot>
i'm trying to figure out how to get the buffer through in add as well +1 ping to krl :)
<pguth2>
cryptix I would like to see what you did there.
Tv` has joined #ipfs
<whyrusleeping>
voxelot: okay, i'm glad i'm not the only one having an issue
<cryptix>
voxelot whyrusleeping: tl/dr; i think you want var buf = new ArrayBuffer(); for modern browsers but dont nail me on it feross/buffer is prop. what you want?
<whyrusleeping>
i had just assumed i was bad at javascript
<whyrusleeping>
ArrayBuffer? hrm
<cryptix>
there is also Uint8Array
<whyrusleeping>
I'm using a Uint8Array and wrapping that in a Buffer
<voxelot>
too many streams, guess we try them all
<cryptix>
iirc gopherjs uses that to represent []byte
<whyrusleeping>
using ArrayBuffer -> 'addFile takes vinyl files'
<lgierth>
i'd like to have the sprint gh issue at monday again
<lgierth>
or just stick with a pad
<jbenet>
lgierth: yeah we should have the gh issue on monday. we've (i've) just slacked on putting it up. I'll make sure to
mildred1 has quit [Read error: Connection reset by peer]
<lgierth>
i can do it too. sprint master eh?
<lgierth>
i just won't be up every sunday night at 3 am utc :)
<lgierth>
*monday
mildred has joined #ipfs
<jbenet>
yeah, i can handle it. i can make sure to delegate it when i wont be able to be there, but i should still own it
<jbenet>
ok who's ready -- whyrusleeping?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
sure
<whyrusleeping>
This last week I worked on getting the ipfs http commands lib refactored and fixed up. We now have better reliability provided by the commands as they are now able to return errors after beginning a stream of data via http trailers.
<whyrusleeping>
I implemented a --create option for ipfs patch, along with a moderate refactor of the involved code.
<whyrusleeping>
Windows builds work again, although NAT traversal is iffy on the platform.
<whyrusleeping>
we shipped 0.3.6! and it was awesome, but the webui was broken, so we fixed that and shipped 0.3.7.
<whyrusleeping>
Since then, I implemented rabin fingerprinting chunking which I still need to address CR on.
<whyrusleeping>
I also started working on implementing the mfs API to allow for a whole bunch of cool new apps to be built, including a dropbox-like app i'm writing in javascript.
<whyrusleeping>
I learned a lot about javascript and I'm kinda getting the hang of how things work.
<cryptix>
mfs stands for? :x
<whyrusleeping>
mutable filesystem
humanbsd has quit [Ping timeout: 244 seconds]
<whyrusleeping>
i'm not set on the name, but its short and it works
<jbenet>
whyrusleeping what did you set out to do last week that didnt get through with?
jbenet_ has joined #ipfs
<whyrusleeping>
i think I kept pace pretty well, I may have aimed to do dht/records stuff, but didnt get to looking at that more
<daviddias>
- We've run successfully an interoperability test between go-ipfs and node-ipfs, using the PING protocol. It works™
<daviddias>
- Friedel is now helping node-ipfs land (woot! \o/), starting by adding tests.
<daviddias>
- network-layer
<daviddias>
- [x] ton of discussions on defining how we are going to modularize (already following what was defined) the libp2p (see images below for future reference)
<daviddias>
- Peer routing
<daviddias>
- [x] standardise the interface for router components
<daviddias>
- [~] make kad-router. kbucket update/split works (with tests), dht registers on swarm, need to finish the peer `query loop`
<daviddias>
 - swarm
<daviddias>
- [x] add ability to open just connection (without a stream) (today!)
<lgierth>
ok can i go next?the supermarket closes in 30 mins
<daviddias>
I need to get kad-router (which is kind of close there, will need to spend a great amount of time actually in tests) done I can wrap with a simple record stuff and call it alpha of libp2p
rtlong has joined #ipfs
<lgierth>
very cool that go/node interop is a thing
<daviddias>
probably going to multiplex my time getting other stuff polished with Juan for CCC
<jbenet_>
daviddias, whyrusleeping: awesome work on getting interfacing with go/node
<daviddias>
it was magical! :D
<whyrusleeping>
:D
<jbenet_>
lgierth: yep! sec
<whyrusleeping>
it was very satisfying to see those pings come back
<daviddias>
we followed the spec, implemented it and it worked at "first"
<daviddias>
we just had to change protocol names
<lgierth>
heh
<jbenet_>
daviddias: so next is getting the dht protocol in yeah?
<whyrusleeping>
yeap, daviddias lets write out what all the protocols should be named in an issue
<whyrusleeping>
and i'll update it in the codebase
<daviddias>
from '/ipfs/spdy/1.0.0' to '/spdystream'
<daviddias>
jbenet_ yep, the kad-router is the dht peer-routing
<jbenet_>
daviddias: yep, awesome. anything else? anything else you wanted to do but didnt get to?
<daviddias>
whyrusleeping:
<daviddias>
I've a bunch of things on my line of sight (blog posts about the muxing extravaganza, doing a mDNS router and more stuff), but right now the priority is really the kad-router so that we can ship that libp2p alpha
<daviddias>
and with that, polish rest of the spec
<jbenet_>
daviddias: sounds good. yep. also, this week probably all the stuff we discussed on the train. but lets talk more later.
<jbenet_>
ok next, lgierth?
<daviddias>
yep
<lgierth>
c&p inc
<lgierth>
- ipfs repo config on the gateways is now managed by ansible. in the past we lost the private key of pluto, which won't happen again.
rtlong has quit [Ping timeout: 246 seconds]
<lgierth>
- neptune has been rebuilt
<lgierth>
- the nginx-related ansible code got a good refactoring (CR needed)
<lgierth>
- pinbot and metrics have been moved to a new host called scruffy (CR needed). that means we now have infrastructure stuff separate from gateway/bootstrap nodes
<lgierth>
- did a bit more work on cjdns-discovery, but that's still WIP
<lgierth>
- checked out other cjdns integration points with daviddias
<lgierth>
- got the dmca blocklist daemon to work today
<lgierth>
- battlemesh preparations and travels
<lgierth>
- didn't get to writing down testnet stuff
<lgierth>
c&p eof
<lgierth>
whyrusleeping: i'll have a few comments re: blocklist
rtlong has joined #ipfs
<whyrusleeping>
lgierth: cool, lets pick a time to figure that out
<lgierth>
yep infra sync maybe
<jbenet_>
lgierth: ok sweet. may be good to chart out a roadmap of cjdns integration points, etc, maybe a meta issue with a list of things
<jbenet_>
lgierth: cool, will be good to have the blocklist ready. +1.
<lgierth>
i'm a bit lost there at the moment. let's discuss that later today
<jbenet_>
lgierth: great, anything else?
<lgierth>
that's it i think
<jbenet_>
ok thanks! next -- krl ?
<krl>
ok
<krl>
made the api test suite run also in browserify
<krl>
append-only-finger tree that persists to ipfs (github.com/krl/aolog), d3 animation of the structure (for demo and blog post)
<krl>
so there's some different things i could focus on, fixing api from browser seems highest prio, and with the CORS stuff landed it should work. Testing of adding a buffer implemented in browser already. We might need to simulate a file upload in phantomjs though.
<krl>
windows installer? making an interface for publishing twitter-like logs on ipfs
<krl>
no release of modular webui yet, want to browser-test the api add first, and re-add logs since this is now fixed (thx whyrusleeping)
<krl>
eof
<whyrusleeping>
krl: would love to get the webui to a point where we can publish it again
<whyrusleeping>
would also (more personally) love to have the add stuff working in the browser
thelinuxkid has joined #ipfs
<whyrusleeping>
i'm learning more about javascript and may be able to help out on some of these things too
<jbenet_>
krl: how is the component-ization going? did you get the webcomponents set up?
thelinuxkid has quit [Client Quit]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<krl>
no, i have not worked on that yet, the question is if we want to wait for that before we release a new webui?
<krl>
that's another relatively major redesign
<whyrusleeping>
gordon was thinking we probably should hold off on webcomponents for now
<whyrusleeping>
but maybe i misheard him
<jbenet_>
krl: makes sense. we can discuss with daviddias later today-- he has some good thoughts on it (and re polymer)
<jbenet_>
whyrusleeping: what was the reason?
thelinuxkid has joined #ipfs
<whyrusleeping>
it makes interop for others a lot more difficult
<whyrusleeping>
essentially, anyone who wants to work with our code or reuse it has to also use polymer
thelinuxkid has quit [Client Quit]
<jbenet_>
whyrusleeping: should discuss that prob on github with the notes + relevant links.
thelinuxkid has joined #ipfs
<jbenet_>
krl: so last week was (a) api test suite run on browserify, (b) aoft/ipfs, and (c) work on modular webui?
<daviddias>
whyrusleeping: gordon mentioned that not all browsers implement the full spec yet (expect chrome), and that the polyfil (named polimer) implements the a shadow dom polyfil that is considerably slow
<jbenet_>
krl: if we don't do components, what's the eta on modular webui?
<krl>
jbenet_: blocked on test add, but after that just pull out webui-common, and add logs
<daviddias>
but what we gain of ease of use, standardization, opportunity to set a new trend, is very important to consider, even if it is a bit slow in firefox (which eventually have the spec also implemented)
<daviddias>
ok
<whyrusleeping>
daviddias: ah, yeah
<jbenet_>
krl: can you pull out webui-common and logs without the ipfs add fix?
<krl>
yeah sure
<jbenet_>
krl: also whyrusleeping fixed the eventlogs, they return again on commands. what's missing there?
<krl>
yeah i saw that, i have not tried it out yet, but assuming it's the same as the old api logs, it shouldn't be a problem
www has joined #ipfs
<jbenet_>
(they work with 'ipfs log tail' but not on my webui afaict)
<jbenet_>
anyway, discuss later on. ok thanks
<jbenet_>
i'll go next
<jbenet_>
This week was very productive in terms of 0.3.6. Other thing so so.
<jbenet>
- [x] SEA -> BER -> MDX team travel (lots of) things
<jbenet>
there we go. (/me spams)
* whyrusleeping
gets back up off the floor
<jbenet>
oh and i got a next iteration of ipfsld done. need to figure out what the node api will be exactly, and need to figure out marshaling details (reflect stuff, validation, link types)
humanbsd has joined #ipfs
<jbenet>
ok done. anyone else?
<whyrusleeping>
jbenet: hows records stuff/ipld?
www has quit [Ping timeout: 246 seconds]
www has joined #ipfs
<jbenet>
whyrusleeping o/
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
oh, there it is
<jbenet_>
ok looks like that's it. see you at discussions.
<jbenet_>
i'm not sure if the contingent of us in maribor will be able to do video (we will try) as the connectivity isn't amazing. but we should be able to do audio
simonv3 has joined #ipfs
* whyrusleeping
can video chat with himself
Eudaimonstro has joined #ipfs
freedaemon has joined #ipfs
hellertime has quit [Read error: Connection reset by peer]
hellertime has joined #ipfs
thelinuxkid has quit [Quit: Leaving.]
thelinuxkid has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
jbenet_ has quit [Ping timeout: 265 seconds]
thelinuxkid has quit [Client Quit]
Leer10 has joined #ipfs
Encrypt has joined #ipfs
www has quit [Ping timeout: 246 seconds]
thelinuxkid has joined #ipfs
<rschulman>
whyrusleeping: How’s that solo video chat treating you?
<rschulman>
:)
<krl>
jbenet: ping
Encrypt has quit [Quit: Quitte]
thelinuxkid has quit [Quit: Leaving.]
thelinuxkid has joined #ipfs
<jbenet>
Krl sorry Internet here is interesting. Can have the discussion here on irc?
<krl>
jbenet: works for me
<daviddias>
Here too
thelinuxkid has quit [Client Quit]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
Ok let's figure this out.
<krl>
there's kind of different levels to this, i'm not sure if an ipfs app directly maps to a webcomponent
<jbenet>
krl what are the strong differences/reasons it shouldn't?
<krl>
if i'm not mistaken, webcomponents are more like 'widgets'
<daviddias>
What is your primary concern if we have them in an WebComponent, would there be anything from the Spec that would limit an IPFS web app?
<krl>
and for example a twitter-like ipfs app would consist of multiple parts/widgets
<daviddias>
A web Component can be a set of several components
<daviddias>
( that's the shadow Dom part)
<jbenet>
How is this different from an npm module (which can be both library (widget) or program (app)
<daviddias>
And each Component exposes an interface to talk with it that's defined but be Spec
<krl>
i don't have a strong argument here, just thinking about how to partition it
<krl>
so the structure would be that the whole webui is one webcomponent, that imports other components?
<daviddias>
Can be that, or can be a page that brings its functionality through web components, the root doesn't need to be a web Component
<daviddias>
Web components are also nice for responsiveness
<jbenet>
Afaict, webcomponents are roughly equivalent to modules. So can s/webcomponent/module/
<krl>
so the risk would be that we're betting on webcomponents becoming standard enough
<krl>
so that people will want to use them to write ipfs apps
<daviddias>
Webcomponent also scopes css and js for a given component ( which is awesome )
<krl>
yes, this would admittedly simplify the model as well
<jbenet>
Afaik they already are well standardized. Used all over happs and Mozilla is implementing the standard.
<jbenet>
(Also isn't tHe other option our own rolled ad hoc thing which isn't standard)?
<krl>
jbenet: i mean more like, people wanting to learn writing ipfs apps having to learn two things at once
<krl>
jbenet: that's a good point
<krl>
but that's one part of it, whose other part could either be a standard html page, or a webcomponent
<krl>
but i'm kind of convinced of the model, and would not have anything against re-working this into webcomponents
<jbenet>
Doesn't that mean learning even more? (Also webcomponents will have a lot of lit / explanations)
<krl>
the question is then what to put time into until camp, if i should release the iframe based thing or go directly for components, and how much i should focus on getting the starlog poc done
<jbenet>
Yeah I think these are two sep qs-- we can merge iframes now and write up how to go to components
<daviddias>
If the PR is ready and doesn't introduce any bug that we would have to solve. Merging it would be fine
<daviddias>
Until we have the web component version
<jbenet>
Agreed that iframes/comps won't add much to camp
<jbenet>
I think the blog is the most important thing to do.
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<krl>
it would be really nice to have ipns/keys to go with the blog
<krl>
i'm not completely utd on what the prios there are
mildred has quit [Quit: Leaving.]
* whyrusleeping
returned from lunch
<daviddias>
whyrusleeping: marination?
<jbenet>
Keys would be nice but adding them right now would need whyrusleeping's help and a migration later to move them into being objects
<jbenet>
That coupled with ccc being next week isn't super promising.
<jbenet>
(But yes keys would be nice)
<krl>
it's not a blocker or anything, would just make a nicer demo
<jbenet>
Yeah. Why don't you spec out what the demo looks like first, in both cases, and let's pick one?
<jbenet>
Having a clear spec will also make it obvious what is achievable or not
<krl>
ok, i'll write it up tomorrow, will also look around at which frontends are out there already
<krl>
like ghost
<krl>
did you look around anything? remember you talked abotut it
thelinuxkid has joined #ipfs
<whyrusleeping>
daviddias: nope, marination is only on wednesdays
<jbenet>
krl didn't find anything conclusive yet.
<jbenet>
krl ok to write it tomorrow but this is next week so have that in mind. Few days left
<jbenet>
Ok times up :/ any last things?
<krl>
nothing in particular
<ipfsbot>
[go-ipfs] whyrusleeping force-pushed feat/mfs from 5a24cd7 to fbc8f7b: http://git.io/vOWXQ
<whyrusleeping>
rschulman: take a peek if you want
<whyrusleeping>
cryptix: my twitter feed is like 30% in german, i blame you
<whyrusleeping>
i have to learn german now
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
hellertime has quit [Read error: Connection reset by peer]
hellertime has joined #ipfs
<rschulman>
lol
<rschulman>
whyrusleeping: If I build from source now, I should be able to use it, no?
freedaemon has quit []
freedaemon has joined #ipfs
headbite has quit [Remote host closed the connection]
<jbenet>
Sorry our wifi cut out.
<jbenet>
whyrusleeping: down for infrastructure chat in 2? (Travel keeps lining up with Mondays :/)
<whyrusleeping>
sure, i can chat now
<whyrusleeping>
rschulman: yeah, if you build that branch from source it should just work for you
<rschulman>
whyrusleeping: Trying. Can’t get go to build it. Command just returns immediately.
<jbenet>
Ok just waiting on lgierth for a sec and we're good to go
jbenet_ has joined #ipfs
<whyrusleeping>
rschulman: how are you building?
<whyrusleeping>
jbenet: same link?
<rschulman>
whyrusleeping: Just git clone, git checkout feat/mfs, go build
<rschulman>
or go install both return instantly
<whyrusleeping>
go build ./cmd/ipfs
<whyrusleeping>
or go install ./cmd/ipfs
<lgierth>
i'm here
<rschulman>
that’s… werid
<jbenet_>
whyrusleeping a/v wont work-- the net is bad
<rschulman>
never seen that before in go.
<rschulman>
huh
<whyrusleeping>
jbenet_: oooh, okay
<lgierth>
down for infra? blocklist is the only thing unless you have more to talk about
<lgierth>
i got the blocklist daemon to work today, it's not dumping a simple newline-delimited list of keys though, but a map of notice-identifier => keys-to-block
<lgierth>
notice-identifier so that we can show more info, or provide a link to more info on why something got blocked
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet_>
whyrusleeping o/
<whyrusleeping>
yeah?
<whyrusleeping>
ah, lgierth is here now
<lgierth>
ideally i'd like to see a 451 response with a link to more info
<whyrusleeping>
okay, so how do we generate the notice identifiers?
<jbenet_>
+1 on HTTP 451
<whyrusleeping>
and what would one look like?
<lgierth>
"opaque" :P
<whyrusleeping>
(i'm already returning http 451 on my branch :D )
<lgierth>
2015-08-03-foobar
<lgierth>
ah good! i looked the other day
<jbenet_>
notice id could be derived from the notice itself (could be its hash)
<jbenet_>
though yeah date + sender work too
<lgierth>
the notices repo has a directory per notice, with a notice.md and a keys file (newline-separated list of keys)
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #1550: WIP: implement mfs API (master...feat/mfs) http://git.io/vORo3
<jbenet_>
whyrusleeping allow/deny is well known networking / web / perms terminology
<whyrusleeping>
i mean, blacklists and whitelists are the canonical terms...
<jbenet_>
whyrusleeping (blocklist is clear in the abstract, but may confuse us as we have things called blocks etc)
<jbenet_>
whyrusleeping: yes, and racism was canonical. it's been brough up to me multiple times by multiple people
<whyrusleeping>
alright, allowlist i can deal with. but blocklist is bad naming on our part
<lgierth>
denylist
<jbenet_>
{allow, deny}list -- let's do it.
<whyrusleeping>
kay
<lgierth>
whyrusleeping: so right now the daemon returns this: [{"uri":"http://dmca.ipfs.io/2015-08-03-foobar","keys":["Qmfoobar","Qmsomething","Qmbarbaz"]}]
<lgierth>
s/daemon/denylist daemon/
dignifiedquire has joined #ipfs
<whyrusleeping>
okay, looks like a good format to me
<lgierth>
ok then i'll give it a bit more polish and get a review
<whyrusleeping>
SGTM, i'll change up my code to parse that format
<jbenet_>
lgierth: can we have it be a route rather than a whole subdomain for it? maybe something like http://i.ipfs.io/lists/dmca ?
<whyrusleeping>
jbenet_: nginx can proxy it as a route, cant it?
<jbenet_>
also, a note on this, the list needs to be signed and we need to verify sigs on it
<lgierth>
sure we can
<lgierth>
mhk. gpg-signed?
<jbenet_>
we cant have a list like this floating around allowing anyone to DOS others by taking out an http server, dns, etc.
<jbenet_>
yeah gpg sgtm
<jbenet_>
(we can use ipfs signed objects, too)
<whyrusleeping>
eta on ipfs signed objects?
<jbenet_>
oh also, there may be multipel of these lists, sometimes people will want to keep another list for other reasons, so people should be able to follow multiple
<lgierth>
well, our denylist will only be valid if it comes from $url
se3000 has joined #ipfs
<jbenet_>
(ex: content may be blocked in particular countries but not others)
<rschulman>
whyrusleeping: First bug report, I just get repeated “404 Command Not Found” as a return. :)
<jbenet_>
lgierth: $url can be tampered with. even after adding ssl. rather have the whole list itself signed.
<whyrusleeping>
rschulman: uhm... did you build and run the new code?
<rschulman>
yeah
<whyrusleeping>
curling which endpoint?
<rschulman>
nope, not curl
<rschulman>
running the binary gets me that on the command line
<whyrusleeping>
which commands?
<lgierth>
jbenet_: of course, just saying there shouldn't be random denylists floating around the net
<rschulman>
ipfs mfs create alone gets me an error to include a name
<rschulman>
anything else gives me the 4040
<rschulman>
e.g. “ipfs mfs create testing”
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
rschulman: are you sure both your daemon and client are running the new code?
<rschulman>
oh bugger
jbenet_ has quit [Ping timeout: 260 seconds]
<whyrusleeping>
rschulman: a warning, its a bit clunky still
<whyrusleeping>
i think the session option needs to be on the mfs command instead of on each subcommand
<jbenet_>
whyrusleeping: just need to finish the ld formatting part and implement the "storing keys" part of keystore.
<jbenet_>
whyrusleeping: we should probably finsh specing that out and implement it into dev0.4.0
<whyrusleeping>
storing keys is gonna require the public/private blockstore distinction
<jbenet_>
yeah we need to be able to figure out not serving something.
<jbenet_>
whyrusleeping: it may be worthwhile for you to spend a bit of time looking at how capabilities work, and how they're implemented in various systems.
<jbenet_>
let's dicuss that in go-ipfs -- should discuss node-ipfs with daviddias soon
<jbenet_>
daviddias: down for node-ipfs?
<daviddias>
I'm down
<whyrusleeping>
standard implementations of capabilities (at least in operating systems) use a check to the kernel based on a UID. other systems that are more relevant to use (distributed systems) use crypto and key sharing for capabilities
<whyrusleeping>
tb discussed later...
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet_>
daviddias let's discuss node-ipfs
jbenet_ has quit [Read error: Connection reset by peer]
Encrypt has joined #ipfs
hellertime1 has joined #ipfs
hellertime has quit [Ping timeout: 240 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<rschulman>
whyrusleeping: When I try to start the daemon now: ERRO[16:04:41:000] error mounting: %!s(<nil>) no such rootfs module=core/commands
<ipfsbot>
[go-ipfs] whyrusleeping force-pushed feat/lists from 21e5e32 to 6d4b7b1: http://git.io/vO0ev
<ipfsbot>
go-ipfs/feat/lists 6d4b7b1 Jeromy: change naming and parse format...
<ipfsbot>
go-ipfs/feat/lists 9e5c5df Jeromy: implement access control lists for gateway...
<whyrusleeping>
rschulman: thats weird... are you using 'ipfs mount' ?
<rschulman>
ipfs daemon --mount
<rschulman>
but yes
<whyrusleeping>
ah, yeah... my changes mess with the /ipns mount
<whyrusleeping>
since ipns used what used to be ipnsfs
<whyrusleeping>
and i've refactored it
<rschulman>
gotcha!
<whyrusleeping>
if you dont need the fuse mounts for now, avoid using them until i fix that
<whyrusleeping>
but this all makes the fuse stuff a little bit more testable
<whyrusleeping>
since it decouples the filesystem stuff from the fuse stuff
<whyrusleeping>
and the ipns stuff
<rschulman>
yay for ipns testing!
<rschulman>
:)
<rschulman>
so should mfs create return a hash?
<rschulman>
is that right?
jbenet_ has joined #ipfs
therealplato has quit [Ping timeout: 256 seconds]
<whyrusleeping>
rschulman: it doesnt return a hash
<whyrusleeping>
its not doing that is it?
<rschulman>
haha, no
<rschulman>
I was just wondering how to distinguish success.
<whyrusleeping>
running 'ipfs mfs' will show you all active mounts and their current hash roots
<whyrusleeping>
they dont persist over daemon restarts yet, but i'm thinking that i might want that functionality
hellertime1 has quit [Ping timeout: 244 seconds]
<rschulman>
gotcha
hellertime has joined #ipfs
jbenet_ has quit [Ping timeout: 260 seconds]
therealplato has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
pfraze has quit [Read error: Connection reset by peer]
hellertime has quit [Ping timeout: 250 seconds]
pfraze has joined #ipfs
hellertime has joined #ipfs
Encrypt has quit [Quit: Sleeping time!]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
therealplato has quit [Ping timeout: 260 seconds]
humanbsd has joined #ipfs
jbenet_ has joined #ipfs
jbenet_ has quit [Client Quit]
therealplato has joined #ipfs
jbenet_ has joined #ipfs
dignifiedquire has quit [Quit: dignifiedquire]
<jbenet_>
whyrusleeping: sorry internet here went down _completely_
<whyrusleeping>
jbenet_: lameeeee
<whyrusleeping>
i'll just sit here sipping my midafternoon coffee
<jbenet>
it seems to be working
<jbenet>
daviddias want to finish our node-ipfs discussion?
<jbenet>
(gist so far: we've got the swarm and most of the network down. next is the DHT and various libp2p components
<whyrusleeping>
SGTM
<daviddias>
sgtm
<daviddias>
through chat or video?
<jbenet>
chat'
<daviddias>
jbenet: do I have irc?
<jbenet_>
yes
<whyrusleeping>
jbenet_, meet jbenet
<daviddias>
nice
<jbenet_>
o/
<jbenet>
\o
mildred has quit [Quit: Leaving.]
<daviddias>
making a quick catch up
<jbenet>
we were discussing also node-ipfs-api and improving it
<daviddias>
ok, jumping to that then
<daviddias>
there is a PR on node-ipfs-api that would be good to be done and merged
<daviddias>
so I can tinker with it lil more freely
<jbenet>
whyrusleeping: one thing might be to add a sharness tests that tests against the node-ipfs-api
<daviddias>
doing a test/codestyle clean up
<jbenet>
(sharness test to go-ipfs)
<whyrusleeping>
jbenet: yeah, i was thinking about that
<jbenet>
and/or the webui (to avoid problems like the last one)
<whyrusleeping>
how would we do that? just git clone down and run some tests?
<jbenet>
daviddias: +1
<daviddias>
helping to make that experience to run it in the browser
<daviddias>
and add a specific set of tests for all of that CORS/future CSRF protection to be part of the main test suite
<daviddias>
so that go-ipfs keeps respecting that
therealplato has quit [Quit: Leaving.]
<jbenet>
whyrusleeping: yeah though it may be a bit more complicated than that-- may need special scripts to test the webui works on phantomjs
<jbenet>
daviddias
<daviddias>
<whyrusleeping>
jbenet: hrm... yeah
<whyrusleeping>
i'm not familiar with phantom yet
<daviddias>
phantom is just an scriptable headless browser
<daviddias>
you don't get fancy stuff like WebRTC
<daviddias>
but you can do XHR and UI tests
<jbenet>
btw, daviddias do you need anything in node dht land from go land?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<daviddias>
well, following what we talked and drawed in the whiteboard
<daviddias>
libp2p will mean some changes on how go-ipfs network is implement
<daviddias>
I've been trying to make sure I communicately as effectively as I can to you both to make sure I don't go in a direction that doesn't fit what we expect
<whyrusleeping>
daviddias: +1
<jbenet>
(+1)
<daviddias>
but so far, I've been having positive feedback from both
<jbenet>
yeah i think so far so great
<daviddias>
so, once libp2p.js is done, I'll need also to rewrite/polish the spec we have so we can reimplment it in go and confortably release it to the wild
<jbenet>
daviddias: sgtm!
<jbenet>
cool, well anything else on this? if not let's say done
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
whyrusleeping: re: go-ipfs -- i'm not confident the net is going to stay up and I'm wiped from travel. if you've enough stuff, it may be best to discuss tomorrow? i dont think i'll be very useful atm. (though we can try)
<whyrusleeping>
jbenet: sure thing
<whyrusleeping>
i'm just hacking away on my thing :)
<jbenet>
(i found the bug in get btw, think i fixed it)
<jbenet>
sgtm
<whyrusleeping>
link to the bug?
<whyrusleeping>
i didnt know there was still one
<jbenet>
yeah the halting thing
<whyrusleeping>
ooooh, whats the issue?
<whyrusleeping>
(is it related to the halting on add?)
<whyrusleeping>
jbenet: question if youre still here
simonv3 has quit [Quit: Connection closed for inactivity]
jbenet_ has quit [Ping timeout: 240 seconds]
humanbsd has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
ughhhh, why does the namesys.Resolve not work anymore??
<whyrusleeping>
(not work -> i dont know how to make it work)
<whyrusleeping>
rschulman: i 'fixed' the ipns mount issue you were seeing
<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to feat/mfs: http://git.io/vO0Xg
<ipfsbot>
go-ipfs/feat/mfs a0eae9b Jeromy: make ipns fuse mount use mfs...
<whyrusleeping>
it needs a lot more testing, but the few things i ran worked
<whyrusleeping>
and it also lets you manipulate the /ipns stuff through the mfs API
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
whyrusleeping it doesnt work?? :/
<jbenet>
whyrusleeping idk. probably because we dont have enough tests
<whyrusleeping>
i mean, i think it works
<jbenet>
whyrusleeping the halting thing is a bug in the tarWriter. that thing was a mess. (i fixed it i'll push it out tomorrow)
<whyrusleeping>
its just not the interface i was expecting
<whyrusleeping>
i cant just hand it a key and have it be resolved
<whyrusleeping>
i have to turn the key into an ipns path