<whyrusleeping>
kevina: hrm... it makes sense. And its something i've wanted to do for a little while, but i'm not sure it makes sense for us to do that right now
* whyrusleeping
thinks
<whyrusleeping>
kevina: yeah, lets punt on doing the zero-copy values for now
<whyrusleeping>
I definitely want that at some point, but lets not rock the boat *too* much here
<whyrusleeping>
the reason i'm more hesitant on this one is that misuse of that feature can lead to some pretty subtle bugs
<Kubuxu>
whyrusleeping: what do you think about this idea: instead of gxifying while fuse dep and it's deps, I would vendor it's deps into it and gxify it as a whole.
kulelu88 has quit [Quit: Leaving]
* whyrusleeping
shrugs
<whyrusleeping>
i'm not gonna be violently against that
Qwertie has quit [Ping timeout: 260 seconds]
anonymuse has quit [Ping timeout: 268 seconds]
Qwertie has joined #ipfs
galois_d_ has quit [Remote host closed the connection]
galois_dmz has joined #ipfs
ianopolous__ has quit [Remote host closed the connection]
c5e3 has joined #ipfs
lentan has quit [Remote host closed the connection]
cubemonkey has quit [Ping timeout: 246 seconds]
mildred2 has joined #ipfs
mildred1 has quit [Ping timeout: 268 seconds]
<whyrusleeping>
kevina: I think we should probably automatically call close if we return 'false' from a NextSync call
<whyrusleeping>
does that seem right to you?
<whyrusleeping>
or should we make it the convention that the user always calls 'Close' on the query when they are done?
lentan has joined #ipfs
lentan has quit [Remote host closed the connection]
lentan has joined #ipfs
<jbenet>
whyrusleeping: the most important thing to do on stream muxing is metrics. measurement is what's going to give us a way to compare, adopt, default to one or the other
lentan has quit [Read error: Connection reset by peer]
<whyrusleeping>
jbenet: yeap
lentan has joined #ipfs
<jbenet>
whyrusleeping: we should be able to generate perf graphs of all this stuff, and how they fail-- doing this blind will cause another screw up like go-utp.
<jbenet>
whyrusleeping: a muxer behaving poorly has the potential to be equally bad, possibly worse.
<jbenet>
whyrusleeping: point is really driven home when you look at perf graphs from hard core datacenter systems. not baby WAN distributed systems.
<jbenet>
also, grass always seems greener on the other muxer.
<jbenet>
muxer/transport/etc
ianopolous__ has joined #ipfs
nonluminous has quit [Read error: Connection reset by peer]
<whyrusleeping>
jbenet: we're pretty sure actually that go-utp wasnt the problem and that it was actually yamux
<whyrusleeping>
havent gone back to verify yet, but its pretty sketchy on yamux's part
<jbenet>
whyrusleeping: im not clear on how that would work, given yamux works with tcp but failed fantastically with go-utp.
<jbenet>
whyrusleeping: my bet is that it's actually both are fucking up.
<whyrusleeping>
jbenet: it had to do with timings
<jbenet>
whyrusleeping: both libraries ARE NOT robustly tested. it's honestly bad for yamux to be where it's at and not have standard perf / failure graphs (of the likes you can find for TCP, SSH, SCTP, QUIC,
<jbenet>
"where it's at" = in production in hundreds of thousands of machines.
<jbenet>
^ that's something, btw, that i dont think {muxado, multiplex, spdystream, etc} can claim. and it's a big deal though. it's at least _that_ battle tested.
<jbenet>
don't be so eager to switch it for something that's "so far so good" in thousands of machines.
<jbenet>
need a _rigorous_ perf review to support a claim like "let's use X (battle testing O(1K)) instead of Y (battle testing O(100K))". otherwise you're just fooling yourself.
<jbenet>
be careful-- one thing that sets apart "engineering" from "hacking" is correctness, rigorous proof, a scientific approach to proving what's right.
<jbenet>
(it's why i remotely feel safe trusting something like TLS 1.3, or Zcash, or Signal, or Tor, ... rigorous _engineers_.
<whyrusleeping>
yeap, exactly why i said i'd add it as an experimental option.
<whyrusleeping>
because currently i have no way of even testing performance. I need the ability to do so before i can start getting any miniscule perf data
anonymuse has joined #ipfs
<whyrusleeping>
but currently, yamux is the biggest pain point of the networking stack, and by far the biggest consumer of memory
Qwertie has joined #ipfs
<whyrusleeping>
to reduce memory much lower than the current usage levels I either need to 'fix' yamux, rewrite yamux, or write a new stream muxer
unfrustratable has joined #ipfs
<vtomole>
I'm thinking about using ipfs as a replacement to Dropbox, anything i should consider before making the switch?
chungy has joined #ipfs
trvon has quit [Ping timeout: 258 seconds]
anonymuse has quit [Remote host closed the connection]
<kevina>
whyrusleeping: sorry I was away for a bit, you are giving me conflicting messages
<pjz>
vtomole: I think anything you post is publicly readable unless you encrypt it manually
<whyrusleeping>
kevina: most recent == closest to what i'm actually thinking
chungy has quit [Ping timeout: 260 seconds]
<whyrusleeping>
also, random aside, but eggs, pickles and sriracha make a pretty tasty meal
<whyrusleeping>
either that or my taste buds are broken
lentan has quit [Remote host closed the connection]
<kevina>
okay, no comment on your taste buds :) I don't even know what a sriracha is, sounds weird :)
<pjz>
whyrusleeping: add peanut butter to get a taste of thai :)
<pjz>
kevina: sriracha is "kickin' chicken" sauce :) mildly spicy
<whyrusleeping>
pjz: peanut butter? ooooo! peanut butter is my favorite topping
<whyrusleeping>
and thai is my favorite type of food
<whyrusleeping>
pjz: youre a genius
chungy has joined #ipfs
herzmeister has quit [Quit: Leaving]
lentan has joined #ipfs
<pjz>
whyrusleeping: if you want even more protein, add some tuna
<whyrusleeping>
unfortunately there is no tuna in my house
<whyrusleeping>
the only protein i have are eggs, soylent, and cliff bars
<whyrusleeping>
and i'm running low on at least two of those things
M[m]4 has joined #ipfs
mguentner has quit [Ping timeout: 252 seconds]
<pjz>
soylent is too scary for me
<pjz>
I just get Clif builder bars or other protein bars and jerky or sausage sticks to snack on
<pjz>
also roasted nuts and wasabi peas
<whyrusleeping>
cliff builder bars are awesome
<pjz>
some are. Peanut butter hides lots of crimes.
tmg has quit [Ping timeout: 250 seconds]
<pjz>
tried out some 'chili lime' beef jerky - it's okay, but a bit sweet.
Qwertie has quit [Ping timeout: 258 seconds]
<whyrusleeping>
ooooh, is it that krave brand?
evil_filip has joined #ipfs
<whyrusleeping>
ugh, youre making me hungry for real food
<kevina>
whyrusleeping: so NextSync() should _not_ call close if it will return false?
herzmeister has joined #ipfs
<whyrusleeping>
kevina: I'm thinking so, yeah
<whyrusleeping>
that way we don't provide an excuse for users to forget to call close
<vtomole>
I found this post:https://www.quora.com/What-are-the-major-issues-with-current-operating-system-design-in-terms-of-scalabilty-and-inter-operability-of-OS-between-different-devices. I just thought it was interesting that the problems the answer-er talks about are what ipfs is solving. I've been thinking about why a new operating system would be needed for this.....
jonnycrunch has joined #ipfs
mguentner has quit [Quit: WeeChat 1.6]
jonnycrunch has quit [Ping timeout: 260 seconds]
<pjz>
whyrusleeping: I just ahd v0.4.4 crash on me
<daviddias>
dignifiedquire: I know I asked this before but I'm not sure if I got the answer: When this happens -- https://travis-ci.org/ipfs/js-ipfs-api/builds/176709791 -- it means that you enabled CI to fail if Sauce fails again, right? (because I know there was a version that wouldn't mind if Browser tests failed for the sake of Sauce Labs)
<dignifiedquire>
daviddias: I made is so that if sauce labs fails some browsers it will not fail ci
<victorbjelkholm[>
daviddias: not sure if related, but I noticed that the sauce labs badges stopped working, maybe you changed some project settings?
<daviddias>
"some browsers"
<dignifiedquire>
that log looks like it completely failed to connect to saucelabs
<daviddias>
victorbjelkholm[: I didn't touch anything with regards to SauceLabs' config options
<dignifiedquire>
not even starting the test
<dignifiedquire>
that's why it's failing
<daviddias>
ok, so, when you say "some browsers" is it "if it passes in Chrome and Firefox we are good, if it fails in any of the two then it fails"?
lamarpavel has quit [Ping timeout: 260 seconds]
azdle has quit [Remote host closed the connection]
<daviddias>
I've disabled that test, so everything else should be just fine
<daviddias>
verifying
arpu has joined #ipfs
ylp has joined #ipfs
<daviddias>
got it
<daviddias>
disabled the add test
<daviddias>
forgot to disable the get test for the dir
<haad>
daviddias: new commit incoming?
<daviddias>
yep
jonnycrunch has joined #ipfs
<daviddias>
just to make sure, note that this is just a issue with js-ipfs-api and go-ipfs, js-ipfs-api + js-ipfs adding directories is super fine
jonnycrunch has quit [Ping timeout: 244 seconds]
<haad>
ok
jonnycrunch has joined #ipfs
<yoshuawuyts>
daviddias: when are you arriving in Berlz?
<daviddias>
yoshuawuyts: late night tomorrow, like landing 11pm
<daviddias>
will you be there for DAppHack? :D
<yoshuawuyts>
daviddias: ya they just emailed me to present on dat tech ✨ - figured since you're there I might as well haha
<yoshuawuyts>
^________^
<lgierth>
daviddias: oh on thursday already
<lgierth>
that's cool
<lgierth>
yoshuawuyts: o/
<daviddias>
yoshuawuyts: sweet! :D
jonnycrunch has quit [Ping timeout: 244 seconds]
<yoshuawuyts>
wooooh! - excited to see y'all again :D
<daviddias>
lgierth: yeah, tomorrow already, I guess I should have shared my flight details with you :) :)
<haad>
daviddias: something went wrong with the latest merge commit to js-ipfs-api. it has a change to swarm.js and updating aegir to latest in package.json is missing
<daviddias>
whyrusleeping: I've been seeing an increase of time to execute the IPNS tests in js-ipfs-api, it just took 52 seconds to complete
<haad>
daviddias: nvm re. aegir, it was already updated in master
<victorbjelkholm[>
daviddias yeah, guessing the increase is because the network in general is growing. More nodes = slower ipns
<daviddias>
victorbjelkholm[: js-ipfs-api tests are made in an isolated network
<daviddias>
disconnected from the global one
<daviddias>
victorbjelkholm[: so although you are right, we might see a increase in delay overall, it should not affect js-ipfs-api tests
<victorbjelkholm[>
Ooh, and it almost takes a minute? Wow
maxlath has joined #ipfs
jonnycrunch has joined #ipfs
Qwertie has quit [Ping timeout: 250 seconds]
<daviddias>
dignifiedquire: "WARNING in DedupePlugin: This plugin was removed from webpack. remove it from configuration."
<daviddias>
you probably know what is happening in Webpack land
<daviddias>
I believe we were getting massive bundle rreduction with the DedupePlugin
<dignifiedquire>
well the message is pretty clear, we can drop the dedupe plugin because it is enabled by default now
brimstone has joined #ipfs
jonnycrunch has quit [Ping timeout: 260 seconds]
Qwertie has joined #ipfs
<daviddias>
"was removed" === "enabled by default"?
* Kubuxu
is wondering why ipfs is not reporting local IP address
<Kubuxu>
connecting the ethernet cable fixes the issue :p
<mg>
victorbjelkholm[: how about any mechanism to estimate the availability of some hash in the network? e.g. how many are hosting some file at the moment
<Kubuxu>
mg: you can estimate it a bit with `ipfs dht findprovs` it will give you some estimate
<mg>
Kubuxu: thanks
<victorbjelkholm[>
mg: you can use "ipfs dht findprovs QmVZYWLJoQDxbMEEhLWGKgzz3eJZKKSSQKPdGZQofjiwE6" to find a list of the peers that can provide a value in the dht
<lgierth>
ah i fixed that thing about publishing go-ipfs multiple times
<lgierth>
i failed to return an error
<lgierth>
do you wanna go first with your gx things?
<Kubuxu>
nah, it shouldn't conflict much
<lgierth>
i'm ready ~now
maxlath has joined #ipfs
Qwertie has quit [Ping timeout: 245 seconds]
brianhoffman has joined #ipfs
<brianhoffman>
hey guys quick question
<haad>
ies, all test passing! heads up daviddias and dignifiedquire, there's a fairly fundamental refactoring PR coming in for how we handle streams (and chunked objects) in js-ipfs-api (mainly in request-api.js), along with pubsub
<brianhoffman>
over at OpenBazaar we have a concern raised about people sending in malicious file paths to IPFS that would serve up files on the system outside of the directory structure intended to share
<brianhoffman>
do you all have any protections in placel for that? is that a valid concern?
Qwertie has joined #ipfs
<brianhoffman>
i.e. someone asks to retrieve /../../../familyPics/pic.png and serves up files above the ipfs share?
<brianhoffman>
thoughts jbenet?
<lgierth>
brianhoffman: unless you're using the /ipfs FUSE mount that's not a concern
<lgierth>
since .. is not a valid multihash
<lgierth>
and .. is also not a valid path namespace
<Kubuxu>
lgierth: he means object with link {'../hax': 'QmXXX}
<Kubuxu>
that you would latter `ipfs get`
<lgierth>
well slashes in link names are not addressable
<lgierth>
ah
<lgierth>
slashes in link names are kinda undefined behaviour
<lgierth>
everything expects that it can safely split a path at a forward-slash
<brianhoffman>
so you’re saying it is possible or is not?
Axman6 has joined #ipfs
<brianhoffman>
we’re trying to decide how much wokr we put into doing validation of paths on our side
<Kubuxu>
just hand crafted block like that and it doesn't sanitise in go-ipfs: ipfs object patch add-link QmbFMke1KXqnYyBBWxB74N4c5SBnJMVAiMNRcGu6x1AwQH '../hax' QmeomffUNfmQy76CQGy9NdmqEnnHU9soCexBnGU3ezPHVH
<Kubuxu>
if slug is fully user defined, then they can do whatever they want
<lgierth>
yeah don't do stupid things and say it's ipfs's fault :D:D
<lgierth>
haha
<lgierth>
you don't need the local fs at all if you have everything stored in ipfs
konubinix has joined #ipfs
<brianhoffman>
got you
<brianhoffman>
yes it’s two separate issues i conflated on my side
<brianhoffman>
but the ipfs concern seems unfounded
<brianhoffman>
that’s good. thank you.
<lgierth>
you can put something like varnish in front of the gateway though if that'd speed things up in your case
<brianhoffman>
this is p2p software so people won’t be setting up varnishes in most cases
<brianhoffman>
anyone setting up a full gateway might though
JustinDrake has joined #ipfs
<victorbjelkholm[>
daviddias: can't get js-ipfs to work between node and browser. It's supposed to work right?
<lgierth>
brianhoffman: yeah the full gateway was what i was thinking. if you have lots of ram but only slow spinning disks it's useful to disk accesses
<brianhoffman>
yeah we probably will set one up for our gateway that we have
<brianhoffman>
not that far along yuet
<brianhoffman>
we will be releasing a Milestone 1 version of the IPFS OpenBazaar soon though
<brianhoffman>
probably will mess around with that then
<JustinDrake>
Hi, Justin here from Duo Search (www.duosear.ch), working on OpenBazaar.
<brianhoffman>
^ fellow I was chatting about before
<brianhoffman>
you can read up and see if I expressed your concern properly and clarify if necessary
<JustinDrake>
I've been poking around with IPFS, love it! Very happy with the decision by the OB team to use IPFS.
<lgierth>
brianhoffman: cool, looking forward to M1! :)
<richardlitt>
I'm guessing it doesn't exist, and this is another example of having links to repos which aren't there?
s_kunk has quit [Ping timeout: 260 seconds]
<daviddias>
it doesn't seem that who wrote that file was really pointing to the repos, just creating placeholders, I guess
rgrinberg has quit [Ping timeout: 258 seconds]
Qwertie has quit [Ping timeout: 246 seconds]
sametsisartenep has quit [Quit: leaving]
rgrinberg has joined #ipfs
Encrypt has joined #ipfs
Qwertie has joined #ipfs
<seharder>
OK, I finally got npm install working. :)
ygrek has joined #ipfs
Encrypt has quit [Quit: Quit]
trvon has quit [Remote host closed the connection]
trvon has joined #ipfs
bastianilso__ has joined #ipfs
pfrazee has quit [Remote host closed the connection]
pfrazee has joined #ipfs
strauberry has joined #ipfs
<kenshyx>
seharder: how? :D
pfrazee has quit [Ping timeout: 245 seconds]
<seharder>
kenshyx: that's a good question.
<seharder>
kenshyx: I'll have to go through it again to see what steps are important.
<seharder>
I have a new error now when trying to init a repo: TypeError: base64 is not a function
<seharder>
On this line: /home/sid/git/js-ipfs/node_modules/libp2p-crypto/src/crypto/util.js:18
<seharder>
return new BN(Buffer.from(str, 'base64'))
<seharder>
Is this still an install issue?
bastianilso___ has joined #ipfs
<lgierth>
pls write the node-gyp solution down in that issue +1
bastianilso__ has quit [Ping timeout: 248 seconds]
<richardlitt>
:/
<richardlitt>
Ok. I am giving up on Heroku and running my bot.
<richardlitt>
Which means it will not be a good bot, because it won't be here much. Heroku keeps setting dynamic bots, but I need the range to be restricted for it to work with IRC; I basically can't have ipfs-sprint-helper connect on 6667, so I can't use Heroku to host it.
<richardlitt>
lgierth: Is there another server I can use to run a small node IRC bot?
<whyrusleeping>
richardlitt: sprint time?
<richardlitt>
Yeah; working on getting the video up now
<whyrusleeping>
woo! thanks
pfrazee has joined #ipfs
<richardlitt>
Connectivity in this café is worse than I expected, taking a bit for me to load.
z0mfg has left #ipfs ["memory allocation failure"]
<lgierth>
ok zoom link?
<richardlitt>
Still trying.
sprint-helper has joined #ipfs
<richardlitt>
First time coming to this cafe, not coming again it seems
<lgierth>
can someone else host the zoom call and record?
<richardlitt>
I can't connect to the call
<kevina>
zoom link?
<richardlitt>
I am on, it is just taking ages
<richardlitt>
Someone else will have to initially host it
<richardlitt>
Stream is not going to be up today, as dignifiedquire is not around. Kubuxu, if you could stream, that would be good.
<lgierth>
please mute everybody
<richardlitt>
daviddias: made you the host
<richardlitt>
please record this
<richardlitt>
I am going to have to drop out, my connexion won't take this
<daviddias>
recording
<richardlitt>
Thank you
<daviddias>
I got this! :D
<richardlitt>
The good news is that sprint-helper totally worked!
<victorbjelkholm[>
Woho! Great job
<richardlitt>
But sadly, only while I ran it on my own terminal
<victorbjelkholm[>
fun with the links twice :)
<richardlitt>
If anyone wants to help me host it on a server that doesn't dynamically assign ports (looking at you, Heroku), would be awesome to do that.
<mib_kd743naq>
is there a downside of creating a block full of links that is much closer to the 1mb soft-limit for blocks in general?
<Kubuxu>
mib_kd743naq: there is balance in blocksize
<Kubuxu>
in case of links you want to fan up quite quickly
<Kubuxu>
so you can fetch data blocks
<Kubuxu>
and first data byte
<Kubuxu>
but you don't want to spend ages downloading link block
pfrazee has quit [Quit: pfraze has joined]
<Kubuxu>
when you want to just traverse one branch
<Kubuxu>
imagine tree of depth 4 but with link block size of 1MiB
lacour has quit [Client Quit]
<victorbjelkholm[>
daviddias, in case you missed: regarding website, nothing to preview yet, rushing to get ipld.io finished first :/
<Kubuxu>
or tree of depth 8 but with blocks of 8K
<Kubuxu>
in case of first you will download 4MiB just to traverse to one byte
<Kubuxu>
in the second it will be just 64k.
<Kubuxu>
math might be wrong but this is the concept
<Kubuxu>
mib_kd743naq: ^^
<mib_kd743naq>
Kubuxu: right, but if you want the whole file, you will end up downloading more in case of excessive tree depth, as each link is nontrivial overhead
<mib_kd743naq>
correct?
<Kubuxu>
yes
<mib_kd743naq>
ok thanks
pfrazee has joined #ipfs
trvon has quit [Ping timeout: 256 seconds]
<mib_kd743naq>
Kubuxu: is the "time to first byte" part important? I feel like I am missing something obvious
JustinDrake has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
JustinDrake has joined #ipfs
JustinDrake has quit [Client Quit]
webi has quit [Quit: leaving]
feross has left #ipfs [#ipfs]
<lgierth>
hackmd.io kept confusing my caret
JustinDrake has joined #ipfs
reit has quit [Ping timeout: 260 seconds]
<Kubuxu>
mib_kd743naq: yes, when you are streaming media for exampel
<mib_kd743naq>
nod, makes sense
JustinDrake has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<daviddias>
and it doesn't throw uncaught exceptions anymore and the error is propagated to the callback/promise as expected
<daviddias>
want to try it out?
<pfrazee>
daviddias: much obliged, 1 moment while I clone and set it up
tidux[m] has joined #ipfs
<victorbjelkholm[>
daviddias: maybe you missed my earlier question so I'll ask again. I seem to remember that js-ipfs on node and in the browser should be able to communicate with each other, is that true? If so, I need to do something special? Can't get it to work with my js-ipfs-passing-data example
<victorbjelkholm[>
I'm using the webrtc signaling server for both of them
<daviddias>
victorbjelkholm[: libp2p-webrtc-star works on nodejs now, but libp2p-ipfs hasn't been updated to enable that transport by default
<victorbjelkholm[>
daviddias: sorry, got stuck on something now, can you verify that "swarm peers" is working and printing peers for you?
<daviddias>
pfrazee: interesting :)
<daviddias>
so, no crypto operations ?
<daviddias>
how are you doing the hashing?
<pfrazee>
daviddias: fully story - the webcrypto deps of libp2p-crypto want to build against open ssl. I have to build from source in electron due to funkiness with the ABI. MacOS doesnt use openssl anymore and tries to keep you from doing it. Thus, my problem
<daviddias>
victorbjelkholm[: in js-ipfs latest?
<victorbjelkholm[>
daviddias: yup, in node
<victorbjelkholm[>
cli
<pfrazee>
daviddias: afaict, the ipfs-api only needs libp2p-crypto to generate keypairs, but it *appears* that I dont need to do that. So my stripped build works
<victorbjelkholm[>
doing some debugging and doesn't even seem to call core/components/swarm#peers for me
<daviddias>
pfrazee: I was just mentioning that on the libp2p call, the benefits of using WebCrypto (reduced bundle size, tons of perf improv in crypto operations), has left us with that ugly dependency install
<daviddias>
victorbjelkholm[: ah, you mean, you do jsipfs daemon and then jsipfs swarm peers doesn't show anything?
<pfrazee>
daviddias: yeah. Tough tradeoff. Since I'm just calling out to the go daemon, I think/hope I can avoid it
haad has quit [Quit: reboot]
<daviddias>
victorbjelkholm[: that might just because you are not connected to anyone?
<victorbjelkholm[>
daviddias: yeah, exactly, even with two jsipfs instances locally
<victorbjelkholm[>
no, outputted the array coming from the call, is empty as well
Guest2 has joined #ipfs
Guest2 has quit [Client Quit]
<daviddias>
victorbjelkholm[: you have to connect them
<victorbjelkholm[>
trying to force-connect them with "/ip4/127.0.0.1/tcp/4002" and get "Error: multiaddr not valid"
<daviddias>
jsipfs id, pick one of the addrs of your local network
<victorbjelkholm[>
ay
<daviddias>
you need to include the /ipfs/QmHash
<victorbjelkholm[>
ofc, thanks... :/
<daviddias>
I guess we need better error logs with suggestions
<daviddias>
works now?
<victorbjelkholm[>
silly me, yeah, everything is working
<victorbjelkholm[>
thanks for catching my silly mistakes
<daviddias>
hey, no problem! Happy that is working 😃)
<victorbjelkholm[>
me too! :D
Guest2 has joined #ipfs
<daviddias>
once that go-ipfs 0.4.5 is released (which means interop with js-ipfs) I'll add update and add mdns again
<kevina>
whyrusleeping: any other place where the datastore Query method is used (other then the Blockstore and providers)?
<daviddias>
so that it is all seemless for the discovery of LAN nodes
ylp has quit [Ping timeout: 260 seconds]
ylp has joined #ipfs
JustinDrake has joined #ipfs
rendar has quit [Quit: Leaving]
brianhoffman has quit [Read error: No route to host]
ulrichard_ has joined #ipfs
mib_kd743naq has quit [Ping timeout: 260 seconds]
ianopolous has joined #ipfs
JustinDrake has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
ulrichard_ has quit [Remote host closed the connection]
Guest2 has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Qwertie has joined #ipfs
ashark has quit [Ping timeout: 244 seconds]
slothbag has joined #ipfs
JustinDrake has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<whyrusleeping>
kevina: I think thats it
PseudoNoob has quit [Remote host closed the connection]
<kevina>
whyrusleeping: okay, everything should be good to go pending review...
<kevina>
there are 5 related pull requests, two of them of in go-ipfs
quim has quit [Ping timeout: 265 seconds]
rgrinberg has quit [Remote host closed the connection]
Encrypt has quit [Quit: Sleeping time!]
Powersource has joined #ipfs
elico has quit [Quit: Leaving.]
JustinDrake has joined #ipfs
vtomole has quit [Ping timeout: 260 seconds]
A124 has quit [Remote host closed the connection]
ulrichard has quit [Remote host closed the connection]
A124 has joined #ipfs
wallacoloo_ has quit [Quit: wallacoloo_]
<Powersource>
does js-ipfs work by going through ipfs.io/ipfs?
galois_d_ has joined #ipfs
ZaZ has quit [Read error: Connection reset by peer]
cubemonk1y has quit [Read error: Connection reset by peer]
galois_dmz has quit [Ping timeout: 260 seconds]
Kane` has quit [Remote host closed the connection]
<Mateon1>
Powersource: No, it runs an actual node.
<Mateon1>
If you want something more lightweight, you can run a service worker and capture network requests. If it's an IPFS request you can try redirecting it to (in order): a browser builtin with fs:/ipfs/... ipfs:// ; a localhost gateway; and ipfs.io.
<Powersource>
Mateon1: but how does it work if it's limited to http?
<Mateon1>
Powersource: It's not limited to HTTP, it uses web sockets (to go-ipfs nodes) and WebRTC to other js nodes
<Powersource>
oh damn I need to read up
Qwertie has quit [Ping timeout: 240 seconds]
<Mateon1>
daviddias: Care to upload the sprint call recording to IPFS?
Qwertie has joined #ipfs
JustinDrake has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<seharder>
daviddias: I just finished testing js-ipfs on ubuntu with ipv4 and ipv6 on port 5001, daemon is not choking. :)