<pinbot>
failed to pin Qmd1usDGx7zDU5zq3M4WQft6ENVQJtzc8XNYy4qzPEgKWE: pin: context deadline exceeded
<whyrusleeping>
thanks pinbot
<tperson>
krl: So if we do everything by the book with ReactJS we could use Jest. I'ved look at it a bit and while it's neat I'm also not keen on jumping into something restrictive.
pinbot-test has quit [Remote host closed the connection]
<tperson>
I did a bit of research with Mocha and some other tools to allow you to test ReactJs components, and it appeared to be a good approch.
<headbite>
have the pinbot do refs first
<whyrusleeping>
headbite: thats what pinbot-test is running
<whyrusleeping>
pushing to pinbot now
<tperson>
That would give us component testing, we would almost probably want to do some e2e testing using something like webdriver.io
<whyrusleeping>
alright, watch carefully kids, lets see if anyone notices pinbot being updated ;)
notduncansmith_ has quit [Read error: Connection reset by peer]
<tperson>
Gone are the days of getting setup quickly with node (and even javascript) and playing around with things.
www1 has quit [Ping timeout: 264 seconds]
notduncansmith_ has joined #ipfs
notduncansmith_ has quit [Read error: Connection reset by peer]
<whyrusleeping>
tperson: :/
<krl>
tperson: by the book must not be the case if we find a simpler model, but i'll look at it
<krl>
think actually the api is more important to get down and tested first
<tperson>
krl: agree, having a solid test for the API would be great.
<krl>
will have to look at that soon, since some things seem to be broken still
<jbenet>
oooooof travis!
<tperson>
We really should test the API via something like Go with strongly typed struchures to make sure we aren't throwing anything extra down the pipe we don't expect. Go should (correct me if I am wrong) should catch extra fields coming down the pipe it doesn't expect, or if it doesn't we probably can easily test for them. It will also help with type checking.
<tperson>
Unless there is a better tool for testing rest APIs. There might be an opportunity to make one.
<Tv`>
tperson: encoding/json and protobuf both happily ignore extra fields
<tperson>
Darn, in this case, that might not be ideal.
<tperson>
mafintosh: Just saw your chromecasts repo <3
<krl>
jbenet: could you give me +w to the node-ipfsd-ctl repo?
<ipfsbot>
[go-ipfs] jbenet pushed 1 new commit to master: http://git.io/vT2eU
<ipfsbot>
go-ipfs/master 1c35377 Juan Batiz-Benet: Merge pull request #1256 from ipfs/fix/init...
notduncansmith_ has joined #ipfs
notduncansmith_ has quit [Read error: Connection reset by peer]
<jbenet>
krl whyrusleeping did we add that option to disable api access blocking?
<jbenet>
(for testing new webui hash)
<krl>
jbenet, yes
<krl>
it's in the PR
<whyrusleeping>
^
<jbenet>
krl: maybe want to submit the go-ipfs branch as we test
<jbenet>
ah
<whyrusleeping>
pinned the new webui on mars
<jbenet>
dddddderp
<whyrusleeping>
it looks small though
patcon has joined #ipfs
<krl>
i'
<krl>
i'd like to make ipfs-geoip more useful too
notduncansmith_ has joined #ipfs
<krl>
make it updatable with structural sharing
notduncansmith_ has quit [Read error: Connection reset by peer]
<krl>
and maybe even provide historical information in a ui
<krl>
could actually be pretty useful as a tool on its own
<whyrusleeping>
krl: +1
<whyrusleeping>
i'll also add you to pinbots happy list
notduncansmith has quit []
<tperson>
So, one thing we should look at for a lot of these node packages is having ipfs-api as a peer dependencies instead of an actuall dependencie.
<krl>
tperson: i'm unfamiliar with the distinction
<tperson>
Best to just read the blog post they made about it, as Im not sure if I'd be able to really explain it very well.
<jbenet>
krl: thanks sorry we couldn't push it out yet!
<ipfsbot>
[go-ipfs] jbenet closed pull request #1235: Fix: Using the `dnslink` feature led to infinite redirects (master...fix/dnslink-issues) http://git.io/vTejA
<jbenet>
whyrusleeping: i'm going to kill go 1.3 from travis for now
<ipfsbot>
[go-ipfs] jbenet pushed 2 new commits to master: http://git.io/vT20y
<Luzifer>
\o/ #1235 - now I can move my swagger-api-docs to IPFS! \o/
patcon has quit [Ping timeout: 246 seconds]
<jbenet>
\o/
<Luzifer>
lulz. just had to wait for my own build to move to the begin of the build queue :D
<Luzifer>
nice to see gobuilder is active use <3
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<ipfsbot>
[go-ipfs] jbenet created add-circleci (+1 new commit): http://git.io/vT2zk
<ipfsbot>
go-ipfs/add-circleci 6fb31b9 Juan Batiz-Benet: add circleci support
<ipfsbot>
[go-ipfs] jbenet force-pushed add-circleci from 6fb31b9 to 05b8506: http://git.io/vT2zm
<ipfsbot>
go-ipfs/add-circleci 05b8506 Juan Batiz-Benet: add circleci support
Tv` has quit [Quit: Connection closed for inactivity]
ei-slackbot-ipfs has quit [Remote host closed the connection]
ei-slackbot-ipfs has joined #ipfs
Wallacoloo has quit [Ping timeout: 256 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<Luzifer>
damn. path does not work as expected :(
rht__ has joined #ipfs
<Luzifer>
oh. wking something broke hostname resolution completely I think... I'm getting an 404 and namesys does not do any DNS resolve...
<rht__>
is there ipfs 'directory' repo where I can browse neat files? otherwise I have to just keep using QmQv4YQNmRPuTTHs4AgBhKEFDdN7eQYeTbSmr8JVWVfury/sintel.mp4 for my testing
<rht__>
(some wikileaks documents maybe?)
<doublec>
rht__: This one contains a bunch of thumbnails and pics of pitcairn island QmWom8ULsNCbtMgNMk82aJT7G6gU55mSGxVaC2i1fuyX3s
notduncansmith has joined #ipfs
<rht__>
gotcha, lmk if you have more
notduncansmith has quit [Read error: Connection reset by peer]
<ipfsbot>
[go-ipfs] jbenet force-pushed add-circleci from 05b8506 to 1fcde90: http://git.io/vT2zm
<ipfsbot>
go-ipfs/add-circleci 1fcde90 Juan Batiz-Benet: add circleci support
<doublec>
If I ipfs add something, is it automatically pinned?
<jbenet>
rht__ maybe we should add some links to github.com/ipfs/examples
<Luzifer>
something being in place some days ago redirecting dnslink-domains into the ipns_hostname thing is not longer there or changed its behavior
<jbenet>
Luzifer: probably unrelated. but i recall the mux is being manipulated there very oddly (different muxes get created instead of all mounting into one)
<rht__>
whyrusleeping: I need to get total file length of the hash
<jbenet>
Luzifer: did you try adding log statments there and it doesnt get called?
<whyrusleeping>
the pbdata.GetFilesize should work. is it not?
<rht__>
nope
<rht__>
returns 0
<whyrusleeping>
odd
<whyrusleeping>
are you calling it on a unixfs file?
<whyrusleeping>
or is it a directory?
<rht__>
I use file specifically QmQv4YQNmRPuTTHs4AgBhKEFDdN7eQYeTbSmr8JVWVfury/sintel.mp4
<rht__>
`ipfs get QmQv4YQNmRPuTTHs4AgBhKEFDdN7eQYeTbSmr8JVWVfury/sintel.mp4`
<Luzifer>
jbenet: the only change in the daemon is the unrestricted api access stuff... also the only change in ipns_hostname is my change... (speaking in terms of changed in 2015-05)
<jbenet>
Luzifer but did you check whether it's called? (trying to bisect the logic)
<whyrusleeping>
rht__: thats odd, i'm seeing the pbdata.Filesize set properly there... let me look into it a bit more
<whyrusleeping>
Luzifer: on my end, the daemon is hanging looking for a block while resolving your entry
notduncansmith has joined #ipfs
<Luzifer>
hmmmh with latest master code?
<jbenet>
Luzifer: also worht trying it again commenting out // corehttp.GatewayOption(writable) -- so that we see if it is or not the odd mux problem
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
Luzifer yeah
flugsio has joined #ipfs
<jbenet>
(ignore that "yeah")
<whyrusleeping>
rht__: do you ever set your pbdata in that reader?
<Luzifer>
hmm the ipns_hostname stuff is called
<Luzifer>
(see latest comment
<jbenet>
whyrusleeping is the manners thing why tests are failing? should we just revert for now?
<Luzifer>
((yeah fmt.Printf debug))
<whyrusleeping>
jbenet: tests failing?
<whyrusleeping>
i dont think so
ei-slackbot-ipfs has quit [Remote host closed the connection]
<Luzifer>
wow. half of namesys/dns.go has been rewritten
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #1270: do http server properly so daemon can shut down (master...fix/daemon-shutdown) http://git.io/vT2yi
<whyrusleeping>
jbenet: that first one, i have no idea whats happening
<rht__>
whyrusleeping: ahhh maybe, I thought I already put it in writetobuf, wait a sec
<jbenet>
whyrusleeping: that's your new perm stuff -- wondering whart's up there
<whyrusleeping>
yeah... i've no clue yet
<rht__>
whyrusleeping: still 0
<whyrusleeping>
rht__: code?
<whyrusleeping>
do you have a branch pushed somewhere?
<Luzifer>
jbenet: error is in namesys/dns.go... "not a key"
<whyrusleeping>
i'd just change n.Namesys.Resolve(ctx, host) to n.Namesys.Resolve(ctx,"/ipns/"+ host)
<jbenet>
we should verify it's a domain before doing that.
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<Luzifer>
jbenet: nope! that adds a redirect... very bad... just no... swagger.luzifer.io => swagger.luzifer.io/ipfs/QmczrnBpAgnhCo87rtZ8VbLfGU5v8g7cKr7Ef2fKRvYniA/
<jbenet>
huh redirect?
<Luzifer>
yeah
<Luzifer>
302 => path resolved from ipns
<jbenet>
the problem is that the resolver no longer matches everything as it used to
<jbenet>
wking's changes make it be an absolute path
<jbenet>
and force things to start with /ipns/
<Luzifer>
*sighs*
<Luzifer>
yeah but somehow it
<Luzifer>
is throwing the hostname into path/path.go
<jbenet>
(hmmm i dont think it should be doing this here)
<jbenet>
Luzifer: i dont understand your redirect loop?
<Luzifer>
not a loop
<Luzifer>
but a redirect
<whyrusleeping>
this is what i meant when i was wanting more tests around resolution
* whyrusleeping
sighs
<ipfsbot>
[go-ipfs] jbenet closed pull request #1270: do http server properly so daemon can shut down (master...fix/daemon-shutdown) http://git.io/vT2yi
<jbenet>
cc wking
<jbenet>
o/
<whyrusleeping>
jbenet: so how does ipns 'dnslink' stuff work now?
<whyrusleeping>
is it no longer /ipns/domain ?
<jbenet>
Luzifer oh because you dont want to redirect from the host-- right. this is a separate problem.
<Luzifer>
whyrusleeping: get domain.tld => dig txt domain.tld => split dnslink=... on = => deliver rest of the entry as a path
<Luzifer>
in short
<whyrusleeping>
okay
<whyrusleeping>
so thats what i remember
<Luzifer>
and somehow at any place it's treating the domain name as a ipfs hash...
<whyrusleeping>
that ipns hostname handler needs to be shortened to return core.Resolve("/ipns/" + host)
rht_ has quit [Ping timeout: 246 seconds]
<jbenet>
Luzifer did you see the line i pointed to?
<jbenet>
that's why
<jbenet>
my patch has to happen
<jbenet>
preventing the redirect is a separate step
<Luzifer>
Lq
<jbenet>
Lq?
<Luzifer>
:q with a typo in the wrong window
<jbenet>
haha
<jbenet>
now to find the redirect...
<ipfsbot>
[go-ipfs] whyrusleeping created fix/shutdown-err (+1 new commit): http://git.io/vT2FF
<ipfsbot>
go-ipfs/fix/shutdown-err c84819f Jeromy: silence error log on shutdown
ei-slackbot-ipfs has quit [Remote host closed the connection]
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #1271: silence error log on shutdown (master...fix/shutdown-err) http://git.io/vT2Fx
ei-slackbot-ipfs has joined #ipfs
<Luzifer>
jbenet: got a solution
<Luzifer>
jbenet: QmdaRVbYUmycA3FXQ7HQfjdLnVoHVikTYb7zeeSaQizZBL < take that. works.
<Luzifer>
based on your solution extended with my solution for #1234
<Luzifer>
and it's not plain water i'm drining other times
<Luzifer>
(okay gased water! not the stuff without gas I heared americans like more)
<jbenet>
Luzifer: i still get one 302 Found with your patch
<Luzifer>
jbenet: what are you calling?
<whyrusleeping>
gased water?
<whyrusleeping>
like, carbonated?
<Luzifer>
sparkling water
<Luzifer>
dunno from what language I got that gased... O_o
<whyrusleeping>
i dont get the appeal, just give me ice cold plain real water any day :P
<whyrusleeping>
lol
gozala has quit [*.net *.split]
wiretapped-cb has quit [*.net *.split]
lachenmayer has quit [*.net *.split]
Skaag has quit [*.net *.split]
<whyrusleeping>
anyways, its finally time for me to sleep
Skaag has joined #ipfs
wiretapped-cb has joined #ipfs
<whyrusleeping>
i cant imagine what time it is for both of you right now
<Luzifer>
okay have a good night :)
<Luzifer>
8:43 am
gozala has joined #ipfs
<whyrusleeping>
lol, okay, so youre up for the morning :P
<whyrusleeping>
gnite!
octalberry has joined #ipfs
lachenmayer has joined #ipfs
Wallacoloo has joined #ipfs
<jbenet>
Luzifer: PR pls? also we should make one function that does this everywhere.
<jbenet>
Luzifer: worth double checking the path, too.
Wallacoloo has quit [Client Quit]
<jbenet>
whyrusleeping niiight
<jbenet>
its almost 3 here
<ipfsbot>
[go-ipfs] Luzifer opened pull request #1272: Fix: dnslink domain resolving was broken; Add: no caching for those (master...fix/dnslink) http://git.io/vT2pZ
<jbenet>
luzifer come to berlin this weekend!
<Luzifer>
public transit in germany just broke down yesterday xD
<jbenet>
oh no! how come?
<Luzifer>
the "Deutsche Bahn" is in strike...
<Luzifer>
so no train transits...
<Luzifer>
and by car it's a hell of a trip... 300km oneway
mappum has quit []
octalberry has quit [Ping timeout: 258 seconds]
mappum has joined #ipfs
<jbenet>
Luzifer: we should be using path.Clean
<Luzifer>
the path module messes up as soon as a trailing slash is present
<jbenet>
even clean?
ei-slackbot-ipfs has quit [Remote host closed the connection]
<Luzifer>
we need that one for http://swagger.luzifer.io/ < r.URL.Path == / if it's stripped there are redirects as hell
ei-slackbot-ipfs has joined #ipfs
<Luzifer>
so yeah just taking "name" instead of path.Join should work fine
<jbenet>
ah right.
notduncansmith has joined #ipfs
<Luzifer>
updated PR
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
Luzifer: URL.Path is guaranteed to always have prefix / ?
<Luzifer>
it had in every case I saw until now
<jbenet>
Luzifer: maybe include a test or something? so we dont screw this up?
<Luzifer>
we would need a whole-stack-test
<Luzifer>
there are tests checking for different things with those dnslink entries but they did not catch this because it was domain parsing :(
<ipfsbot>
[go-ipfs] jbenet deleted fix/shutdown-err at bd30866: http://git.io/vT2hp
<jbenet>
Luzifer we should do a test like your curl test
Wallacoloo has joined #ipfs
<jbenet>
Luzifer with sharness
<Luzifer>
but thats a bigger test including DNS mock
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
sigh the permissions think broke
lgierth has joined #ipfs
pfraze has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
pfraze has quit [Remote host closed the connection]
lgierth has quit [Quit: Ex-Chat]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
rht__ has joined #ipfs
<cryptix>
hello ppl
<rht__>
whyrusleeping: just saw ur msgs https://botbot.me/freenode/ipfs/msg/39765345/ +- 5 min, thanks. iirc there is supposed to be a precomputed metadata that has the info of the total size of a given node?
octalberry has joined #ipfs
<rht__>
whyrusleeping: NodeStat.CumulativeSize?
RzR has quit [Excess Flood]
RzR has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
atomotic has joined #ipfs
domanic_ has quit [Ping timeout: 255 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Wallacoloo has quit [Ping timeout: 256 seconds]
chrisr_ has joined #ipfs
lgierth has joined #ipfs
www has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<rht__>
I also wonder IF ipfs makes the web easier to index? (or this is an orthogonal concern because crawlers generally traverse through hyperlinks, not node links)
Wallacoloo has joined #ipfs
notduncansmith has joined #ipfs
hellertime has quit [Quit: Leaving.]
notduncansmith has quit [Read error: Connection reset by peer]
robsyme has left #ipfs [#ipfs]
u7654dec has joined #ipfs
<rht__>
on web vs nodehash-links: the point is, however, semantic similarity != byte sequence similarity (though their kolmogorov complexity could be correlated)
ei-slackbot-ipfs has quit [Remote host closed the connection]
ei-slackbot-ipfs has joined #ipfs
u7654dec has quit [Read error: Connection reset by peer]
saebekassebil has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
chrisr_ has quit [Ping timeout: 264 seconds]
lgierth has quit [Quit: Ex-Chat]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
rht__ has quit [Quit: Page closed]
RzR has quit [Ping timeout: 276 seconds]
Wallacoloo has quit [Quit: Leaving.]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
atomotic has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
atomotic has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<Blame>
The point in multiple DHTs is rally to call attention to the fact that they all work on the same premise, and really on provide optimizations for specific network usage cases and topologies
<zorun>
Blame: I'm just generally happy to see people using a DHT in real life (in this case, http://ring.cx/ ), I'm not developping one
<Blame>
*really
<zorun>
interesting
<Blame>
Essentially the idea is to make the developer overhead from "idea involving DHT" -> "thing that works involving DHT" as easy as possible
<Blame>
but without assuming every use case need the same DHT mechanism
<zorun>
actually, I'm quite interested in latency (I did a research project a few years ago, where we took into account latency for routing decision)
<zorun>
and I wondered if we could use the same ideas to favour low-latency peers in a DHT
<Blame>
I'm building a DHT as part of my dissertation that embeds nodes into a hyperbolic point space based on inter-node latency
<zorun>
but since Kademlia's original paper already mentions this idea, I guess it's not very new :)
<zorun>
oh, interesting
<Blame>
then nodes greedily route along sorest routes
<zorun>
so basically, you shape the keyspace according to latency?
<whyrusleeping>
zorun: we plan on using a coral like clustering system to improve network latency
<Blame>
essentially. We move servers around in real time to respond to latency
<zorun>
hmm, ok, interesting
<Blame>
because latency changes a lot over time
<Blame>
so it is a realtime distributed embedding of the latency graph into a metric space
<whyrusleeping>
wking: let me know when you start working on a docker driver
<zorun>
yes, that was my question, it seems difficult to change the keyspace over time
domanic_ has quit [Ping timeout: 264 seconds]
<Blame>
you need to be a bit more aggressive about backups, but otherwise it seems to work fine in simulation
<Blame>
we will see about application soon
jholden has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<wking>
whyrusleeping: hopefully starting the Docker-registry backend tomorrow. I'll try to post what I can of an adder-API proposal today.
<whyrusleeping>
okay, cool
<wking>
whyrusleeping: jbenet said you'd done some initial work on one with him, but he didn't have a link to any code you'd been working on
<wking>
do you have a repository somewhere, or should I start from scratch?
<Blame>
zorun: the callbacks in openDHT would not be hard to add, but would not be very exciting in go or python
<whyrusleeping>
wking: that depends on what exactly youre trying to do
<whyrusleeping>
which is why i wanted to chat about it
<cryptix>
hello guys
<whyrusleeping>
cryptix: heyo!
mildred has joined #ipfs
vonzipper has quit []
vonzipper has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
lgierth has quit [Quit: Ex-Chat]
<ipfsbot>
[go-ipfs] whyrusleeping force-pushed refactor/importer from f6aa8b2 to 89d7625: http://git.io/vTrp7
<ipfsbot>
go-ipfs/refactor/importer 89d7625 Jeromy: change pinning to happen in a callback
pfraze has joined #ipfs
<ipfsbot>
[go-ipfs] whyrusleeping force-pushed rm-testing from 7dc767e to 3188dc1: http://git.io/vT2Pa
<ipfsbot>
go-ipfs/rm-testing 3188dc1 Jeromy: move core mock into its own package
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]
Tv` has quit [Ping timeout: 255 seconds]
pfraze has quit [Remote host closed the connection]
Tv` has joined #ipfs
kyledrake has joined #ipfs
<kyledrake>
g'day
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
kyledrake: hey there!
<whyrusleeping>
long time no see
<kyledrake>
Hi! I'll be around all day today.
<whyrusleeping>
awesome, whatcha up to?
<whyrusleeping>
jbenet: my desktop node had no bootstrap peers for a few days and i didnt notice because it managed to bootstrap itself through mdns and my laptop
<kyledrake>
whyrusleeping Today I'm doing the IPFS integration on Neocities. First the data model, then the async workers
<kyledrake>
whyrusleeping after that I have to figure out how to stop IPFS from "portscanning" to stop pissing off our hosting provider.
<kyledrake>
I high-level believe I can use iptables for this, but I'm not sure yet.
<whyrusleeping>
kyledrake: if the port scanning thing is an issue for you i can work on that today
<kyledrake>
whyrusleeping our DC operator gave us one day to "fix it" or they were going to turn off the server :/
<whyrusleeping>
kyledrake: do you have the list of ip's from your isp?
<whyrusleeping>
similar to the one posted in the issue
<kyledrake>
if you send me your email I can give you the letter they sent us
<whyrusleeping>
why at ipfs.io
<Tv`>
jbenet: stumbled on this line from irc logs: 13:49 <jbenet> also-- this isn't there yet, but whyrusleeping and i are talking about (a) making most of our use of leveldb into ipfs objects --and thus blocks-- (like ipfs pin data for example) -- and (b) swapping out boltdb for leveldb later on.
<kyledrake>
sent
<Tv`>
jbenet: you really don't need leveldb or bolt; there's so little non-block stuff in the datastore already, flatfs etc can handle that just fine
<Tv`>
jbenet: i wish i had realized that when we were talking about Mount etc, i would have just replaced all of leveldb with flatfs
<Tv`>
but i didn't
<whyrusleeping>
Tv`: you think so? even the dht records?
<Tv`>
whyrusleeping: i mean, all designs are to be adjusted if you jump multiple magnitudes bigger
<Tv`>
whyrusleeping: but what i'm seeing now is someone lighting a match in a blast furnace
<Tv`>
it's just not gonna matter
<whyrusleeping>
kyledrake: those are *all* docker addresses being dialed...
* whyrusleeping
sighs
notduncansmith has joined #ipfs
<jbenet>
(Mid travel will have more bw in a min) I think we should have a simple iptables script first
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
That will be way faster than trying to do this filtering correctly, in the config and so on
<kyledrake>
whyrusleeping That's odd. I'm not running docker on that machine.
<jbenet>
and we can play with what subnets to filter out there first
<jbenet>
kyledrake: it's docker on other people's computers
<Luzifer>
whyrusleeping: what are "docker addresses"?
<kyledrake>
ah.
<Tv`>
gotta say lxd was so nice, you just associate a "foo:" prefix with a image service url
<whyrusleeping>
172.17.*
<Luzifer>
nah thats not exclusively docker. I've got two "DC"s running on that range
<Luzifer>
(in reality sub-parts of DCs but nevermind)
<jbenet>
Somebody with iptables or whatever experience here?? Their help would be very useful
<Tv`>
*ahem*
* whyrusleeping
feigns ignorance
* jbenet
goes through "security"
<jbenet>
Much theater. So fake. Wow slooow
<Luzifer>
:D
<Luzifer>
at frankfurt airport I was carrying multiple metal objects through the metal-scanner... they didn't care... xD
<Luzifer>
at hamburg airport I had a pack of "fishermens" in my pocket and they nearly got a heart attack and brought multiple hand scanners....
<Tv`>
for the record, you are not allowed to use a cellphone in the waiting area of a security checkpoint
<Luzifer>
you aren't? nobody told me...
<Tv`>
because terrorists can tell other terrorists deep secrets
<Tv`>
same reason why you are not allowed to leave the checkpoint area (backwards) once you enter
<Tv`>
enforcement is a different matter, of course
<Tv`>
in the meanwhile, Oakland airport employees were smuggling pot past security checkpoints
<Luzifer>
:D
<mmuller_>
To be fair, they should't be checking for pot.
<Tv`>
yeah, sure, just pot is a substance for which many people would take high risks
<whyrusleeping>
i managed to carry a switchblade through security from san diego to seattle once
<whyrusleeping>
not sure how i pulled that off
<Tv`>
so if that's risked with pot, you can imagine what kind of risks are taken to smuggle the real bad stuff
<Tv`>
whyrusleeping: there was a time when i always had a leatherman of some sort on me, even when flying.. those days are gone
<jbenet>
Anyway if someone made a sh script with iptables rules that blocked ipfs from dialing certain subnets that be awesome
<mmuller_>
There was one guy who accidentally brought a gun on an entire flight in his carry-on.
tilgovi has joined #ipfs
<Luzifer>
hmm maybe I should implement ipns targeting in my dockerproxy... that way I could finish my api-docs-in-ipfs project... :(
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
ei-slackbot-ipfs has quit [Remote host closed the connection]
<jbenet>
or-- is it that the error needs the change?
mildred has quit [Quit: Leaving.]
www has quit [Ping timeout: 276 seconds]
<jbenet>
whyrusleeping: the bitswap hang is concerning. this is -- i belive -- after bumping down the instances on travis to 200x200
<jbenet>
whyrusleeping: looks like it isn't making forward progress or something?
<jbenet>
not sure.
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
tilgovi has quit [Ping timeout: 246 seconds]
Wallacoloo has joined #ipfs
patcon has quit [Ping timeout: 256 seconds]
tilgovi has joined #ipfs
octalberry has joined #ipfs
<jbenet>
whyrusleeping i _want_ to merge 1218. idk about the bitswap hang-- i think in general it works better than master
<jbenet>
whyrusleeping so we could merge and go from there? -- i do worry about regressions.
<krl>
jbenet: is the ipns spec thing still in your pipeline btw?
<jbenet>
yes :/
<jbenet>
:(
<jbenet>
need to try to do less.
<krl>
jbenet: let's just chat about it when we meet
<jbenet>
sgtm.
<krl>
have some thought also, as you might have expected :)
<jbenet>
cryptix, krl: let's get together in berlin tomorrow. I'll be free to hang out probably around 3pm.
<jbenet>
(maybe earlier)
<krl>
brb
<krl>
but sounds good
hellertime has quit [Quit: Leaving.]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
therealplato has joined #ipfs
<ipfsbot>
[go-ipfs] jbenet pushed 1 new commit to master: http://git.io/vToxT
<ipfsbot>
go-ipfs/master 5853eac Juan Batiz-Benet: Merge pull request #1218 from ipfs/refactor/bitswap...
nsh has quit [Excess Flood]
nsh has joined #ipfs
mildred has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
tperson krl or whyrusleeping could you look at kyledrake's json problem? And the iptables script? Would be huge for them to have both of those fixed.
<nell>
kyledrake why does your name sound familiar
<nell>
are you @kyledrake
<kyledrake>
nell ja.
<nell>
Oh cool, I follow ya on twitter thats how I knew ya
<nell>
I'm @themodernend
<kyledrake>
added :)
<whyrusleeping>
kyledrake: whats higher priority for you, the add json thing, or the netscan?
<kyledrake>
whyrusleeping netscan by a mile.
<nell>
Fancy seeing you interested in ipfs.
<whyrusleeping>
kay
<kyledrake>
whyrusleeping I'm implementing a parser for the current output standard in the code right now, so it's not a show stopper.
<nell>
I'm working on a metaverse project and see IPFS as a possible infrastructure module to building it.
<nell>
If you're familiar with piratebox [which I'm sure you are] it works similarly to that, but the offline [mesh] networks will host content relevant to immersive 3D graphics environments
<kyledrake>
nell ah, cool
<nell>
It's inspired by the metaverse detailed in Snow Crash, mixed with Gibson and cypherpunk ethos :P
<whyrusleeping>
one thing i was thinking was that since each person has their own keyspace for storage in ipfs, anyone who wants their own 'place' in the metaverse just needs to have the files under their ipns entry in a special location
<whyrusleeping>
like /ipns/<myhash>/.metaverse
<nell>
Hah kyledrake I think I'll use coinpunk to test the use of cryptocurrency in mixed reality, it seems almost too perfect.
mildred1 has joined #ipfs
<whyrusleeping>
nell: i just really like the idea of having my terminals in a room that i can walk around and look at/type on
<kyledrake>
nell sorry just got back from fixing a salad
mildred has quit [Ping timeout: 256 seconds]
<kyledrake>
I don't recommend using Coinpunk, it's quite out of date at this point.
<kyledrake>
You're better off making something with bitcoinjs
<nell>
It was a great proof of concept at the time
<kyledrake>
Yeah. I had to choose between things due to my workload unfortunately. I could update that, but there's a lot of people doing wallet stuff so plenty of choices now.
<whyrusleeping>
Tv`: ping
* nell
wonders if anyone made a docker container for bitcoinjs
<whyrusleeping>
jbenet: i dont think that iptables is the right fix for the port scanning issue...
<whyrusleeping>
:/
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]
<whyrusleeping>
because you can pretty easily block outgoing traffic to a given subnet:
mildred1 has quit [Ping timeout: 256 seconds]
<whyrusleeping>
iptables -A OUTPUT -d 192.168.0.0/16 -j REJECT
<whyrusleeping>
which... works
<whyrusleeping>
but its a very invasive way to solve an issue with our code
mildred has joined #ipfs
<kyledrake>
hmm.
<kyledrake>
I could try getting my ISP to stop complaining about "net scans" from that machine too.
<whyrusleeping>
well, in either case, i would explain to them whats actually happening
<whyrusleeping>
and then i would use iptables to block places that are 'scanned' frequently
<whyrusleeping>
like the 172.17 range
<whyrusleeping>
and then 192.168 range
<kyledrake>
Works for me.
<whyrusleeping>
the next step for us would be a config setting to specify ranges not to dial
<whyrusleeping>
which is a very good choice between iptables and actually being smart about things
<alexandria-devon>
hey Juan - a bunch of our guys are getting "no Namesys on IpfsNode - can't resolve ipns entry" when they do "ipfs init" on mac os x 10.10.3 and 10.10.2
<alexandria-devon>
known bug?
lgierth has joined #ipfs
<kyledrake>
alexandria-devon I just noticed that one too.
<kyledrake>
oh, right
<kyledrake>
alexandria-devon: run ipfs daemon &, then run it.
<alexandria-devon>
k
<kyledrake>
The example doesn't say to start the daemon first. :)
<whyrusleeping>
alexandria-devon: thats new...
<alexandria-devon>
that was it
<alexandria-devon>
i think there was no reliance on IPNS in previous builds
www has joined #ipfs
<alexandria-devon>
makes sense that daemon needs to be running to resolve IPNS names
* whyrusleeping
is so confused
<whyrusleeping>
but you have to init before you can run the daemon?
EricJ2190 has joined #ipfs
Wallacoloo has quit [Quit: Leaving.]
<alexandria-devon>
you do init, and it gives you a line to run
<alexandria-devon>
so the order should be, do init, then run daemon in another terminal window, then do the line the init gave you to run
<whyrusleeping>
oh
<whyrusleeping>
that...
<whyrusleeping>
wking: something broke in resolution
<whyrusleeping>
alexandria-devon: that is a bug, thanks for bringing it to our attention :)
* whyrusleeping
gets to patching that
<alexandria-devon>
1) ipfs init 2) ipfs daemon in a new window 3) ipfs cat /ipfs/[hash]/readme
<alexandria-devon>
or fix bug ;)
<alexandria-devon>
thanks for your help guys
inconshreveable has joined #ipfs
<whyrusleeping>
alexandria-devon: not a problem at all
<whyrusleeping>
i have the bug fixed locally, hardest part is writing more tests
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<zorun>
Blame: hmm? I fail to see how the language impacts the usefulness of their callback system
<Blame>
zorun: adding a callback wrapper to blocking code would be trivial in either language.
<Blame>
so, given a library that blocked, I could easily wrap the code in a thread that waited for the response and called the callback
<zorun>
ah, sorry, we're not on the same frequency
<whyrusleeping>
Blame is on 25.3Mhz i beleive
<zorun>
what I call "callback" is the fact that you can tell the DHT « Please keep me informed whenever the value at this key changes »
<zorun>
whyrusleeping: :)
<whyrusleeping>
zorun: i call those notifications
<wking>
whyrusleeping: Are you ok handling the "no Namesys on IpfsNode" issue, or did you want me to pick that up?
<zorun>
right, naming is hard
<whyrusleeping>
wking: ive got it, i'll have you CR in a few
<wking>
ok, thanks
<whyrusleeping>
name resolution is complicated...
<wking>
;)
<Blame>
aha. yeah with the current design of UrDHT you would have to a a dedicated "poller" function equipped with a mapping of "things to check"->"functions to call"
<zorun>
ok, actually, the authors call it a "listen" operation... I should use that :)
<wking>
is there going to be a daemonless sharness test in your PR?
<zorun>
Blame: "A listen operation, similar to get, that informs the requesting node of changes of values at a provided key for a few minutes. This avoid the need to pool for changes every seconds." from https://github.com/savoirfairelinux/opendht/wiki
<Blame>
yeah I saw that. I'm planning a polled-Queue primitive that should do most of those jobs
<ipfsbot>
[go-ipfs] whyrusleeping created fix/resolution (+1 new commit): http://git.io/vTKKm
<ipfsbot>
go-ipfs/fix/resolution 1babd9d Jeromy: fix offline full path resolution bug
<Blame>
"listen" looks like it requires more state keeping then I want for the first version of UrDHT
mildred has quit [Ping timeout: 255 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<wking>
whyrusleeping: GitHub forgets about comments made in the "files" view after rebase, but I think it might preserve comments made to old commits.
<whyrusleeping>
github preserves comments made in the files tab
<wking>
I'm happy to switch to email-based patch submission and review if folks get tired of GitHub forgetting their stuff ;)
<wking>
whyrusleeping: I don't think it does. Let me drag up an example
<whyrusleeping>
thats why i comment there, because it keeps them if you rebase
<whyrusleeping>
it drops them if you delete the line that the comment is on
<whyrusleeping>
but otherwise, they stay
<wking>
whyrusleeping: oh ok. I just hit the "if you delete that line" case frequently ;). I expect commits made on old comments to survive in that case
<wking>
*comments made on old commits
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]