<alu>
thats someone from australia [opposite corner of the world] in my room!!
<alu>
on my 2d/3d web page!
<jbenet>
alu that's pretty cool
<alu>
Now I have a question, is that person also seeding the room?
<alu>
he doesn't have ipfs
<alu>
so how can he become a peer ?
<jbenet>
alu: are they accessing via gateway.ipfs.io, or what?
<alu>
yes
<jbenet>
alu: then no, for now they're just viewing (once we have node-ipfs they will be able to seed too just from the gateway). we should talk about how to make this work with ipfs now, like people installing ipfs locally.
<jbenet>
alu: i presume they're already installing something?
www1 has quit [Ping timeout: 252 seconds]
<jbenet>
whyrusleeping: 512k window size? does it always allocate those buffers? cause * N conns?
<jbenet>
hey rht___ are the latest goprocess changes in go-ipfs/master now?
border has joined #ipfs
<rht___>
actually not yet because godeps
<jbenet>
rht___ ah fun. godeps is painful.
<jbenet>
whyrusleeping: what needs to happen to gx?
<jbenet>
can we ditch godeps for good? (i'm borderline to using a shell script :) )
<jbenet>
also, rht___ mind if i reassign some other issues? #143 is lower prio, #1418 has kbala and lgierth on it atm (may still need help later, but atm may be more trouble coordinating changes)
<jbenet>
#1353 would be good to have, but needs a bit of work discussing before doing. (we could do something like git's ref logs (which are just text files) or make them proper ipfs objects.
<rht___>
sure, I pick #143 just to settle if /dev/urandom is fine (and why gpg is not using it)
Leer10 has quit [Ping timeout: 255 seconds]
<rht___>
(but golang's openpgp uses /dev/urandom ... stuff like this)
Leer10 has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<rht___>
gx: every update to deps will have to change the hashes of the import statement in the sources?
<jbenet>
rht___ i think we came up with a way to prevent the hashes from changing everywhere (though i personally dont mind that too much)
<rht___>
but shouldn't the versioning (especially that it is not human readable. code is meant to be readable) be localised to the package manager json file?
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
alu: i emailed the janusvr folks. i'll keep you posted if anything good comes up.
<jbenet>
rht___ yeah, maybe. and maybe not. the "package.json" may be too far out.
<rht___>
unless there is auto import tool (e.g. goimports), then I'd have to copy the hash manually
<jbenet>
it may be that a single file is all you need per module (think npm but even more modular)
<jbenet>
rht___ yeah granted, that'd be annoying :)
joshbuddy has quit [Quit: joshbuddy]
<rht___>
> it may be that a single file is all you need per module ...
<rht___>
this is only possible if the package manager has a knowledge that the file doesn't depend on other parts of the module
queue has quit [Ping timeout: 244 seconds]
<Tv`>
jbenet: status update! i burned a bunch of time for Gophercon, needed these two days to just catch up, but have time for ipfs again during the rest of this week; lmk if you have something lined up.
notduncansmith has quit [Read error: Connection reset by peer]
<Tv`>
akshully, not sure if you want to close pulls once they're in dev or master
<tperson>
I love git, it would appear every command you run is track, not only your source tree (git reflog)
<tperson>
jbenet: Decided as part of this readonly gateway I'm going to start writing the api test spec. As I feel there needs to be better test coverage for the API.
<Tv`>
fuse/readonly/ipfs_test.go:39: too many arguments in call to importer.BuildTrickleDagFromReader
<tperson>
When playing around with the ipfs-blob-store I was getting some errors when running against the read-only branch.
<ipfsbot>
[go-ipfs] tv42 force-pushed s3 from f857e0d to ba6686b: http://git.io/vmnzd
<ipfsbot>
go-ipfs/s3 4cf9ead Tommi Virtanen: Remove dead code
<ipfsbot>
go-ipfs/s3 95299bf Tommi Virtanen: core tests: Stop assuming internals of Config
<ipfsbot>
go-ipfs/s3 611ddd9 Tommi Virtanen: Fix BuildTrickleDagFromReader call args
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
www has quit [Ping timeout: 246 seconds]
<ipfsbot>
[go-ipfs] tv42 force-pushed s3 from ba6686b to dda3c6a: http://git.io/vmnzd
<ipfsbot>
go-ipfs/s3 b0dd235 Tommi Virtanen: Remove dead code...
<ipfsbot>
go-ipfs/s3 77f6941 Tommi Virtanen: core tests: Stop assuming internals of Config...
<ipfsbot>
go-ipfs/s3 e233d4f Tommi Virtanen: sharness: Stop assuming leveldb Datastore...
<whyrusleeping>
maybe make the logo look incomplete?
tilgovi has quit [Ping timeout: 240 seconds]
pinbot has quit [Ping timeout: 246 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
jbenet: the tests run much faster now :)
<whyrusleeping>
sharness tests*
<whyrusleeping>
those shutdown changes i made helped it seems
<tperson>
jbenet: whyrusleeping Do you think a API Spec tester should be implemented as a go test, or as a standalone program that can be compiled and downloaded?
<tperson>
I'm cleaning towards a standalone
<tperson>
leaning*
<whyrusleeping>
standalone would be nice if we were planning on implementing the API exactly the same in other languages
<whyrusleeping>
i guess it would also be nice to have something outside the codebase to assert that it doesnt change too
simonv3 has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
tomdanner has quit [Remote host closed the connection]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
pfraze has quit [Remote host closed the connection]
<kyledrake>
alu Right now I'm just describing how to use it with the Docker image. What I want to do eventually is have a guide to use IPFS to share data between machines in a docker cluster using etcd as a shared IPFS hash reference
<alu>
heh
<alu>
funny, you catch me right when im making a neocities page
<alu>
trying to get a pretty / basic front end up
<alu>
ima just use darkengine fuck it
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<alu>
its like neocities for Janus, a giant apartment complex to go with it
<alu>
you get a room and can customize it w/ the editor, design how you want your portal to look, etc
<alu>
pretty neat
<alu>
but yeah, rooms are webpages and links are portals. my room is hosted on neocities by appending the code into index.html.
<jbenet>
hey guys, im trying to fix the docker image. i'll try using alpine linux-- one thing to do would be remove the requirement of go compilation and just use one of our binaries from go builder.
<jbenet>
i guess we'd have to install both 32 and 64 archs?
<jbenet>
lgierth Luzifer: you two are probably asleep, but thoughts here would be useful o/
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
simonv3 has quit [Quit: Connection closed for inactivity]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<Luzifer>
jbenet: nah. why both archs? shouldn't be required. I'd suggest to build `FROM scratch`
<Luzifer>
hmh will not work because docker hub cannot download things before building
<Luzifer>
okay alpine sounds good
<Luzifer>
if you want to I can build a docker file for alpine
karlwestin has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<Luzifer>
down to 28.89MB \o/
<ipfsbot>
[go-ipfs] Luzifer opened pull request #1481: WIP: Migrated Dockerfile to use alpine linux (master...alpine) http://git.io/vmczu
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
joshbuddy has joined #ipfs
tperson has quit [Quit: Connection closed for inactivity]
karlwestin has quit [Quit: karlwestin]
reit has joined #ipfs
karlwestin has joined #ipfs
karlwestin has quit [Client Quit]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<jbenet>
mafintosh: does playback work with chromecasts anymore?
<mafintosh>
jbenet: yea
<mafintosh>
jbenet: i just used it to broadcast stuff
<jbenet>
mafintosh it's not working so hot atm. new chromecast.
<mafintosh>
jbenet: oh maybe they changed something then
<jbenet>
mafintosh: any idea how to debug the cast?
<mafintosh>
jbenet: if not then its probably something with discoveribility that has changed
<jbenet>
and of course the moment you get involved it just works
<mafintosh>
lol
<jbenet>
(well maybe, it's attempting to work)
<alu>
man if i dont find a place to live after DEFCON
<alu>
i might just move to thailand
<alu>
also im testing the docker image myself
<alu>
would like to know if theres been any updates
keroberos has quit [Max SendQ exceeded]
keroberos has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
joshbuddy has quit [Quit: joshbuddy]
<jbenet>
Luzifer: that's great, thank you!
atomotic has joined #ipfs
<jbenet>
Luzifer: "Check checksum for downloaded asset" would be sweet. I wonder how best to do this-- have a pre-commit hook that builds IPFS and updates the hash? or check it from elsewhere? (i guess from gobuilder?)
<jbenet>
Luzifer: know if go is reproducible?
Tv` has quit [Quit: Connection closed for inactivity]
zorun_ has quit [Ping timeout: 256 seconds]
<mafintosh>
jbenet, daviddias how is the node spdy transport module working out for you guys? is it stable enough to use for "production" things?
notduncansmith has joined #ipfs
<jbenet>
daviddias went to sleep a bit ago. i think there's a bug daviddias found ran into earlier today, but take a look. you can probably fix them.
<mafintosh>
jbenet: cool thanks
notduncansmith has quit [Ping timeout: 265 seconds]
zorun has joined #ipfs
www has joined #ipfs
kbala has quit [Quit: Connection closed for inactivity]
karlwestin has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
reit has quit [Ping timeout: 240 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
rht___ has quit [Quit: Connection closed for inactivity]
konubinix has joined #ipfs
<cryptix>
gmorning
<Luzifer>
jbenet: it isn't I'd check against hash table of gobuilder.
karlwestin has joined #ipfs
<Luzifer>
jbenet: but what we could do: Sign all tags… (So don't use lightweight tags but signed ones) in that way gobuilder would verify the code before building a tag…
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]
reit has joined #ipfs
martinBrown has quit [Quit: -]
martinBrown 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]
<Luzifer>
jbenet: Git-Cop is trolling me. I've used its forced message and it continues to flame me about it. >:(
<lgierth>
nevermind the make step, they just download golang
therealplato has joined #ipfs
<Luzifer>
jep. I really really really don't like their approach by downloading some random golang
<lgierth>
i just really like to be able to build images from local changes or other branches
notduncansmith has joined #ipfs
<lgierth>
what about a Dockerfile.dev or something like that?
notduncansmith has quit [Read error: Connection reset by peer]
<lgierth>
yeah but you download a random ipfs ;)
<Luzifer>
for other projects I use a jenkins job fetching the code, executing the `centurylink/golang-builder:latest` providing me a binary and then bunding them into a `FROM scratch` container… but thats not possible on dockerhub
<Luzifer>
lgierth: random? *throws code* they are not random!!!!!!!!111eleven! they are build by my build system!!!! :P
<lgierth>
:>
<lgierth>
all your entropy are belong to us
tomdanner has joined #ipfs
* Luzifer
throws entropy at lgierth
<lgierth>
i might just be missing something -- i don't understand why it's bad to build within `docker build`
<lgierth>
me having next to no experience with golang
<Luzifer>
the container is several hundred MB bigger
<lgierth>
rm -rf
<Luzifer>
but tbh I don't have an opinion on this… jbenet wanted to have a container without building inside, I built the container… ;)
<Luzifer>
my servers are having all that layers for golang:latest nevertheless
<lgierth>
ok let's wait and see what his reason was
<lgierth>
i gotta go to ikea again
<Luzifer>
wow. at this time?
<Luzifer>
just must be crazy ;)
<Luzifer>
(at least the IKEA at schnelsen is hell that late)
<lgierth>
we have 4 ikeas in this city
<lgierth>
:)
<Luzifer>
hm we only have 2 ;)
<lgierth>
it's fine here if you know exactly what you want
tomdanner has left #ipfs [#ipfs]
<Luzifer>
yeah but all those people are wandering around like really slow zombies… If they were not there I'd need 5-10min to get all the things I want…
<whyrusleeping>
i need to add two separate 'blockstores' now
<Tv`>
huh
<whyrusleeping>
one to store blocks we're intending to share, and one for things that are private, such as private keys and pin lists
<Tv`>
you might want to look at the s3 branch first
<whyrusleeping>
since jbenet wants to store everything in the blockstore
* whyrusleeping
does a git fetch
<Tv`>
i'd hate to have tons of merge conflicts
<whyrusleeping>
lol
<whyrusleeping>
what doesnt kill you only makes you more pissed off at someone else
<Tv`>
ideally, let's bring that in first, clean up, then you get to have two sections in config instead of just Datastore
<Tv`>
or something
<whyrusleeping>
:+1:
<Tv`>
because i fear you wouldn't want those "private" things on s3, necessarily, etc
border has quit [Ping timeout: 246 seconds]
<Tv`>
or you might, but with a separate bucket or acl
<whyrusleeping>
yeah
<whyrusleeping>
i'd want them on disk still
<whyrusleeping>
but thats just me
<Tv`>
then you really want two sections in config
domanic has quit [Ping timeout: 246 seconds]
* whyrusleeping
is excited for Tv`s s3 changeset
domanic has joined #ipfs
karlwestin has quit [Quit: karlwestin]
<Luzifer>
whyrusleeping: nice quote! I think I'll keep that one :D "17:42:51 <@whyrusleeping> what doesnt kill you only makes you more pissed off at someone else"
karlwestin has joined #ipfs
<whyrusleeping>
Luzifer: :D
karlwestin has quit [Quit: karlwestin]
<Tv`>
whyrusleeping: i was actually amazed to see how little i needed to do to get it working
<Tv`>
i mostly did the config stuff as a cleanup, even that wasn't strictly necessary
atomotic has joined #ipfs
<whyrusleeping>
Tv`: thats good news
Taek has quit [*.net *.split]
hellertime has quit [*.net *.split]
ralphthe1inja has quit [*.net *.split]
mappum has quit [*.net *.split]
uhhyeahbret has quit [*.net *.split]
jonl has quit [*.net *.split]
krl has quit [*.net *.split]
Luzifer has quit [*.net *.split]
ralphtheninja has joined #ipfs
krl has joined #ipfs
hellertime has joined #ipfs
uhhyeahbret has joined #ipfs
Taek has joined #ipfs
jonl has joined #ipfs
mappum has joined #ipfs
Luzifer has joined #ipfs
atomotic has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<jbenet>
lgierth Luzifer: the thing about building is the go compiler is big. We could remove it afterwards, but then every user has to do that. I almost think there should be two dockerfiles, one for building, one for downloading latest _release_.
nessence has joined #ipfs
tilgovi has joined #ipfs
ryan___ has quit [Remote host closed the connection]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<lgierth>
jbenet: i don't understand the issue with removing the go compiler. it's an rm -rf away isn't it?
joshbuddy has joined #ipfs
<lgierth>
and we don't neccessarily build FROM golang
<lgierth>
anyhow i'm fine dockerfile.dev and dockerfile.release
<lgierth>
i'm thinking which one i'd use for the gateways though, and that'd be one which lets me build from an arbitrary commit
<lgierth>
nevermind the last line, tagged release images are fine for the moment
<jbenet>
Yeah we can rm rf. But many people will not want to wait through compiling (or may not be able to-- golang compile is a mem hog) so I think having both is probably the right way
<jbenet>
whyrusleeping: we don't _need_ two blockstores. That's just one way of inplementing that. If color works, that should be enough for an implementation. (Can do either way as blockstore may be faster)
<jbenet>
(But it is more complicated because now we have to worry about two blockstores at a higher level)
mildred has joined #ipfs
<Tv`>
jbenet: do you want to configure where what kinds of blocks are stored?
<Tv`>
if you do, you either need two Datastore config sections, or more complex config that can talk about mounts
<Tv`>
(and for both, the cleanups in the s3 branch would be a good thing)
mildred has quit [Read error: Connection reset by peer]
mildred has joined #ipfs
<lgierth>
jbenet: Luzifer: anyhow this whole discussion is not a blocker regarding the blog posting :)
ryan___ has joined #ipfs
<jbenet>
Tv`: it's just for deciding which blocks are served and which aren't. The more complicated question is: to whom.
patcon has quit [Ping timeout: 244 seconds]
<jbenet>
lgierth: maybe we can make the dockerfile in repo build from src (using ADD) and move the release dockerfile to its own repo?
<Tv`>
jbenet: I understood there were privacy implications in leaking existence of keys for one of those colors
<Tv`>
I figured one way to use s3 would be to just let it be public
<Tv`>
Hence, wrong color objects there would be bad
tilgovi has quit [Ping timeout: 250 seconds]
notduncansmith has joined #ipfs
simonv3 has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
Encrypt_ has joined #ipfs
<whyrusleeping>
jbenet: what do you mean 'if color works' ?
<whyrusleeping>
jbenet: i really think that two blockstores is the only reasonable way to do it.
mildred has quit [Ping timeout: 248 seconds]
mildred has joined #ipfs
Encrypt_ has quit [Quit: Quitte]
<whyrusleeping>
go-peerstream as a gx package: QmUgvTbsHH1h6BJ9t7obEpomqQ2f1iizDNHFLNJhYzJqD1 (assuming my internet holds out well enough for an ipfs transfer)
<jbenet>
I mean if we trust it then why not trust it to handle the coloring of what to serve
<whyrusleeping>
because its inneficient as all hell to enumerate that entire set constantly
<whyrusleeping>
or keep it in memory
<whyrusleeping>
wont scale at all
<jbenet>
Tv`: I would want to store my private data on S3 (possibly encrypted)
<jbenet>
whyrusleeping: that's orthogonal, you asked me why I said "if it works". Also see above where I say that it's likely faster to do this.
<jbenet>
(This being two blockstores)
nsh has quit [Ping timeout: 246 seconds]
<jbenet>
"Keep it in memory"? Why is that a concern?
<whyrusleeping>
ah, i thought you were being skeptical of the fact that it worked
nsh has joined #ipfs
<Tv`>
jbenet: sure, private data on s3 might still be a separate bucket, a different acl setting, etc
<Tv`>
(the current s3 branch just shoves everything under one "config" = bucket,acl,etc, but one could write a datastore config widget for saying "handle /blocks like this")
nessence has quit [Remote host closed the connection]
notduncansmith has quit [Read error: Connection reset by peer]
<daviddias>
sprintchicken: working on integrating the new spdy-transport on swarm, trying to understand why it is sending a RST_STREAM frame, writing more tests and refactoring during that process
tilgovi has quit [Quit: No Ping reply in 180 seconds.]
tilgovi has joined #ipfs
<lgierth>
anybody able to ping 128.199.219.111?
<lgierth>
that's the gateway in singapore
Encrypt_ is now known as Encrypt
<lgierth>
and i just lost it a few minutes ago. digitalocean ui says it's alive
<jbenet>
lgierth: nope cant ping
<jbenet>
lgierth: may be hosed-- i've seen this happen and have to reboot via DO
<lgierth>
i'll give that a try
<lgierth>
ah hey digitalocean's "live" stats stop 15 minutes ago too
<jbenet>
lol
<lgierth>
yeah that power cycle seems to hang
* lgierth
files support ticket
<lgierth>
The hypervisor in this case is under a very high load. Our platform engineers are working to reduce the load as a first step before doing a reboot of the machine itself, as a reboot is usually the last step.
<lgierth>
-- John Doe, Customer Success Engineer
domanic has quit [Ping timeout: 256 seconds]
<whyrusleeping>
sprint checkin: sketched out ideas more concretely for live streaming/pubsub, finished dev0.4.0 rebase, working on gx a bit now
therealplato has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<whyrusleeping>
jbenet: go-peerstream as a gx package: QmPfGxidTpHSe4Pq6PHwrPAbqHkRZfJbUsFquA9mym6eVw
<whyrusleeping>
(all dependencies included as well)
<lgierth>
saturn still unresponsive, another update said it looked like a slightly bigger issue in SGP1
<whyrusleeping>
lgierth: lame...
<lgierth>
let me take it out of dns
tilgovi has quit [Ping timeout: 240 seconds]
joshbuddy has quit [Quit: joshbuddy]
tilgovi has joined #ipfs
hellertime has quit [Ping timeout: 246 seconds]
domanic has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<lgierth>
back it is
<lgierth>
since 20:12
<lgierth>
gonna reinstate dns
simonv3 has quit [Quit: Connection closed for inactivity]
<lgierth>
i disabled it by setting the [h.]gateway.ipfs.io record from "@" to an arbitrary "disabled"
tilgovi has quit [Ping timeout: 244 seconds]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<cSmith>
lgierth: is it in nyc3?
<cSmith>
I've had that issue too, box was really sluggish and eventually just died. moved it over to a different DC and had no issues
<lgierth>
cSmith: nope sgp1
<lgierth>
had a support engineer confirm issues with the hypervisor
<cSmith>
oh, interesting
* cSmith
is moving to lunanode soon anyways
<lgierth>
no idea how often similar issues happened in the past, our monitoring has been up for a week only
<jbenet>
kbala: ok i found it-- the synchronization is off, i'll comment
<lgierth>
Bat`O: i'm a go newb too
<lgierth>
sublime and a terminal
<lgierth>
with gosublime
<kbala>
jbenet thanks
<Bat`O>
lgierth: i'll try that thx
<jbenet>
kbala: was this from whyrusleeping's code
<jbenet>
?
<kbala>
jbenet: no this is mine, only the transport function is from whyrusleeping
<jbenet>
kbala: did you add the `done` channel? we can go through the sync together
notduncansmith has joined #ipfs
<jbenet>
(or the `toDeliver` channel)
notduncansmith has quit [Read error: Connection reset by peer]
<kbala>
jbenet yeah i did
www has quit [Ping timeout: 264 seconds]
Leer10 has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
www has joined #ipfs
pfraze has quit [Remote host closed the connection]
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
<cryptix>
Bat`O: /debug/vars and /debug/pprof are stdlib go. you can scrape and analyse them with 'go tool pprof' see https://blog.golang.org/profiling-go-programs for the gory details
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]
therealplato has quit [Ping timeout: 244 seconds]
therealplato has joined #ipfs
notduncansmith has joined #ipfs
notduncansmith has quit [Read error: Connection reset by peer]