<ipfsbot>
go-ipfs/fix/bootstrapping 52da1e3 Jeromy: fix bootstrapping bug and add real test for bootstrapping...
joshbuddy has quit [Quit: joshbuddy]
<whyrusleeping>
jbenet: one thing i noticed is that we're not really using the whole 'port zero' thing for sharness tests
<whyrusleeping>
we're randomly picking ports
<whyrusleeping>
could be a cause of some of the random sharness failures
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #2210: fix bootstrapping bug and add real test for bootstrapping (master...fix/bootstrapping) https://github.com/ipfs/go-ipfs/pull/2210
jhulten has quit [Ping timeout: 264 seconds]
<jbenet>
whyrusleeping: hm it's possible
<whyrusleeping>
jbenet: check that PR
gaboose has quit [Ping timeout: 255 seconds]
redfish has quit [Quit: leaving]
<slothbag>
I just compiled that latest boostrapping commit.. still cant connect to webui and still not getting any peers
<whyrusleeping>
>.>
<whyrusleeping>
slothbag: i was able to get a reliable repro, and after that commit it works
<whyrusleeping>
are you sure you rebuilt?
<slothbag>
lemme try again
<jbenet>
whyrusleeping doesnt work for me either :)
<whyrusleeping>
you guys are high
* whyrusleeping
checks again
<jbenet>
ok i got bootstrapped now
<jbenet>
it took a whiiiile
<whyrusleeping>
#notMyProblem
<whyrusleeping>
is your connection decent?
<jbenet>
whyrusleeping very
<jbenet>
bootstrappers may not be rebooting or something
<whyrusleeping>
nah
<whyrusleeping>
your node probably picked all 0.3.* nodes for the first round
<whyrusleeping>
waited for a timeout
<whyrusleeping>
then picked another set
<slothbag>
i have an older build (0.3.11) that bootstraps instantly to 70+ peers... latest master sat there with nothing
<whyrusleeping>
maybe
<jbenet>
ok wfm. whyrusleeping see my comments
<whyrusleeping>
yeah, its the selecting 0.3.* nodes
<jbenet>
makes sense
<whyrusleeping>
i just added print statements and saw the behaviour youre describing
<whyrusleeping>
jbenet: why do you want the test_empty thing?
zorglub27 has quit [Quit: zorglub27]
<jbenet>
whyrusleeping to _make damn sure_ mdns is not making things happen in that test.
<whyrusleeping>
hrm...
<whyrusleeping>
mdns doesnt always happen immediately though
<whyrusleeping>
and iptb disables it by default
Akaibu has joined #ipfs
<whyrusleeping>
and then i explicitly disable it on the bootstrapper
<jbenet>
i saw.
joshbuddy has joined #ipfs
<jbenet>
im still not convinced those things may not break etc. this doesn't hurt anything
<jbenet>
it verifies expected truth before testing another condition
<jbenet>
it's an important control.
<whyrusleeping>
test_expect_success "no shooting stars seen looking out the window for ten seconds" '
<whyrusleeping>
i guess my hesitation is that testing something that isnt likely to happen even in an error case seems odd
<jbenet>
there's way to much there that could go wrong, the sharness config setup could break
<jbenet>
the json parsing could break
<jbenet>
mdns could break
<jbenet>
iptb could break
<jbenet>
just confirm there's nothing beforehand or the test you're doing is useless.
<jbenet>
tests recently didnt catch _this major of a bug_ when we both thought that they certainly would, and you dont want to add 3 lines?
<whyrusleeping>
and even if any of that did, that test wouldnt fail
<whyrusleeping>
i'm saying that mdns wouldnt connect those nodes, we would have to put a wait for 20seconds or something
<whyrusleeping>
and then it wouldnt ever fail in CI, because mdns doesnt work on travis
<jbenet>
no, mdns often kicks in immediately for me.
<whyrusleeping>
often
<whyrusleeping>
but not always
<jbenet>
so this test would often be useless.
<whyrusleeping>
i have to wait a few seconds for it to kick in when i'm messing with it
<whyrusleeping>
right
<whyrusleeping>
if we want to test that mdns isnt doing anything, maybe we should have a little mdns listener thing?
<whyrusleeping>
or something
<whyrusleeping>
i just want to be sure we don't get a false sense of security from the test
<jbenet>
which is why we test that there are no peers connected first
<whyrusleeping>
that wasnt where i was going
<whyrusleeping>
ensuring there are no peers connected first GIVES us that false sense of security
<whyrusleeping>
if i turn mdns on, its still going to pass
<jbenet>
not always, i dont even think often. but WITHOUT that check it's even worse.
<jbenet>
the check is not making it worse, it's narrowing the possibilities of a false positive
<jbenet>
if we need to really do more, sure i'm all for it
<jbenet>
but that at the very least should be there.
<Codebird>
I wrote a little perl script that converts between them.
<Codebird>
(mostly - it doesn't work with a few, I can't figure out why)
<ansuz>
they can't include the most recent hash in the page itself, because that would change the hash of the page, yea?
<Codebird>
Correct. But I think the site owner can find it via their control panel.
<ansuz>
ahh
reit has joined #ipfs
<Codebird>
The bottom two contain weird adult artwork. They are at the bottom because I didn't include them at first. then I figured they are excellent examples of image-intensive sites, which is in itsself useful.
<ansuz>
human porn creeps me out a lot less than animal porn
<whyrusleeping>
yeah....
<ansuz>
but anyway
<whyrusleeping>
but this is the internet
<ansuz>
good work
IlanGodik has quit [Quit: Connection closed for inactivity]
<whyrusleeping>
yeah, this is really good
<whyrusleeping>
its nice to have a bunch of links to browse
<whyrusleeping>
keep up the good work!
<Codebird>
I can provide many more.
<Codebird>
But I thought that would be sufficient.
<whyrusleeping>
i'd be interested in as big of a list as you can get :D
<Codebird>
Did I give you my base58/base32 conversion script?
jgraef has quit [Remote host closed the connection]
<redfish>
is it possible to specify external ip for ipfs node for nodes behind a punched NAT? (Like -externalip for bitcoind.)
<Codebird>
What do you want a big list for?
<Codebird>
I have no automated means of generating one - I have to do them all individually.
<whyrusleeping>
redfish: it generally discovers that automatically
<whyrusleeping>
but adding that as an option doesnt seem like a bad idea
<whyrusleeping>
we ask the peers we connect to what ip address they see us at
<whyrusleeping>
and note that down and use it
<redfish>
@whyrusleeping: oh, so there's automatic decentralized ip discovery, cool. It's strange that I see not a single connection to my node on port 4001 and ~30ish outgoing connections. Do nodes exchange peers?
<whyrusleeping>
redfish: yeah, they do
<whyrusleeping>
your router might be restrictive and not allowing incoming dials?
<redfish>
@whyrusleeping: k, I'll leave the node alone, and later check again if anybody connected to it (I already checked reachability via telnet.)
<redfish>
is there a config option for incoming dials?
dignifiedquire has quit [Quit: Connection closed for inactivity]
<jbenet>
Codebird I appreciate the willingness to help. But read our code of conduct. You cannot post links like that (porn or copyrighted content) here. You've been warned.
<jbenet>
I want everyone here to understand that this sort of stuff in our official public community spaces is unacceptable.
Not_ has joined #ipfs
Not_ has quit [Remote host closed the connection]
<Codebird>
Adult artwork removed from list in future versions. I was just being silly. The only copyrighted content on there is from neocities and the xkcd mirror, and one that looks suspicious but I can't retrieve.
Not_ has joined #ipfs
joshbuddy has joined #ipfs
_rht has joined #ipfs
patcon has quit [Ping timeout: 260 seconds]
<Codebird>
It probably wasn't the best idea to compile a list at one in the morning.
Quiark has joined #ipfs
joshbuddy has quit [Quit: joshbuddy]
speewave has joined #ipfs
disgusting_wall has joined #ipfs
speewave has left #ipfs [#ipfs]
* achin
is having more problems getting the ipfs gateways to load my data :(
<achin>
for the most part it loads OK, but sometimes i'll get a context-deadline-exceeded error
redfish has joined #ipfs
Not_ has quit [Ping timeout: 240 seconds]
borgtu has quit [Remote host closed the connection]
<ipfsbot>
[go-ipfs] whyrusleeping closed pull request #2210: fix bootstrapping bug and add real test for bootstrapping (master...fix/bootstrapping) https://github.com/ipfs/go-ipfs/pull/2210
<ipfsbot>
go-ipfs/add-arguments-to-pin-ls 19e0927 Christian Couder: pin/pin: replace isPinned() with isPinnedWithType()...
<ipfsbot>
go-ipfs/add-arguments-to-pin-ls 9039ff9 Christian Couder: pin: refactor 'pin ls'...
<ipfsbot>
go-ipfs/add-arguments-to-pin-ls e425bb6 Christian Couder: pin: make 'ipfs pin ls' accept path args...
<Quiark>
random_hash, I think the design is pretty solid but client is still pretty alpha...
simpbrain has joined #ipfs
simpbrain has quit [Client Quit]
random_hash has quit [Quit: Page closed]
Guest25__ has quit [Read error: No route to host]
se3000 has quit [Read error: No route to host]
simpbrain has joined #ipfs
computerfreak has quit [Remote host closed the connection]
pskosinski has quit [Read error: Connection reset by peer]
zorglub27 has joined #ipfs
IlanGodik has joined #ipfs
ljhms is now known as ettjattelangtnam
ettjattelangtnam is now known as ljhms
pskosinski has joined #ipfs
Tv` has quit [Quit: Connection closed for inactivity]
yellowsir has quit [Quit: Leaving.]
<Codebird>
I've studied the architecture a bit and agree with Quiark, with the notable exception of IPNS, which really needs redoing.
yellowsir has joined #ipfs
rendar has joined #ipfs
<yellowsir>
i can't connect to my own peer using `ipfs swarm connect` ... anyone who can help me debug this?
dignifiedquire has joined #ipfs
<lidel>
yellowsir, are you using the latest build? there were some fixes related to bootstrap yesterday
<yellowsir>
ok, i will update
pevaso has quit [Ping timeout: 240 seconds]
<yellowsir>
nice, it worked
* Codebird
extends the list, even if not clear why.
<M-davidar>
I think ipfs.pics intends adding video support in the future fwiw
chriscool has quit [Ping timeout: 265 seconds]
M-oddvar has quit [Quit: node-irc says goodbye]
screensaver has joined #ipfs
corvinux has joined #ipfs
mildred has joined #ipfs
corvinux has quit [Remote host closed the connection]
corvinux has joined #ipfs
corvinux has quit [Quit: IRC for Sailfish 0.9]
<lidel>
support for .gifv (.webm / .mp4 loops) would be nice
mildred has quit [Ping timeout: 240 seconds]
mildred has joined #ipfs
<lidel>
hm.. something is still fishy with 0.4.0-dev under docker, I try to ipfs cat a resource that is at public gateway and got a lot of traffic (~300KB/s incoming, 50KB/s outgoing) but no progress (http://hastebin.com/ronehanute.sm)
zorglub27 has quit [Ping timeout: 240 seconds]
rombou has joined #ipfs
rombou has left #ipfs [#ipfs]
patcon has quit [Ping timeout: 272 seconds]
mildred has quit [Ping timeout: 260 seconds]
zorglub27 has joined #ipfs
m0ns00n has joined #ipfs
rombou has joined #ipfs
ljhms is now known as ljhmsljhms
ljhmsljhms is now known as ljhms
Guest has joined #ipfs
rombou has quit [Ping timeout: 260 seconds]
_rht has quit [Quit: Connection closed for inactivity]
Guest has quit [Ping timeout: 264 seconds]
dignifiedquire has quit [Quit: Connection closed for inactivity]
<lidel>
mapping release to latest was nice touch tho ;-)
spm_draget has joined #ipfs
<spm_draget>
Greetings. I am involved in i2p (invisible internet project) and recently found out about ipfs. Curious to find out what ideas could be exchanged.
rombou has joined #ipfs
voxelot has joined #ipfs
<jbenet>
lidel: the docker tag name is on the right.
<jbenet>
lidel: yes. latest was already in use, and should be "release"
<jbenet>
spm_draget i'm sure! and also interface. there was some work (by cryptix i believe) in making an ipfs transport over i2p
<lidel>
jbenet, ok, then it looks fine.. weird
<lidel>
sadly /api/v0/version does not return commit hash : {"Version":"0.4.0-dev","Commit":"","Repo":"3"}
<jbenet>
lidel: i believe it has to be a proper built release for that.
<jbenet>
lidel: ask whyrusleeping
<lidel>
it would be nice to confirm the image was built from the last hash of release branch etc
<jbenet>
indeed.
<jbenet>
lidel: that is most certainly NOT from the release branch. release is at 0.3.11
<lidel>
yeah..
simonv3 has joined #ipfs
<Kubuxu>
If versions is semver it could have been included in version metadata.
<Kubuxu>
like 0.4.0-dev+gf32a6672
<Kubuxu>
or 0.4.0-dev+git.32a6672
chriscool has quit [Quit: Leaving.]
chriscool has joined #ipfs
_rht has joined #ipfs
pierrebeaucamp has joined #ipfs
pierrebeaucamp has quit [Remote host closed the connection]
<jbenet>
lidel: this is a problem with docker hub. im not sure what's up.
<jbenet>
lidel: i'll trigger a rebuilt.
<jbenet>
rebuild*
chriscool has quit [Quit: Leaving.]
ppham has joined #ipfs
libman has joined #ipfs
Matoro_ has quit [Ping timeout: 256 seconds]
<cow_2001>
where's the function that calculates the /ipfs/$hash?
<Kubuxu>
cow_2001: It is complicated.
<cow_2001>
oy vey
<cow_2001>
oh here's a handy specs.md file
<Kubuxu>
you can use: ipfs add -n [file] to calculate it and add file.
<jbenet>
but also involves file importing into ipfs objects
<Kubuxu>
plain hash is sha2-256 but it isn't raw data.
<lidel>
jbenet, i pulled :release (there were changes indeed) and recreated container, still {"Version":"0.4.0-dev","Commit":"","Repo":"3"}
<lidel>
I start to wonder if it may be something on my end
<jbenet>
lidel: very confusing :(
<jbenet>
lidel: im not sure.
<lidel>
could anyone do `docker pull jbenet/go-ipfs:release` and confirm the Digest is sha256:52b8228d0d918c05ed2b121343218c9810c159a79b7c345f0afcf111cb46f985
<lidel>
?
<jbenet>
cc Luzifer might know something
<achin>
this reminds me to put "fix docker on my machine" to my TODO list :/
Matoro_ has joined #ipfs
ppham has quit [Remote host closed the connection]
reit has quit [Ping timeout: 272 seconds]
chriscool has joined #ipfs
ugjka has quit [Quit: Leaving]
ugjka has joined #ipfs
silotis has quit [Remote host closed the connection]
silotis has joined #ipfs
<whyrusleeping>
Blame: ping
<whyrusleeping>
You around?
neurosis12 has joined #ipfs
neurosis12 has joined #ipfs
ploopkazoo has quit [Quit: ZNC - 1.6.0 - http://znc.in]
Matoro_ has quit [Ping timeout: 256 seconds]
<Blame>
whyrusleeping: pong. Im Mobile. Eta at keyboard 1 hr
ploopkazoo has joined #ipfs
neurosis12 has quit [Remote host closed the connection]
<jbenet>
Luzifer: any idea what i'm doing wrong? { dh-latest: gh-release, dh-release: gh-release, dh-master: gh-master }, and yet dh-release seems to be pulling gh-master
<jbenet>
OH
<jbenet>
wow.
<lidel>
:-D
<jbenet>
Luzifer: should we just remove that line, then?
<jbenet>
Oh no
<jbenet>
it's used in the installation. hmmmm
<jbenet>
maybe the dockerfile should build with go-- that's painful of course, but it might be the right thing to get Dockerfiles per version, etc.
<Luzifer>
we could build the binary inside the docker container from source, that would fetch the current branch… otherwise… at least install git into the container and try to fetch the current branch from git… but that may break if gobuilder did not build the branch before dockerhub
rombou has joined #ipfs
<jbenet>
maybe we could have two Dockerfiles, one for dev (builds with go) and one for relase (does what this does, but pulls `release`
<jbenet>
the pain i think was not having to compile yourself.
<jbenet>
golang requires >512MB mem to compile
<Luzifer>
building inside the dockerfile isn't painful and also it doesn't put that much overhead to the size of the dockerfule
<Luzifer>
*file
<Luzifer>
dockerhub does the compile… not the user
<Luzifer>
the user only downloads the results
<Luzifer>
and beside dockerhub having some serious issues with failing builds because of their build system (that type of issue not being there but sending emails to the user) they can manage to build go binaries inside the container
<lidel>
maybe you could ADD .git/HEAD and then set VERSION to the value from the file
<Luzifer>
(size of the docker image of course)
<Luzifer>
lidel: yeah but see my comment above, this may introduce race-conditions and then builds fail because of magic happenings
<jbenet>
Luzifer: ok im down with whatever here
<lidel>
cut -d/ -f3 .git/HEAD → master
<Luzifer>
jbenet: I'll prepare a PR with building in the container?
<lidel>
Luzifer, no need for build, it would be fetched from gobuilder as it is now
<Luzifer>
lidel: if gobuilder already has that tag / branch built
ppham has joined #ipfs
<Luzifer>
hmm :/ we're still using alpine 3.2… my PR switching to 3.3 was not merged yet… :/ that propose will again contain a switch to 3.3 as we need golang 1.5
<lidel>
Luzifer, oh, i thought gobuilder triggers docker hub? you say it is github who triggers builds at both at the same time?
<Luzifer>
dunno whether the github-trigger was deactivated
<jbenet>
Luzifer: thank you yes that'd be great.
<jbenet>
github.com/ipfs/go-ipfs has a "Docker" service hook, but unclear whether it's triggering dockerhub/jbenet/go-ipfs as dockerhub/ipfs/go-ipfs also exists... (we really should be switching to that and off my account. i believe there was password issue but that may have been resolved.
<lidel>
Luzifer, I recently did similar setup for my pet project and there is a bug in Docker when you try to pipe build binary from one container directly into the target one: https://github.com/docker/docker/issues/15785 (the fix is in 1.10)
ppham has quit [Remote host closed the connection]
f[x] has joined #ipfs
<jbenet>
i'll merge when tests finish
<jbenet>
I'm so sick of that damn travis test getting stuck. it makes testing so much more painful. :(
<Luzifer>
hm somehow you need to bring that dockerfile to release too… :/
<jbenet>
whyrusleeping: the ports thought is a good idea-- where did you notice randomization vs /0 ?
ppham has joined #ipfs
<jbenet>
Luzifer: indeed. we could cherry pick it and rebase master... though that's sketchy. maybe cherry pick and hope git applies it fine in the merge later