<lgierth>
bret: they aren't that bad, i shouldn't have said "really bad", just don't be disappointed with their performance when compared to x86 :) i'm not sure the RNG will make a difference if you use urandom instead of random
<daviddias>
whyrusleeping: how to add files to a mfs folder and or create the files there automatically?
<daviddias>
can see the dir making commands
<daviddias>
but "ipfs files cp <source> <dest> - copy files into mfs" doesn't accept the source to be out of mfs
<whyrusleeping>
ipfs files cp /ipfs/QmBlah <dest>
<whyrusleeping>
daviddias: o/
<daviddias>
ah! got it was expecting that if it was a local file, for it to know how to add
<daviddias>
would that make sense?
<whyrusleeping>
hrm.... maybe
<whyrusleeping>
how would it differentiate a local path from another path within mfs?
<daviddias>
seems that mfs always requires for a leading slash
<whyrusleeping>
making the assumption that not having a leading slash means 'add from your local filesystem' feels wrong to me
<daviddias>
Will just update the err message for now
<whyrusleeping>
daviddias: sounds good
<whyrusleeping>
also, making that do an implicit add would make that command bizarrely complicated
<davidar>
.seen jbenet
<multivac>
davidar: I last saw jbenet at 2015-10-09 - 21:48:40 in here, saying Thanks multi ac
<ipfsbot>
[go-ipfs] diasdavid opened pull request #1826: Update help text to avoid confusion of where files can be added from (dev0.4.0...patch-2) http://git.io/vCnI2
<daviddias>
whyrusleeping: that is also a good reason :)
<davidar>
ipfspics-Vincent (IRC): o/
<bret>
lgierth: do you recommend any small x86/x64 boards? Do the new arm chips improve that performance at all?
<davidar>
bret (IRC): Intel has that Edison thing
<lgierth>
bret: i'll be able to recommend (or not) the intel nuc in the next few weeks
<whyrusleeping>
ipfspics-Vincent: i'm not sure how to go about 'third party' bootstrap nodes. Right now we manage all the bootstrap nodes under our infrastructure repo
<lgierth>
pretty expensive obviously, compared to little arm boards
<davidar>
.w Intel Edison
<multivac>
[WIKIPEDIA] Intel Edison | "The Intel Edison is a tiny computer offered by Intel as a development system for wearable devices. The system was initially announced to be the same size and shape as an SD card and contain a dual-core Intel Quark x86 CPU at 400 MHz communicating via Bluetooth and Wi-Fi. A later announcement changed..." | https://en.wikipedia.org/wiki/Intel_Edison
amstocker has joined #ipfs
<lgierth>
would be interesting to test one of these little edisons out
<bret>
I've hear nuc's have a few issues. Mostly relating to display compatibility issues. Not a huge deal when you run headless
atgnag_ is now known as atgnag
<lgierth>
we have one cjdns'er having their pentium 2 run circles around a raspi1
<lgierth>
bret: yeah i heard that too, haven't had an issue so far
<bret>
A coworker got an Edison to use like an arduini. Not sure if he ever got it up and running. Do those run a full Linux system?
<ipfspics-Vincent>
whyrusleeping: Are the bootstrap servers currently in the same datacenter?
<lgierth>
ipfspics-Vincent: nope -- spread all over digitalocean
f[x] has quit [Ping timeout: 255 seconds]
<whyrusleeping>
ipfspics-Vincent: lgierth is the infra guy and is really the best one to talk to about inf
<whyrusleeping>
we could have a listing of 'community bootstrap nodes'
<bret>
Just as a repo?
<lgierth>
whyrusleeping: on that note, did you check the telemetry yet? :P ssh -L 8080:metrics.i.ipfs.io:80 root@<any-solarnet-host>.i.ipfs.io
<ipfspics-Vincent>
It might not help that much though, we are also on DO
<davidar>
ipfspics-Vincent (IRC): a few people have asked if ipfs.pics is/will be open source?
<ipfspics-Vincent>
davidar: We want it to be. We will rebase the image server, with our caching and basic image analysis, on docker and release it
<ipfspics-Vincent>
Maybe under AGPL
<davidar>
ipfspics-Vincent (IRC): cool :)
patcon has quit [Ping timeout: 240 seconds]
<ipfspics-Vincent>
davidar: It's just that agpl is not so simple to implement. We are just not sure how we can inform the user about the libre source code on an image server, so there's that
jimbeam has quit [Ping timeout: 246 seconds]
ygrek has joined #ipfs
<ipfspics-Vincent>
whyrusleeping lgierth: could also be interesting to have people in independant jurisdictions controlling the bootstrap nodes
TheWhisper has quit [Quit: Leaving]
TheWhisper has joined #ipfs
TheWhisper has left #ipfs [#ipfs]
TheWhisper has joined #ipfs
TheWhisper has quit [Client Quit]
TheWhisper has joined #ipfs
amstocker has quit [Ping timeout: 265 seconds]
jimbeam has joined #ipfs
Not_ has joined #ipfs
mercora has joined #ipfs
<mercora>
Hey again ;) i now run the daemon on a pi and it works as expected with the api option... but sadly it panics after a while rather consistently at some seamingly random point
<lgierth>
on the pi?
<mercora>
is there an known issue or something i could do to help mitigate this?
<davidar>
ipfspics-Vincent (IRC): presumably a link to the code on the homepage would be sufficient?
<mercora>
lgierth: yes thats seems likely, thanks again :)
ygrek has quit [Ping timeout: 260 seconds]
<ipfspics-Vincent>
davidar: I'm not sure what to do if someone links directly to an image. Technically I would not follow the license in that case
amstocker has joined #ipfs
<ipfspics-Vincent>
davidar: 5d d) If the work has interactive user interfaces, each must display Appropriate Legal Notices; however, if the Program has interactive interfaces that do not display Appropriate Legal Notices, your work need not make them do so.
<davidar>
ipfspics-Vincent (IRC): I think what that clause means is: if someone releases agpl code that includes some way of displaying legal information, you're not allowed to take it out
<davidar>
I don't think it means that you have to display such notices everywhere
<davidar>
Nobody would expect legal notices within an image file
<davidar>
ipfspics-Vincent (IRC): In any case, even if you release the code under the agpl, you as the author still retain copyright, so you can do whatever you want :)
<ipfspics-Vincent>
I think we would have to redirect people first to a page where they could acknowledge their rights before they could see it
<davidar>
ipfspics-Vincent (IRC): you could if you wanted to, but I'm pretty sure the agpl doesn't *require* you to
<ipfspics-Vincent>
davidar: but we would like outside contribution also : )
<davidar>
Although, if you did, I think everyone who used your code would also have to do the redirect
<deltab>
what about people who post the images: what do they interact with?
<davidar>
The gpl doesn't really say much about what you *have* to do with your code, it just says what other people *can't* do
<davidar>
If you don't want to put legal notices in, that's fine
<davidar>
But if you do, anyone forking the code has to honor that
<davidar>
*interactive legal notices
<davidar>
"if the Program has interactive interfaces that do not display Appropriate Legal Notices, your work need not make them do so."
<ipfspics-Vincent>
standard gpl doesn't need that, just agpl
<davidar>
s/gpl/agpl
<multivac>
davidar meant to say: The agpl doesn't really say much about what you *have* to do with your code, it just says what other people *can't* do
<davidar>
ipfspics-Vincent (IRC): if you're concerned, look at some other agpl projects to see what they do
Leer10 has joined #ipfs
<deltab>
or ask the FSF, or opensource.stackexchange.com
<sonatagreen>
when recursively adding a directory with lots of files, most of which haven't changed since the last time i added, will it run noticeably faster with the -q flag?
rozap has quit [Ping timeout: 244 seconds]
<deltab>
time it and see
<deltab>
try it each way a few times, to even out the effects of caching
<sonatagreen>
It seems to be about the same either way.
crossdiver has joined #ipfs
kevin`` has joined #ipfs
jbenet has joined #ipfs
<sonatagreen>
here's the bit of shellscript i just wrote to update my site:
<sonatagreen>
ipfs name publish `cat ../ipfs-site-hash`
<sonatagreen>
this seems suboptimally complicated
TheWhisper has quit [Quit: Leaving]
TheWhisper has joined #ipfs
amstocker has quit [Ping timeout: 265 seconds]
Guest73396 has joined #ipfs
tenodynia has quit [Remote host closed the connection]
<whyrusleeping>
sonatagreen: any particular reason youre writing it out to a file?
<whyrusleeping>
'ipfs name publish `ipfs add -rq _site | tail -n1`' should work just fine
<whyrusleeping>
and you can run 'ipfs name resolve' to get the hash back
<sonatagreen>
well, 'ipfs name resolve' seemed slow, so i wanted to cache it
Vilhelm has joined #ipfs
<whyrusleeping>
oooh, okay
<whyrusleeping>
oh, and unpinning the old site.. hmm
<sonatagreen>
well, i /could/ use a bash variable to store the old hash
<whyrusleeping>
what we really need is a way to pin /ipns/ hashes
<achin>
this is a pretty core use-case that ipfs should eventually support (unpinning old hashes)
<sonatagreen>
yeah, ideally i should be able to do all that in one line
<whyrusleeping>
yeah, we can unpin old hashes just fine, but the doing it automatically is the hard part
<sonatagreen>
(except the jekyll)
<sonatagreen>
maybe something like 'ipfs name add -r foo/'?
<achin>
i don't think it needs to be automatic, but i think storing hashes in a text file is the first thing to fix
<whyrusleeping>
achin: yeah, having an add log is on the todo list
<whyrusleeping>
the hardest part is getting everyone to agree on what it should look like
<achin>
sure, it's not clear exactly how it should look and work
<sonatagreen>
or maybe 'ipfs name publish [hash] --unpin-previous-hash'
<achin>
i think tags might be a solution for this
<sonatagreen>
but i like the idea of pinning an ipns. ('whatever hash this currently points to, keep that pinned.')
<achin>
where you can tell your local node a human-meaningful string that should be associated with a particular hash
<achin>
HASH=$(ipfs add thing); ipfs pin rm $(ipfs tag get my-site); ipfs tag set my-site $HAHS
ygrek has joined #ipfs
<kpcyrd>
are the files cached when requested through a gateway? if yes, how long? is it seeded by the gateway during that time?
Guest73396 has quit [Ping timeout: 240 seconds]
<achin>
any node that has a block of data will make it available to everyone
<kpcyrd>
same question if I just do `echo ohai | ipfs add`
<achin>
in IPFS the closest thing to "seeding" is when you pin something
<achin>
the gateways won't pin stuff that is requested through it (it'll just be cached, i'm not sure for how long)
<achin>
(did that answer your question?)
<kpcyrd>
maybe. I'm preparing some experiments and I'm trying to avoid common issues. I'm trying to understand when ipfs decides to store something and when it's throwing something away.
<achin>
there are a bunch of commands that will cause ipfs to download things: ipfs ls, ipfs get, ipfs object get are some examples
<achin>
what exactly is downloaded depends on the command
<achin>
for example, "ipfs get", when given a hash of a directory, will download the entire directory. but "ipfs object get <hash>" will only get the 1 mergkenode referenced by that hash
<achin>
ipfs also stores things when you "ipfs add" something (but obivously in this case, nothing is downloaded)
<achin>
right now, nothing every gets deleted automatically. you have to run a "garbage collection" command
<achin>
"ipfs repo gc". anything that is pinned (either recursively, directly, or indirectly) will be saved
<achin>
everything else is deleted
<achin>
so in your first question, is core question is: how often do the public IPFS gateways run the repo-gc command
devbug_ has joined #ipfs
devbug_ has quit [Ping timeout: 252 seconds]
<kpcyrd>
is /ipns/ a reference during gc? in that case I could just update the name and run gc to get rid of all old files
<sonatagreen>
is there still the bug where having /ipns mounted causes your site to not update reliably?
<sonatagreen>
oh dear, crash bug when adding my site:
<sonatagreen>
ERRO[00:02:59:000] multipart: unexpected line in Next(): "9c58843ea1fec1db077e2dcc8565c7f9ba2ca6d47fb5a80216bd2e0f3d42--\r\n" module=commands/http
<sonatagreen>
err, by crash bug i mean that particular command exited unsuccessfully
<sonatagreen>
the daemon and my computer are still running
<achin>
kpcyrd: no, an ipns is not something that will get cleaned up during a GC
<achin>
kpcyrd: as far as i'm aware, once you publish something, you can't unpublished it (except maybe by shutting down your node for 24 hours? whyrusleeping would know the answer to this)
<kpcyrd>
achin: including the tree it's referencing?
<achin>
that depends if the tree is pinned or not
leeola has joined #ipfs
<kpcyrd>
1) pin the new tree 2) update ipns 3) unpin old tree 4) run gc
<kpcyrd>
got it
<achin>
yep, that should do it
<kpcyrd>
thanks
Not_ has quit [Ping timeout: 256 seconds]
<sonatagreen>
srsly tho, wat do re: my error
<achin>
sonatagreen: i'd make sure you are running the latest version, and if so, file a bug report on github
<amstocker>
it would be nice to have the history stored in some kind of reference chain in case notifications are missed
gaboose has joined #ipfs
<davidar>
amstocker (IRC): yeah, my personal preference would be to store the actual messages as normal ipfs objects, and just use pubsub to send notifications
<amstocker>
yeah I 100% agree
<amstocker>
davidar: I am planning on trying to dig into SSB and see how it could best be integrated with ipfs
<davidar>
amstocker (IRC): cool
<davidar>
Lots of activity in the aggregation thread recently
<amstocker>
although im a node.js noob
<amstocker>
yeah I think it is something a lot of people are interested in
<amstocker>
it would be the foundation of any web2.0 style application
<amstocker>
like ports of reddit, twitter, etc.
<davidar>
Yeah, definitely
<davidar>
But now people are talking to me as if I actually know stuff about CRDTs :p
<amstocker>
haha
<amstocker>
i've read the paper jbenet passed around but thats about it
<sonatagreen>
this is the most /surreal/ bug I've ever seen
captain_morgan has quit [Ping timeout: 250 seconds]
<sonatagreen>
I am seriously tempted to hypothesize that the file is haunted
hugs-bison has joined #ipfs
<sonatagreen>
someone please figure this out so as to affirm my sanity
Eudaimonstro has quit [Remote host closed the connection]
<davidar>
sonatagreen (IRC): replied
Eudaimonstro has joined #ipfs
<davidar>
sonatagreen (IRC): BTW, I wanted to refine my comment from the other day. Whilst I don't agree we should necessarily avoid reimplementing anything that has already been done elsewhere, I do agree that we should be conservative in adding functionality to avoid scope creep
<sonatagreen>
Ideally I want to end up with a Modern Internet Suite
<sonatagreen>
ipfs for data storage, possibly whatever Matrix is doing for ephemeral realtime communications, maybe something like i2p to replace ip addresses
<sonatagreen>
each problem solved by a tool that's optimized for that use case, but everything working together as a seamless whole
<davidar>
Yeah, definitely
<davidar>
We just need to make sure ipfs has the right primitives to support all that stuff
<davidar>
Personally I think the daemon, gateway, and CLI should all be separated too
<sonatagreen>
I think one of the major unsolved problems is a sensible code sandbox. The modern web's killer app is the ability to safely-ish download and run untrusted code.
<davidar>
Yeah, Haskell is a bit if a double edged sword. It forces you to think through all of the details before it'll compile, which means you end up with a robust program, but it sucks for quickly hacking together a prototype
<sonatagreen>
goodnight
<davidar>
Night
azariah has quit [Ping timeout: 244 seconds]
azariah has joined #ipfs
amstocker has quit [Ping timeout: 255 seconds]
<ipfsbot>
[go-ipfs] rht opened pull request #1827: Metalint (master...metalint) http://git.io/vCnKy
Rajarshi has joined #ipfs
patcon has joined #ipfs
Vilhelm has quit [Ping timeout: 264 seconds]
<daviddias>
.seen whyrusleeping
<multivac>
daviddias: I last saw whyrusleeping at 2015-10-11 - 04:14:44 in here, saying or by just turning off your node and waiting 24 hours
amstocker has joined #ipfs
<daviddias>
multivac: I need to know the timezone of that 04:14:44 :)
ipfspics-Vincent has quit [Ping timeout: 246 seconds]
<davidar>
daviddias (IRC): UTC, i think
<daviddias>
thank you davidar :)
<daviddias>
4 hours ago then
<davidar>
yeah
chriscool has joined #ipfs
<daviddias>
.ask whyrusleeping which is the query string value I have to pass for a `ipfs files rm -r <path>`, not being able to get it right
<multivac>
daviddias: I'll pass that on when whyrusleeping is around.
chriscool has quit [Ping timeout: 264 seconds]
<ipfsbot>
[node-ipfs-api] diasdavid created feat/files-api (+2 new commits): http://git.io/vCnPw
<ipfsbot>
node-ipfs-api/feat/files-api a983384 David Dias: add test-usn to be able to use own local node (good for when testing new IPFS versions)
<ipfsbot>
node-ipfs-api/feat/files-api 718bdde David Dias: files api tests
slothbag has joined #ipfs
r1k0 has quit [Ping timeout: 246 seconds]
hjoest has quit [Ping timeout: 264 seconds]
hjoest has joined #ipfs
azariah has quit [Ping timeout: 250 seconds]
<whyrusleeping>
daviddias: sup
<multivac>
whyrusleeping: 2015-10-11 - 06:41:05 <daviddias> ask whyrusleeping which is the query string value I have to pass for a `ipfs files rm -r <path>`, not being able to get it right
hjoest2 has joined #ipfs
hjoest3 has joined #ipfs
hjoest4 has joined #ipfs
hjoest2 has quit [Read error: Connection reset by peer]
<multivac>
ion: I'll pass that on when sonatagreen is around.
vik_ has quit [Quit: Leaving]
hjoest has quit [Ping timeout: 264 seconds]
hjoest has joined #ipfs
<zignig>
.seen jbenet
<multivac>
zignig: I last saw jbenet at 2015-10-09 - 21:48:40 in here, saying Thanks multi ac
<zignig>
o/
Leer10 has quit [Ping timeout: 268 seconds]
Leer10 has joined #ipfs
od1n has quit [Ping timeout: 246 seconds]
zz_r04r is now known as r04r
Spinnaker has quit [Ping timeout: 252 seconds]
Spinnaker has joined #ipfs
Encrypt has joined #ipfs
gaboose has quit [Quit: Connection closed for inactivity]
G-Ray has joined #ipfs
NeoTeo has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
atrapado has joined #ipfs
Spinnaker has quit [Quit: sinked]
NeoTeo has joined #ipfs
mercora has quit [Quit: leaving]
mercora has joined #ipfs
Encrypt has quit [Quit: Quitte]
tsp has quit [Remote host closed the connection]
hjoest2 has joined #ipfs
hjoest has quit [Ping timeout: 264 seconds]
sseagull has joined #ipfs
sseagull has joined #ipfs
NeoTeo has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
Encrypt has joined #ipfs
<daviddias>
whyrusleeping: mornin' :)
<daviddias>
I meant through the HTTP API
<daviddias>
can you give me a curl example of how would you do a `ipfs files rm -r`? Not being able to get it to do the `-r`
G-Ray has quit [Ping timeout: 240 seconds]
voxelot has joined #ipfs
eater has quit [Ping timeout: 240 seconds]
eater has joined #ipfs
<ipfsbot>
[go-ipfs] jbenet closed pull request #1779: Use common progressbar function for cat and get (master...latest-progressbar) http://git.io/vcoOu
<jbenet>
daviddias: the sharness tests should have one i think
<daviddias>
oh! Nice, thanks jbenet . Now I see (I think) that it is actually the node-ipfs-api that is not encoding the { r : true} to a query string properly
chriscool has quit [Ping timeout: 250 seconds]
<daviddias>
got it!
G-Ray has joined #ipfs
demize has joined #ipfs
vanila has joined #ipfs
<vanila>
hi ipfs community
NeoTeo has joined #ipfs
G-Ray has quit [Quit: Konversation terminated!]
captain_morgan has joined #ipfs
Not_ has joined #ipfs
<ipfsbot>
[node-ipfs-api] diasdavid pushed 2 new commits to feat/files-api: http://git.io/vCCcZ
<ipfsbot>
node-ipfs-api/feat/files-api 1f810a9 David Dias: add readme documentation for files
<ipfsbot>
node-ipfs-api/feat/files-api 22d0d0e David Dias: fix rm -r bug
<whyrusleeping>
daviddias: g'mornin
<daviddias>
whyrusleeping: Hey o/ :)
atrapado has quit [Quit: Leaving]
<ipfsbot>
[node-ipfs-api] diasdavid opened pull request #74: feat/files-api (master...feat/files-api) http://git.io/vCCC7
devhoag has joined #ipfs
hellertime has quit [Quit: Leaving.]
<vanila>
i was thinking how nice it would be to have a VPS or something that ran an ipfs node 24/7
<vanila>
what sort of dangers are there to be aware of with such a setup?
<whyrusleeping>
vanila: thats what we do
captain_morgan has quit [Ping timeout: 256 seconds]
Encrypt has quit [Quit: Quitte]
<vanila>
how do i find the ipfs version of my neocities site?
NeoTeo has quit [Read error: Connection reset by peer]
Gaboose_ has quit [Ping timeout: 246 seconds]
domanic has joined #ipfs
NeoTeo has joined #ipfs
gaboose has joined #ipfs
Rodya has joined #ipfs
hellertime has joined #ipfs
HostFat_ has joined #ipfs
HostFat has quit [Ping timeout: 240 seconds]
rendar has quit [Quit: std::lower_bound + std::less_equal *works* with a vector without duplicates!]
HostFat has joined #ipfs
HostFat__ has joined #ipfs
<daviddias>
dignifiedquire was trying to get it working simply by removing that line
<daviddias>
it fired up the node
HostFat_ has quit [Ping timeout: 240 seconds]
<daviddias>
we shouldn't have to specify that in the file since it is always run in the node context
HostFat_ has joined #ipfs
HostFat_ has quit [Remote host closed the connection]
HostFat has quit [Ping timeout: 255 seconds]
hjoest2 has quit [Read error: Connection reset by peer]
hjoest3 has joined #ipfs
HostFat__ has quit [Ping timeout: 255 seconds]
<daviddias>
.tell voxelot that I only received his Oct 8 messages today (for some random reason, dunno why)
<multivac>
daviddias: I'll pass that on when voxelot is around.
nekomune has joined #ipfs
r04r is now known as zz_r04r
od1n has quit [Ping timeout: 240 seconds]
<ipfsbot>
[go-ipfs] miolini opened pull request #1830: Fix data race in commands/http/handler/ServerConfig (master...master) http://git.io/vCWz5
kytv_ has joined #ipfs
kytv has quit [Remote host closed the connection]
<Rodya>
hmm, does setting datastore.path not do anything? seems to still require ~/.ipfs/datastore to exist and still writes stuff to it... nothing got written to the directory i set it to
<davidar>
whyrusleeping (IRC): ooh, did you fix add performance?