digitalwestie_ has quit [Remote host closed the connection]
calavera has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
zootella has joined #ipfs
voxelot_ has quit [Ping timeout: 240 seconds]
analogv0id has joined #ipfs
wiedi has quit [Read error: Connection reset by peer]
zootella has quit [Quit: zootella]
jokoon has quit [Quit: Leaving]
cristian_ has joined #ipfs
wiedi has joined #ipfs
cristian_ has quit [Ping timeout: 244 seconds]
<noffle>
palkeo: o/
Oatmeal has quit [Ping timeout: 250 seconds]
<ipfsbot>
[go-ipfs] whyrusleeping created fix/assert-gopath (+1 new commit): https://git.io/v2n85
<ipfsbot>
go-ipfs/fix/assert-gopath dd3c42c Jeromy: assert that you must build from inside your gopath...
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #2399: assert that you must build from inside your gopath (master...fix/assert-gopath) https://git.io/v2n4G
<ipfsbot>
[go-ipfs] whyrusleeping pushed 3 new commits to master: https://git.io/v2n41
<ipfsbot>
go-ipfs/master 40b21bf Christian Couder: bin: add check_go_version...
<ipfsbot>
go-ipfs/master 6c959c8 Christian Couder: Makefile: add go_check target...
silotis has quit [Remote host closed the connection]
silotis has joined #ipfs
<rhalff>
I wonder, would it make sense to have dictionary words be unique known hashes and gradually all there combinations/links also, perhaps by making their hashes constant and non versionable?
<rhalff>
e.g. both hello & world have unique hashes and hello with a link to world also has it's unique hash, but non-versionable. making the hash for 'hello world' also a fixed hash etc.
<rhalff>
not sure if I'm totally on topic, I didn't dive too deep into ipfs yet.
<rhalff>
basically you can then pick any sentence and retrieve what ever document also contains/links that sentence.
Arakela007 has joined #ipfs
voxelot_ has quit [Ping timeout: 240 seconds]
pfraze has quit [Remote host closed the connection]
ygrek has quit [Ping timeout: 240 seconds]
jaboja has joined #ipfs
calavera has joined #ipfs
HastaJun has joined #ipfs
cristian_ has joined #ipfs
calavera has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
silotis has quit [Remote host closed the connection]
cristian_ has quit [Ping timeout: 244 seconds]
silotis has joined #ipfs
ike_ has quit [Quit: Connection closed for inactivity]
Arakela0_ has joined #ipfs
Arakela007 has quit [Ping timeout: 244 seconds]
zootella has joined #ipfs
ecloud is now known as ecloud_wfh
devbug has joined #ipfs
* jbenet
lgierth: im not sure whether the pin will be on the /ipfs or /ipns path. we should be pinning every version ever.
<jbenet>
!pin /ipfs/QmNu7McSMSi3vp6avTZrgW6MKdgeDkGjJGY6W5zPjnh9Xg the website as of 2016-02-24
<pinbot>
now pinning /ipfs/QmNu7McSMSi3vp6avTZrgW6MKdgeDkGjJGY6W5zPjnh9Xg
<jbenet>
Kubuxu: you mean "dnslink=/ipns/QmAAA...AAA/ovh-kubuxu/current" ? yeah this is by design
<jbenet>
Kubuxu: paths, not hashes ;)
<Kubuxu>
I mean by having one directory with all versions.
<jbenet>
Kubuxu: it also means "/<other-protocol>/..." is supported too.
<jbenet>
Kubuxu: yes, right. I'm actually writing about that now. thinking of how to explore a version graph this way, as the root changes.
<Kubuxu>
And having in this directory 'current' directory.
<jbenet>
Kubuxu: with timestamping guarantees (to a blockchain or similar timestamping service)
<Kubuxu>
I just figured out that chaining isn't the best solution. At least from usability perspective.
jaboja has quit [Ping timeout: 244 seconds]
<jbenet>
Kubuxu: i'm using "master" to follow the git nomenclature, "latest" was also a good name.
<jbenet>
Kubuxu: right. but the root will need to be sharded for sure.
<jbenet>
Kubuxu: i also see this as an index on the commit chain-- so i would still chain the commits for git-like traversal, but also keep a fast access index
<Kubuxu>
Right, thinking of it, it should be really easy to have a chain and index.
<jbenet>
and <index-root>/<commit-index>/target/index.html -- where <commit-index> is some key in the index. ints are weird, because ints imply order and you can only order if you have a total order. meaning chain forks are complicated (without timestamps.)
<jbenet>
but i know a lot of people like ints because they make reasoning about totals easier.
<jbenet>
can also maintain indexes by time-- <index-root>/2016/02/24/0/target/index.html
<jbenet>
and so on-- can do whatever you want. yay ipld.
<Kubuxu>
Is current IPLD working?
<Kubuxu>
The _examples seem outdated.
<jbenet>
the spec is shipped. need an implementation
<Kubuxu>
So go-ipld is really place holder now.
<jbenet>
Kubuxu: no, it's close to done. mildred has done a great job with it. it just needs tweaks i think
<jbenet>
given shipped spec
<Kubuxu>
Ok, thanks for that.
<Kubuxu>
I will look into making first few CRDT on top of IPLD, I will try to make pinning cluster on top of only IPFS.
<Kubuxu>
One question: can I send data directly via go-ipfs/go-p2p channels, without running custom daemon?
<haad>
morning all
<Kubuxu>
morning
reit has quit [Ping timeout: 244 seconds]
ygrek has joined #ipfs
<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to fix/assert-gopath: https://git.io/v2cRx
<ipfsbot>
go-ipfs/fix/assert-gopath 65519a0 Jeromy: move gopath check to separate script and use realpath for symlink handling...
<jbenet>
Kubuxu: you will be able to, right now no. there's a PR for it from AtnNn -- but i havent looked at it in a bit. we should think through as a "mounted transport". i've wanted something like "ipfs transport (or xport) {dial, listen, mount, ...}" we should look carefully at how plan9 did network ops with fs-like semantics and do something similar in the API
<whyrusleeping>
unix domain sockets are pretty nifty in that regard
<Kubuxu>
jbenet: thinking about it, storing links to all other commits in each commit will have annoying scaling. If you don't plan chain merging consider making links to 2^n commits back.
Senji has joined #ipfs
<Kubuxu>
This way you have O(log N) scaling with O(log N) size scaling. In comparison with O(1) and O(N).
<Kubuxu>
re: transport. So I will use plain TCP now and switch over to IPFS when it is here.
<Kubuxu>
Also is go-ipfs-api getting any love or how it looks like?
elima_ has joined #ipfs
<jbenet>
Kubuxu: it needs so much love. please turn it into js-ipfs-api
<whyrusleeping>
Kubuxu: yeah, it needs love
<jbenet>
whyrusleeping: the end is nigh
<jbenet>
whyrusleeping: also that guy with the hockey stick is a jerk. poor bot. he'll regret it later, when the robot overlords watch this video later.
<whyrusleeping>
on a scale of one to ten, i'm kinda worried
<whyrusleeping>
yeah
<whyrusleeping>
that hockey stick guy is screwed
jaboja has joined #ipfs
ulrichard has joined #ipfs
gvbgduh has quit [Quit: gvbgduh]
ygrek has quit [Ping timeout: 244 seconds]
diffalot has joined #ipfs
mildred has joined #ipfs
ipfsrocks has quit [Quit: Lost terminal]
TheWhisper_ has joined #ipfs
TheWhisper has quit [Ping timeout: 244 seconds]
ygrek has joined #ipfs
jaboja has quit [Ping timeout: 244 seconds]
gvbgduh has joined #ipfs
hashcore has joined #ipfs
Aeon has quit [Ping timeout: 276 seconds]
<tmg>
telling that they didn't show what happens when the robot falls onto the box it wants to pick up
jaboja has joined #ipfs
Aeon has joined #ipfs
Aeon has quit [Changing host]
Aeon has joined #ipfs
zz_r04r is now known as r04r
devbug has quit [Ping timeout: 250 seconds]
computerfreak has joined #ipfs
computerfreak has quit [Remote host closed the connection]
mrichter has joined #ipfs
computerfreak has joined #ipfs
reit has joined #ipfs
mrichter has quit [Ping timeout: 240 seconds]
ianopolous2 has quit [Ping timeout: 244 seconds]
<dignifiedquire>
jbenet: still around?
reit has quit [Quit: Leaving]
reit has joined #ipfs
GonZo2000 has quit [Ping timeout: 264 seconds]
jaboja has quit [Ping timeout: 252 seconds]
Oatmeal has quit [Ping timeout: 240 seconds]
s_kunk has joined #ipfs
<jbenet>
dignifiedquire: yeah briefly, but will go soon
Pharyngeal has quit [Ping timeout: 240 seconds]
ygrek has quit [Ping timeout: 244 seconds]
Pharyngeal has joined #ipfs
Encrypt has joined #ipfs
joshbuddy has quit [Read error: Connection reset by peer]
joshbuddy has joined #ipfs
Oatmeal has joined #ipfs
wking has quit [Ping timeout: 248 seconds]
Tv` has quit [Quit: Connection closed for inactivity]
<daviddias>
so apparently I am able to create repos on IPFS org that I'm not the admin of
<daviddias>
and now I can't add myself
<daviddias>
blaaah
hellertime has quit [Quit: Leaving.]
<dignifiedquire>
daviddias: :D
<daviddias>
ahahah
hellertime has joined #ipfs
computerfreak has quit [Quit: Leaving.]
pfraze has joined #ipfs
zorglub27 has quit [Ping timeout: 250 seconds]
cow_2001 has quit [Quit: Offense is taken, not given. No need to disarm the world. Just make yourself bullet proof. Sticks & stones people. Block. Unfollow. Laugh.]
cow_2001 has joined #ipfs
Senj has joined #ipfs
Senji has quit [Disconnected by services]
leer10 has joined #ipfs
<dignifiedquire>
damn it I'm sooo close
<lgierth>
jbenet: it'll resolve first -- i was just testing
<lgierth>
dignifiedquire: yeah! let's go
reit has quit [Ping timeout: 244 seconds]
<dignifiedquire>
lgierth: so, my running node as it pinned:
<dignifiedquire>
IPFS_PATH=~/.ipfs-4 ipfs pin ls | grep QmRn43NNNBEibc6m7zVNcS6UusB1u3qTTfyoLmkugbeeGJ
<lgierth>
i'll just pull it into a fresh fs-repo and downgrade that
<dignifiedquire>
ah are you pulling it out manually?
<dignifiedquire>
nice
<lgierth>
:)
<dignifiedquire>
lgierth: do you still need my node or can I shut it down for now?
elima has joined #ipfs
<lgierth>
dignifiedquire: you can shut it down
<dignifiedquire>
daviddias: lgierth whyrusleeping voxelot_ planning to do first episode dignified hacks on Friday evening my time, what do you think?
joshbuddy has joined #ipfs
<whyrusleeping>
dignifiedquire: i wont be able to make it
<lgierth>
me neither :(
<lgierth>
will be on the way to amsterdam
<voxelot_>
dignifiedquire: i should be in the office in the morning for me, i'll tune in, just let me know when
<whyrusleeping>
i'll be on my way to seattle with a trailer full of things
ygrek has joined #ipfs
<dignifiedquire>
all these people running around in the world
chriscool1 has quit [Read error: Connection reset by peer]
<dignifiedquire>
will see if daviddias is around, if he can't Make it either I'll look for another time, otherwise there will be recordings plus a new episode next week :)
chc4 has joined #ipfs
<whyrusleeping>
wooo!
<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to feat/dag-refactor: https://git.io/v2W5u
<ipfsbot>
go-ipfs/feat/dag-refactor cba00b9 Jeromy: fixes from review...
<dignifiedquire>
I just saw I can upgrade my connection from 100mbit to 400mbit Oo
<dignifiedquire>
I am very very tempted
<Kubuxu>
If you have more than 5Mbit upload you don't need it, at least for streaming.
<dignifiedquire>
I have 4.5 atm
<dignifiedquire>
I really don't need it, but it would be still cool to have
<dignifiedquire>
lgierth: whyrusleeping would monday evening be better? if we end up changing the schedule I could do it after the community hangout
<whyrusleeping>
monday evening could work
s_kunk has quit [Ping timeout: 244 seconds]
voxelot_ has quit [Ping timeout: 240 seconds]
elima has quit [Ping timeout: 276 seconds]
ygrek has quit [Ping timeout: 252 seconds]
zorglub27 has quit [Ping timeout: 240 seconds]
ianopolous has joined #ipfs
elima has joined #ipfs
fiatjaf has quit [Remote host closed the connection]
TheWhisper has quit [Read error: Connection reset by peer]
zorglub27 has joined #ipfs
TheWhisper has joined #ipfs
devbug has quit [Ping timeout: 244 seconds]
chc4 has quit [Ping timeout: 240 seconds]
<lgierth>
dignifiedquire: ok i think we have a proper bug -- this time i'm fetching from biham, on biham, and it's getting stuck again
<lgierth>
dignifiedquire: i'll definitely be there on monday hehe
<lgierth>
i might have ansuz with me
analogv0id has joined #ipfs
analogv0id has quit [Remote host closed the connection]
chc4 has joined #ipfs
<noffle>
geez github notifications and code reviews take up so much time -- almost noon already
ygrek has joined #ipfs
zorglub27 has quit [Ping timeout: 250 seconds]
voxelot_ has joined #ipfs
cristian_ has quit [Ping timeout: 244 seconds]
zorglub27 has joined #ipfs
fuzzybear3965 has joined #ipfs
cristian_ has joined #ipfs
fuzzybear3965 has left #ipfs [#ipfs]
fil_redpill has joined #ipfs
<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to feat/dag-refactor: https://git.io/v2lqx
<ipfsbot>
go-ipfs/feat/dag-refactor ec3959a Jeromy: use an option type to simplify concurrency...
<noffle>
!!!
<ipfsbot>
[js-ipfs-api] xicombd opened pull request #218: Avoid calling the cb twice when statusCode > 400 (master...fix/double-cb-onres) https://git.io/v2lOC
analogv0id has joined #ipfs
danielrf has quit [Quit: WeeChat 1.4]
analogv0id has quit [Ping timeout: 276 seconds]
ugjka has quit [Ping timeout: 240 seconds]
Senji has quit [Ping timeout: 255 seconds]
mildred has joined #ipfs
<noffle>
does anyone have any strong feelings against 'ipfs <cmd>' showing the subcommands without needing to explicitly type 'ipfs <cmd> --help'?
<ipfsbot>
[js-ipfs] xicombd created feature/http-api-config (+1 new commit): https://git.io/v2lna
<ipfsbot>
js-ipfs/feature/http-api-config 286ae40 Francisco Baio Dias: Add /api/v0/config endpoint
cow_2001 has quit [Quit: Offense is taken, not given. No need to disarm the world. Just make yourself bullet proof. Sticks & stones people. Block. Unfollow. Laugh.]
<alu>
whyrusleeping: what are the common troubleshooting routines I should know?
fleeky has quit [Remote host closed the connection]
<alu>
I love using IPFS in this metaverse development workflow but lately it's been hit or miss ;(
fleeky has joined #ipfs
<whyrusleeping>
alu: what do you mean by ping the hash?
<whyrusleeping>
do you have a few different machines running?
<whyrusleeping>
alu: btw, someone saw your business card in my pile of stickers and thought it was hilarious
<alu>
(´・ω・`)
<alu>
in the code I have ipfs.io so I have to change that too now
<dignifiedquire>
alu: good luck for your presentation, excited to see what will come out of your work!
<whyrusleeping>
alu: yeah, let me know anything weird
<whyrusleeping>
if you see more hanging don't hesitate to file an issue about it
montagsoup has joined #ipfs
<alu>
whens 0.4.0 ? I will change the links in the html now but this seems like a meh fix.
<dignifiedquire>
also more tests are coming
<dignifiedquire>
alu: when the big bad tests all pass
cow_2001 has joined #ipfs
<dignifiedquire>
we are trying to make sure it's always a hit before releasing it
<alu>
exciting
<whyrusleeping>
alu: agreed its a meh fix...
<dignifiedquire>
whyrusleeping: I get the feeling multireq is doing more harm than good :/
<whyrusleeping>
dignifiedquire: agreed.
<whyrusleeping>
lgierth is on it, hopefully we can bang this out today/tonight
<dignifiedquire>
ok cool
ansuz has quit [Ping timeout: 250 seconds]
ianopolous has joined #ipfs
danielrf has joined #ipfs
<palkeo>
noffle: o/
<noffle>
palkeo: \o
<palkeo>
cool :)
ashark has joined #ipfs
<palkeo>
you remember the issue I opened about « could not resolve name » ?
<palkeo>
I'd like to understand better :)
cristian_ has quit [Ping timeout: 244 seconds]
<noffle>
did my response work / make sense?
<palkeo>
I have not tried it yet
<palkeo>
but I don't understand : why my node would resolve its own IPNS address on the network ?
<noffle>
each node has it's own public key
<noffle>
but that doesn't mean many computers couldn't be running nodes with that same public key
<noffle>
so we need to check the network to try and make sure we have the freshest value
<palkeo>
that should not happen, right ?
<noffle>
it's a legitimate use case
<noffle>
nothing prevents it
<noffle>
(probably not common though, I agree)
<palkeo>
ok, but the routing system would still work even with multiple nodes with the same public key ?
<noffle>
yes
<palkeo>
ok
<noffle>
peers are identified by their pubkey *and* their network address
<palkeo>
and also, « the freshest value » but value of what ?
<palkeo>
oooh ok
<noffle>
of their ipns entry. each node can map their pubkey to point to an /ipfs/Qmfoobar address
<noffle>
mounting /ipns locally means looking up this value on the network
<palkeo>
oh
<palkeo>
the IPNS name for my node is my public key
<noffle>
correct
<palkeo>
and it maps to a multihash, containing all the things I have published under my public key ?
<noffle>
but if you've never published a value for the name to map to, the network won't find it
<noffle>
exactly :)
<palkeo>
cool, I see :)
<palkeo>
yes, so I need to resolve it
<palkeo>
makes perfect sense !
<palkeo>
thanks !
<noffle>
\o/
<palkeo>
hmmm… isn't that dangerous ? say, I have 2 nodes with the same public key
<palkeo>
A and B
<palkeo>
A publish something, goes down
<palkeo>
a day later, B is powered on, publish something, goes down
<noffle>
what's the danger?
<palkeo>
a day later, A is powered up again, and I suppose that the network won't remember what B had published, and because it's down… so A will publish a new multihash containing the old value it had, ignoring what B published
<palkeo>
is that right ?
<noffle>
yep
<noffle>
ipns doesn't offer consistency
<palkeo>
ok
<noffle>
it's best effort
<palkeo>
and what happens if B stays up
<noffle>
you can also possibly just get old values too
<palkeo>
will A automatically merge what it had with what B published ?
<noffle>
some % of the DHT will see A's value, and some % will see B's value
<noffle>
if they're both publishing
<palkeo>
yes, ok… they are signed anyway so seems legit
mildred has quit [Ping timeout: 244 seconds]
<palkeo>
oh, and do you have a project to make statistics on what is passing by the network ?
<palkeo>
could one make a "virtual" node that would advertise as multiple nodes that would have addresses that are uniformly distributed in the address space
<palkeo>
so it could see all the requests that are made
<palkeo>
for any multihash
<palkeo>
and spy for all the content that is requested
<palkeo>
(for a search engine, whatever)
<palkeo>
but also make statistics
<palkeo>
and have an idea about the peers in the whole network (maybe the peer count, their country…)
<ipfsbot>
[js-ipfs-api] dignifiedquire closed pull request #218: Avoid calling the cb twice when statusCode > 400 (master...fix/double-cb-onres) https://git.io/v2lOC
<montagsoup>
uniform distribution would be extremely hard since hashes are made to avoid that
<whyrusleeping>
yeah, uniform distribution isnt easy
<whyrusleeping>
especially since one 'distribution' of hashes, will only be uniformly distributed relative to one key
<montagsoup>
you would have to find private keys to make just the right public keys that make just the right hashes
<palkeo>
say the address space is from 0 to 1000, and I want 100 public keys. I would generate a bunch of private keys, derive the public key, and take the one that is closest to 10, then to 20 then to 30…
<palkeo>
so it would be approximately uniformly distributed, right ?
<whyrusleeping>
not exactly
<whyrusleeping>
being close to '30' is possible
<whyrusleeping>
but being close to 30 doesnt mean ANYTHING about being close to 29 or 31
<palkeo>
what ? the addresses are public keys right ?
<whyrusleeping>
right
<whyrusleeping>
xor is the distance metric
<palkeo>
ooooh
<palkeo>
I see
<whyrusleeping>
yeah
<whyrusleeping>
yay math!
<palkeo>
:)
<palkeo>
why did you choose xor and not the classic distance metric ?
<palkeo>
(I never though of xor this way, it's cool)
<montagsoup>
you can perform look ups in logarithmic time instead of linear
<montagsoup>
every bit you process splits the remaining space to test in half
<palkeo>
can't you do the same dichotomic search with the canonical distance metric ?
<brimstone>
are all multihashes 46 characters long and prefixed with Qm?
zorglub27 has joined #ipfs
<montagsoup>
brimstone: no just the ones using the same hash function
<montagsoup>
the hash itself describes what function it uses
<montagsoup>
It's called a multihash.
<palkeo>
hmmm… I see that xor distance is the same as in kademlia. Do you have good articles about kademlia to recommend ? (or should I just read IPFS sourcecode ? ^^)
<montagsoup>
the wikipedia article is alright. that's all i've read so far
<palkeo>
ok
<palkeo>
and in the case of IPFS, the nodes storing a « value » for a key (a multihash), have the list of addresses of nodes that actually store the data ?
<brimstone>
montagsoup: oh neat, what part of the multihash identifies the hashing function?
<alu>
dumb question but my friend just downloaded ipfs to a VM and has the webgui already working out of the box
<alu>
and im like o yeah i forgot how cool that webgui is lemme try it
<alu>
but its not showing up on my end
<montagsoup>
brimstone: it's the first few bits. i'll try to find the documentation for you.
<brimstone>
montagsoup: sure a link would be great
<devbug>
what's the general consensus on using IPFS as a backing store for other distributed tech?
<rendar>
devbug? other like what?
calavera has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
vanila has quit [Quit: Leaving]
<devbug>
rendar: package management and build caching
calavera has joined #ipfs
<whyrusleeping>
Quick announcement, we're doing a bit of work on the gateways, if things are more weird than usual, i apologize. I'll update again when things should be running smoothly
<whyrusleeping>
devbug: backing package managers with ipfs is something we're very interested in
<devbug>
whyrusleeping: :)
<devbug>
i actually need some smart people to rubber duck off of too
analogv0id has joined #ipfs
<devbug>
the idea is to build a dag of packages (and their build artifacts)
<devbug>
(this is specific to C/C++ (and C-compatabile) source packages)
<whyrusleeping>
i'm listening
<devbug>
so foreword: this spun out of a simple idea of better managing dependencies for yeti (a game engine)
<devbug>
thus, it's biased towards the "game developer" way of looking at builds
<devbug>
assumptions: platforms are basically homogenous (at least in feature set), compilation options don't vary too much between users, etc.
ulrichard has quit [Remote host closed the connection]
<devbug>
anyways, idea is: you define your dependencies as "packages" w/ build matrices
<devbug>
so, like: "I need LuaJIT on windows, mac, and linux, for x86/x86_64, using these build options in 3 configurations (debug, dev, release)"
Encrypt has quit [Quit: Quitte]
<devbug>
this creates a whole bunch of inbound edges to your node
chriscool has quit [Read error: No route to host]
<devbug>
these edges orignate from other nodes, which point to other nodes and so on and so forth
<devbug>
until you get sha256(tarball(package contents)) (or any CAS scheme)
sobr__ has joined #ipfs
<devbug>
think refs/tags/xxx and some CAS objects
<devbug>
if you run with this concept, however, it would be possible to cache builds:
kvda has joined #ipfs
<devbug>
they're just functions of some node w/ extra inputs
<devbug>
so if you make the dag granular enough to where you're tracking inputs/outputs of your build artifacts (game engine shared lib, game engine runtime exe, eg)
<devbug>
you could replace parts of the graph w/ cached parts (i.e. binary packages)
sobr_ has quit [Ping timeout: 255 seconds]
<devbug>
the first part (source packages) is pretty easy to figure out implement. it's like a bunch of orphaned git commits
<devbug>
the latter part, not so much
<devbug>
so far, I've been thinking "sha256(json(options))", where the ordering is well-defined and such
<devbug>
i.e. cached builds are refs to output objects
pfraze has quit [Remote host closed the connection]
<devbug>
the issue is then verifying that the artifacts are indeed correct
<lgierth>
dignifiedquire: geoip is gonna come to v03x too -- right after we've figured out this gateway hang issue
Soft has quit [Read error: Connection reset by peer]
reit has joined #ipfs
<whyrusleeping>
devbug: i like where youre going with that
cow_2001 has quit [Quit: Offense is taken, not given. No need to disarm the world. Just make yourself bullet proof. Sticks & stones people. Block. Unfollow. Laugh.]
cow_2001 has joined #ipfs
<whyrusleeping>
devbug: using the cached artifacts would require an understanding of how the source inputs map to the outputs
<devbug>
and verification would require either: trusting someone (who signed it), or forming a consensus
<whyrusleeping>
yeah
<devbug>
i have a feeling that verifiable builds aren't a thing with msvc tho
<whyrusleeping>
but in a gamedev environment, you generally can trust the people youre working with
<devbug>
only sometimes ;)
rendar_ has joined #ipfs
<devbug>
i.e. two different links might be ordered differently
<whyrusleeping>
i *think* you can verify msvc builds
ashark has quit [Ping timeout: 244 seconds]
<whyrusleeping>
i remember something in the project options a while back
<whyrusleeping>
way back when i wrote c#
rendar has quit [Ping timeout: 255 seconds]
<whyrusleeping>
dignifiedquire: ping?
<whyrusleeping>
ping pls?
cristian_ has joined #ipfs
Akaibu has joined #ipfs
Soft has joined #ipfs
calavera has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<Akaibu>
whyrusleeping: might do some good popping this onto the ipfs awesome repo
<whyrusleeping>
Akaibu: yeah, wanna file a PR over there?
<whyrusleeping>
Akaibu: also, if the ipfs.io gateways are a little flaky, we're working on them
<whyrusleeping>
just a heads up
<Akaibu>
i'm not the one making this whyrusleeping :)
<Akaibu>
its some channer
<whyrusleeping>
ah, gotcha
<Akaibu>
also,sure whyrusleeping, in a moment, lgierth: i'm not sure how exactly that is relevant, if you would mind, go explain how that might be an issue in the 8ch thread i linked, the dev reads all the messages on there