avastmick has quit [Read error: Connection reset by peer]
rgrinberg has joined #ipfs
avastmick has joined #ipfs
ianopolous has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
gmcquillan has quit [Ping timeout: 240 seconds]
<noffle>
sugarpuff: you can get most of the way there today with ipfs. using orbit-db you could run a redis server for pubsub and a signalling server for webrtc, then use js-ipfs for comments & storage
avastmick has quit [Read error: Connection reset by peer]
<noffle>
but redis-less pubsub is coming soon too
herzmeister has quit [Ping timeout: 240 seconds]
avastmick has joined #ipfs
mossman93 has quit [Ping timeout: 240 seconds]
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
cemerick has joined #ipfs
wallacoloo has joined #ipfs
<ipfsbot>
[js-ipfs] nginnever created add from master (+0 new commits): https://git.io/vo44q
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
zdm has quit [Ping timeout: 244 seconds]
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
ipfsrocks has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
dguttman has quit [Quit: dguttman]
stoogenmeyer has quit [Remote host closed the connection]
cketti has quit [Quit: Leaving]
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
A124 has quit [Quit: '']
avastmick has quit [Read error: Connection reset by peer]
stoogenmeyer has joined #ipfs
avastmick has joined #ipfs
A124 has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
stoogenmeyer has quit [Remote host closed the connection]
stoogenmeyer has joined #ipfs
mgue has quit [Ping timeout: 250 seconds]
clownpriest has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
ipfsrocks has quit [Ping timeout: 272 seconds]
avastmick has quit [Read error: Connection reset by peer]
tymat has joined #ipfs
avastmick has joined #ipfs
mgue has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
<ipfsbot>
[js-ipfs-api] noffle opened pull request #296: Add files.get command and tests (master...ipfs-files-get) https://git.io/vo4E6
avastmick has quit [Read error: Connection reset by peer]
vijayee has joined #ipfs
<sugarpuff>
noffle: but it sounds like for it to be reliable a server needs to be constantly running, and a constantly running server costs $, so it's pretty much what i said earlier, you have to pay someone to host your comments
avastmick has joined #ipfs
<noffle>
sugarpuff: you need energy to store and serve data, so that has to come from somewhere, distributed or not
<noffle>
ideally it's your blog readers, if you have enough of them
<sugarpuff>
i know, that's the conclusion i've come to, that you can't get rid of the server
<sugarpuff>
at least not for blog comments
<sugarpuff>
*random blog comments
<sugarpuff>
in P2P terms, you have to have at least 1 seeder online at any given moment :P
<noffle>
yes, the store+serve energy needs to come from somewhere
<noffle>
though it need not be a seeder (doesn't NEED full data set to serve partial data)
<sugarpuff>
noffle: partial data isn't useful for anyone
<noffle>
sugarpuff: if you could only read your newest 50 emails, would that be *useless* to you?
<noffle>
definitely still lots of utility
avastmick has quit [Read error: Connection reset by peer]
<sugarpuff>
noffle: blog comments != email
<sugarpuff>
blog comments = other people's email
<noffle>
do you read 100% of the blog comments on every blog?
<noffle>
would they make no sense if you had only 99/100 of them?
<sugarpuff>
noffle: i require them to be there
<noffle>
bet you could go even lower
<noffle>
and still have utility
avastmick has joined #ipfs
<sugarpuff>
noffle: would you prefer to read a book with 1 page missing or 0 pages missing?
<noffle>
sugarpuff: you're starting to get it now :) -- context and what is missing matters
<noffle>
some things are ok
<noffle>
others aren't
<noffle>
but usually never 100% is needed for value
<sugarpuff>
i'm just saying, i wouldn't be OK with any comments missing. at the very least i would like to know how many were missing
ryjen has joined #ipfs
<noffle>
many years of using centralized services has given you certain expectations
<noffle>
that's fair
Kane` has quit [Remote host closed the connection]
Kane` has joined #ipfs
<noffle>
there are trade-offs like these in distributed vs centralized systems
Kane` has quit [Client Quit]
avastmick has quit [Read error: Connection reset by peer]
<noffle>
if consistency is essential to you, you'll need to ensure you have 1+ nodes running to make sure everything is consistent
<noffle>
i.e. a "server"
Kane` has joined #ipfs
<substack>
freedom is about accepting that sometimes the trains will be late
avastmick has joined #ipfs
<substack>
you *can* make the trains run on time with grim efficiency, but at what cost?
ryjen has quit [Ping timeout: 252 seconds]
<sugarpuff>
it's not like we have no freedom if servers are still a thing
dmr has joined #ipfs
<sugarpuff>
just call it a dedicated peer if it makes you happy :)
<noffle>
sugarpuff: the nice thing about ipfs is that there's no priveleged peer like there is with a centralized system (the "server") -- any node can act as that dedicated peer keeping consistency. users can donate toward that consistency goal /wo your knowledge or asking too
<noffle>
I think that's a big nicety over the centralized model: infrastructure transparency
<noffle>
we all get to be the cloud
<sugarpuff>
noffle: i agree that that's a property of content-addressable addressing
wallacoloo has quit [Quit: wallacoloo]
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
Senji has quit [Disconnected by services]
Senj has joined #ipfs
Senj is now known as Senji
Senji has quit [Max SendQ exceeded]
Senji has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
mgue has quit [Quit: WeeChat 1.5]
avastmick has quit [Read error: Connection reset by peer]
xioustic has quit [Read error: Connection reset by peer]
PrinceOfPeeves has quit [Quit: Leaving]
mgue has joined #ipfs
dignifiedquire has quit [Quit: Connection closed for inactivity]
<ipfsbot>
[go-ipfs] whyrusleeping deleted feature/test-repeated-ip at 32cdbbd: https://git.io/vo4gL
sydbarrett74 has joined #ipfs
HostFat has quit [Quit: Leaving]
pfista has quit [Ping timeout: 260 seconds]
stoogenmeyer has quit [Remote host closed the connection]
dguttman has joined #ipfs
dguttman has quit [Client Quit]
stoogenmeyer has joined #ipfs
levelheaded has joined #ipfs
ryjen has joined #ipfs
dguttman has joined #ipfs
stoogenmeyer has quit [Ping timeout: 240 seconds]
dguttman has quit [Client Quit]
ryjen has quit [Ping timeout: 252 seconds]
pfista has joined #ipfs
MadcapJake has quit [Read error: Connection reset by peer]
jrusso has joined #ipfs
jrusso is now known as MadcapJake
trn has quit [Quit: quit]
trn has joined #ipfs
go1111111 has quit [Ping timeout: 276 seconds]
avastmick has joined #ipfs
rgrinberg has quit [Ping timeout: 272 seconds]
cemerick has quit [Ping timeout: 240 seconds]
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
dguttman has joined #ipfs
avastmick has joined #ipfs
dguttman has quit [Ping timeout: 252 seconds]
ipfsrocks has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
wallacoloo has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
herzmeister has quit [Ping timeout: 250 seconds]
avastmick has quit [Read error: Connection reset by peer]
Kane` has quit [Ping timeout: 250 seconds]
avastmick has joined #ipfs
M-nd has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
Kane` has joined #ipfs
dguttman has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
Tv` has quit [Quit: Connection closed for inactivity]
dguttman has quit [Ping timeout: 258 seconds]
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
Senji has quit [Disconnected by services]
Senj has joined #ipfs
avastmick has quit [Read error: Connection reset by peer]
avastmick has joined #ipfs
pfista has quit [Ping timeout: 246 seconds]
avastmick has quit [Read error: Connection reset by peer]
<achin>
i had stubbed out the API part so i wouldn't hit API rate limits during development
Senj has quit [Ping timeout: 250 seconds]
Senji has joined #ipfs
atrapado has quit [Ping timeout: 244 seconds]
<achin>
in related news, i've started using nodejs for some minor webdev stuff at work. i can't tell if i like it or hate it. it sure it different from what i'm used to
<M-davidar>
achin: resist the dark side :p
<achin>
:D
apiarian has joined #ipfs
ygrek has joined #ipfs
dguttman has joined #ipfs
apiarian has quit [Ping timeout: 250 seconds]
dguttman has quit [Ping timeout: 246 seconds]
apiarian has joined #ipfs
ligi has joined #ipfs
ligi has quit [Ping timeout: 250 seconds]
ylp1 has quit [Ping timeout: 258 seconds]
rgrinberg has joined #ipfs
PrinceOfPeeves has joined #ipfs
stoogenmeyer has joined #ipfs
stoogenmeyer has quit [Remote host closed the connection]
stoogenmeyer has joined #ipfs
ligi has joined #ipfs
ylp1 has joined #ipfs
computerfreak has joined #ipfs
<circa2050>
Kubuxu: i see, thanks
dguttman has joined #ipfs
ygrek has quit [Ping timeout: 244 seconds]
ligi has quit [Ping timeout: 252 seconds]
dguttman has quit [Ping timeout: 240 seconds]
<lgierth>
Kubuxu: we're waiting for DTLSv3?
rgrinberg has quit [Quit: WeeChat 1.5]
<Kubuxu>
I mean, you won't be able to do TLS on utp.
rgrinberg has joined #ipfs
<Kubuxu>
Unless you build TCP on utp.
<Kubuxu>
TLS requires sequencing and reliability AFAIK
<lgierth>
yeah
<lgierth>
but doesn't utp give us both?
<lgierth>
i was assuming you were mentioning DTLS for non-reliable transports
<lgierth>
anyhow you're right it's nicer to do utp-over-dtls than tls-over-utp
<lgierth>
just to have moar crypto
<lgierth>
if you could point me to some DTLS 2 vs DTLS 3 that'd be great :D
<lgierth>
while i look into gateway issues *again* :)
<Kubuxu>
There is no DTLSv1.3 yet, they usually come out after TLS1v1.x
<Kubuxu>
Also I am looking forward to hacking in Berlin, ehmry found place for me, it will be great.
ligi has joined #ipfs
stoogenmeyer has joined #ipfs
rombou has joined #ipfs
corvinux has quit [Ping timeout: 272 seconds]
chungy has joined #ipfs
ligi has quit [Ping timeout: 250 seconds]
stoogenmeyer has quit [Remote host closed the connection]
ligi has joined #ipfs
ygrek has quit [Remote host closed the connection]
cemerick has quit [Ping timeout: 250 seconds]
computerfreak has quit [Quit: Leaving.]
computerfreak has joined #ipfs
stoogenmeyer has joined #ipfs
ligi has quit [Read error: No route to host]
stoogenmeyer has quit [Remote host closed the connection]
slothbag has quit [Quit: Leaving.]
cemerick has joined #ipfs
matoro has quit [Ping timeout: 250 seconds]
matoro has joined #ipfs
pfista has joined #ipfs
<lgierth>
Kubuxu: is there a way to check for offline mode through the api?
musgravejw has quit [Read error: Connection reset by peer]
<Kubuxu>
Hmm I don't think so.
<Kubuxu>
Also I think it would be best to separate offline mode from ephemeral mode.
<lgierth>
looking into splitting off the gateway and wondering if i can have it use the api exclusively
<lgierth>
but the resolve function should return an error right?
dguttman has joined #ipfs
stoogenmeyer has joined #ipfs
Akaibu has joined #ipfs
<r0kk3rz>
ephemeral mode?
<Kubuxu>
lgierth: yeah, it will fail right away
<Kubuxu>
r0kk3rz: when there is no daemon running but you run `ipfs cat`
<lgierth>
welll the terminology is confusing because ephemeral mode is called offline mode in the commands code
<r0kk3rz>
Kubuxu: so it just reads the cache, which is different to offline mode how exactly?
<lgierth>
offline means it doesn't do networking
<Kubuxu>
No, if there is no such block `ipfs cat` will connect to network
ylp1 has quit [Quit: Leaving.]
<Kubuxu>
or not
<Kubuxu>
No it won't.
<Kubuxu>
but offline != ephemeral really
dguttman has quit [Ping timeout: 252 seconds]
<lgierth>
yeah
<lgierth>
i agree
stoogenmeyer has quit [Ping timeout: 258 seconds]
<r0kk3rz>
yeah i think i get it now, 'ephemeral' is more like the js client
<lgierth>
ephemeral is when there's no daemon running, and the command just goes and reads directly from the repo (which it can because there's noone else and it can take the repo lock)
<lgierth>
offline is there's a daemon but it won't connect to anybody
<lgierth>
and then in the code there's also the concept of "online"
<r0kk3rz>
ephemeral should connect if it needs to, so then its like an on-demand daemon
<r0kk3rz>
which would be pretty nice for mobile use
<r0kk3rz>
i guess systemd can also on-demand the actual daemon too
<r0kk3rz>
is there a github issue for this?
mildred has quit [Ping timeout: 244 seconds]
<noffle>
There's also ipfs-embedded-shell for Go
<noffle>
which is an ephemeral node
pfista has quit [Ping timeout: 244 seconds]
Soft has quit [Quit: WeeChat 1.6-dev]
stoogenmeyer has joined #ipfs
chungy has quit [Ping timeout: 264 seconds]
stoogenmeyer has quit [Remote host closed the connection]
chungy has joined #ipfs
ola has joined #ipfs
<r0kk3rz>
ephemeral mode will also have rpc api? or can it only be daemonless
stoogenmeyer has joined #ipfs
stoogenmeyer has quit [Ping timeout: 252 seconds]
computerfreak has quit [Remote host closed the connection]
<Kubuxu>
no that is offline node
<Kubuxu>
but it is really only about internal naming
M-eternaleye has quit [Changing host]
M-eternaleye has joined #ipfs
M-eternaleye is now known as eternaleye
<lgierth>
Kubuxu: Lisbon Team Week Dates are July 13-21st with travel dates July 12th and July 22nd
dguttman has joined #ipfs
pfista has joined #ipfs
<Kubuxu>
So I should be in Berlin for at least half of their stay.
<lgierth>
i'm telling you because your very much invites :)
<Kubuxu>
Oh wait.
<Kubuxu>
It is in Lisbon right?
<lgierth>
ye
<Kubuxu>
:(
<Kubuxu>
I have to go to Poland 16-23 ...
<lgierth>
maybe stay in lisbon ~12th to 16th?
gm has joined #ipfs
gm is now known as Guest66138
seg_violation has joined #ipfs
j12t has quit [Remote host closed the connection]
Guest66138 has quit [Read error: Connection reset by peer]
<Kubuxu>
I will think about it but I can't promise anything.
<lgierth>
sure sure! juan will ask about it too i guess :)
<lgierth>
whyrusleeping: remember what the problem with the current writable gateway code is? (the stuff behind the Gateway.Writable option)
* lgierth
feels an itch on the backspace-finger
stoogenmeyer has joined #ipfs
stoogenmeyer has quit [Remote host closed the connection]
<lgierth>
the gateway extraction is gonna be so much gx yak shaving
<lgierth>
uses all of go-ipfs/{blocks,core,importer,merkledag,routing,path,unixfs}
j12t has joined #ipfs
<Kubuxu>
it is just that go-ipfs API is bad, IMO
<Kubuxu>
such simple code shouldn't plug in so many places, it just wants to get file in 99% of cases
<Kubuxu>
at least it is my view.
<lgierth>
yeah i think we can do it completely through the api
<lgierth>
or commands
<Kubuxu>
commands == API right now :p
<lgierth>
yeah i just mean not the http api
<lgierth>
but the underlying thing
Tv` has joined #ipfs
* Kubuxu
wonders how flights work, there are flights in same date, that take similar and differ in price by factor of 8...
dignifiedquire has quit [Quit: Connection closed for inactivity]
<kevina>
I need to run now but so far I have not seen anything that addresses my question " Without any sort of header how would you know what you are getting when you retrieve a block?"
computerfreak1 has joined #ipfs
<Kubuxu>
You don't.
<Kubuxu>
You need IPLD context to know that.
<Kubuxu>
If you try to read the block directly you just get "Unrecognized format"
Hory_ has quit [Read error: Connection reset by peer]
<kevina>
So what is the timeline before IPLD gets implemeted?
<Kubuxu>
Next few months for release, but don't quote me on this
gmcquillan has quit [Quit: gmcquillan]
computerfreak2 has joined #ipfs
gmcquillan has joined #ipfs
computerfreak3 has joined #ipfs
circa2050 has quit [Quit: leaving]
stoogenmeyer has quit [Remote host closed the connection]
<kevina>
We need to get back to this, at a later date as I need to go (Sorry), some of this could effect the filestore.
computerfreak1 has quit [Ping timeout: 244 seconds]
zorglub27 has quit [Ping timeout: 264 seconds]
computerfreak has quit [Ping timeout: 276 seconds]
<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to master: https://git.io/voBjF
ianopolous has quit [Remote host closed the connection]
palkeo has joined #ipfs
palkeo has quit [Changing host]
palkeo has joined #ipfs
stoogenmeyer has quit [Read error: Connection reset by peer]
stoogenmeyer has joined #ipfs
gmcquillan has quit [Quit: gmcquillan]
gmcquillan has joined #ipfs
gmcquillan has quit [Client Quit]
gmcquillan has joined #ipfs
gmcquillan has quit [Remote host closed the connection]
gmcquillan has joined #ipfs
ianopolous has joined #ipfs
ipfsrocks has quit [Ping timeout: 264 seconds]
<whyrusleeping>
ughhhh
<whyrusleeping>
i was working on a bug last night
<whyrusleeping>
and could reproduce it easily
<whyrusleeping>
like, in ten seconds
<whyrusleeping>
and now i'm coming up on an hour of it hiding from me
galois_d_ has joined #ipfs
galois___ has joined #ipfs
galois_dmz has quit [Ping timeout: 240 seconds]
galois_d_ has quit [Ping timeout: 258 seconds]
arpu has quit [Ping timeout: 252 seconds]
<xioustic>
Is it accurate to assume full IPLD implementation will be a hard fork of peer protocol as well?
j12t has joined #ipfs
<whyrusleeping>
xioustic: the peer protocol?
<whyrusleeping>
there is very unlikely to be a hard fork because of IPLD
<whyrusleeping>
we've taken some good precautions to make sure that transitioning to new protocols won't leave people out in the cold
livegnik has quit [Quit: No Ping reply in 180 seconds.]
livegnik has joined #ipfs
<ipfsbot>
[go-ipfs] whyrusleeping created fix/send-msg-race (+1 new commit): https://git.io/voRtm
<ipfsbot>
go-ipfs/fix/send-msg-race 63c8a9e Jeromy: pass reference to reader instead of using the one on the object...
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #2856: pass reference to reader instead of using the one on the object (master...fix/send-msg-race) https://git.io/voRt4
arpu has joined #ipfs
<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to master: https://git.io/voRtx
<Kubuxu>
Problem is that between cancel() and `select` the function sometimes finishes or becomes completely closed.
<Kubuxu>
even the `<-done` is sometimes successful
<whyrusleeping>
hrm....
<Kubuxu>
maybe I should just change the test to just require, closing and closed to be closed.
<whyrusleeping>
i think thats just a bad test...
<whyrusleeping>
honestly
<Kubuxu>
yeah...
<Kubuxu>
It was sometimes crashing with 100ms of delay :p
<Kubuxu>
Should I just delete it?
matoro has quit [Ping timeout: 276 seconds]
<whyrusleeping>
mmmm
<whyrusleeping>
i think its useful... but we should probably relax that check
<whyrusleeping>
also, i hate mondays
<whyrusleeping>
i'm so distracted today
<ipfsbot>
[go-ipfs] Kubuxu created feature/test-race-blockstore (+1 new commit): https://git.io/voRm0
<ipfsbot>
go-ipfs/feature/test-race-blockstore 7d2f943 Jakub Sztandera: Remove blockstore test with context...
<Kubuxu>
It is Wednesday.
<whyrusleeping>
uh
<whyrusleeping>
damn
<whyrusleeping>
youre right
rombou has quit [Ping timeout: 252 seconds]
Encrypt has quit [Quit: Quitte]
<Kubuxu>
I could also make the check literally check nothing by just checking all channels ::p
<Kubuxu>
s/checking/allowing
<ipfsbot>
[go-ipfs] whyrusleeping closed pull request #2822: Change logic of stdin input to be last and only if required (master...feature/stdin-stable) https://git.io/voLWo
<Kubuxu>
Excuse my English in that commit message, I don't know what happened.
ipfsrocks has joined #ipfs
<ipfsbot>
[go-ipfs] Kubuxu force-pushed feature/test-race-blockstore from 7d2f943 to 0a4d48c: https://git.io/voRYz
<ipfsbot>
go-ipfs/feature/test-race-blockstore 0a4d48c Jakub Sztandera: Remove blockstore test with context...
<voxelot>
noffle: i added the http tests which show the fail to js-ipfs PR 323, thanks for having a look!
<noffle>
voxelot: <3 !
stoogenmeyer has quit [Remote host closed the connection]
<whyrusleeping>
Kubuxu: lol
<whyrusleeping>
at least comment whats going on though
galois___ is now known as galois_dmz
<whyrusleeping>
maybe some intrepid contributor down the road will glance at it and say "oh, its so obvious, if we just do it this way it will all work"
stoogenmeyer has joined #ipfs
<Kubuxu>
Hmm, right.
neurrowcat has joined #ipfs
matoro has joined #ipfs
jaboja has joined #ipfs
kevina has quit [Quit: Leaving]
stoogenmeyer has quit [Ping timeout: 260 seconds]
stoogenmeyer has joined #ipfs
<whyrusleeping>
Kubuxu: also, i think we do need that timeout in the dht message code
<kevina>
I think I graph database would be cool. But a really good first step would be just to separate the metadata from the payload
step21_ is now known as step21
<kevina>
Storing the data separately from the links does affect block limit, if your concern is with storage. As the links will be stored in a different location from the data.
<ipfsbot>
[go-ipfs] Kubuxu opened pull request #2857: Remove failing blockstore with context test (master...feature/test-race-blockstore) https://git.io/voRcW
<Kubuxu>
It doesn't if we start using raw nodes
<kevina>
Kubuxu: so is this a sure thing now?
<Kubuxu>
almost
mgue has quit [Ping timeout: 260 seconds]
<Kubuxu>
also you have to remember that flat repo will be preferred repo for some use cases
<Kubuxu>
like for example archiving
ianopolous has quit [Ping timeout: 258 seconds]
<Kubuxu>
where data is lying in glacier storage
<whyrusleeping>
ansuz: i feel like cjdns isnt making me any more anonymoose than i was before
jaboja has quit [Read error: Connection reset by peer]
stoogenmeyer has quit [Remote host closed the connection]
<kevina>
Kubuxu: to be clear the move to using raw nodes to encode files larger than will fit in a single block is a sure thing?
<Kubuxu>
I don't know.
<kevina>
If it is not, than most of my points still hold.
<Kubuxu>
This doesn't change the fact that if we don't use raw nodes there will be some metadata before the data inside the block.
<Kubuxu>
Like multicodec ID.
<kevina>
Kubuxu: Not if the metadata is stored seperatly.
<whyrusleeping>
the metadata could also be stored in the links on the parent
<Kubuxu>
that is how raw nodes are supposed to work AFAIk
steefmin has quit [Ping timeout: 246 seconds]
<Kubuxu>
but otherwise we want to keep nodes self identifying, right
<whyrusleeping>
yeahhh... we havent really thought too hard on the raw nodes thing
stoogenmeyer has joined #ipfs
Oatmeal has joined #ipfs
<whyrusleeping>
but self identity of data is one of our core tenets
<whyrusleeping>
so we have to have really good reasons to depart from that
mgue has joined #ipfs
steefmin has joined #ipfs
<kevina>
whyrusleeping: do you understand what I am getting at?
<kevina>
What I am saying is that unless raw nodes are used it might make sense at some point in the future to _store_ the data seperatly from the metadata in repo.
<whyrusleeping>
kevina: right
<whyrusleeping>
that makes sense
<whyrusleeping>
its quite a bit more complicated
<whyrusleeping>
but for use-cases that demand it, its a great idea
<kevina>
Thanks, I think I have already solved most of the complexities in the filestore code.
<kevina>
And my main point is this separation should be considered when setting the block size.
<kevina>
Maybe the datasize could be capped at 256KiB-4KiB.
<kevina>
And the size of the block could be capped at 256KiB.
stoogenmeyer has quit [Remote host closed the connection]
<kevina>
So that no matter how the block is stored in the repo things are basically aligned on the drive.
<kevina>
With or without the metadata.
stoogenmeyer has joined #ipfs
<kevina>
(That is assuming a 4KiB block size on the drive.)
ipfsrocks has quit [Ping timeout: 252 seconds]
<ipfsbot>
[go-ipfs] whyrusleeping closed pull request #2856: pass reference to reader instead of using the one on the object (master...fix/send-msg-race) https://git.io/voRt4
erde74 has joined #ipfs
reit has quit [Ping timeout: 250 seconds]
<whyrusleeping>
234 nodes in the network at this moment
<whyrusleeping>
thats like, double what it was a couple weeks ago
Senji has quit [Disconnected by services]
Senj has joined #ipfs
Senj is now known as Senji
ansuz has quit [Ping timeout: 240 seconds]
_mak has joined #ipfs
<ipfsbot>
[go-ipfs] whyrusleeping created deps/libp2p/3.3.4 (+1 new commit): https://git.io/voR0i
<ipfsbot>
go-ipfs/deps/libp2p/3.3.4 de56a47 Jeromy: update go-libp2p to 3.3.4...
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #2858: update go-libp2p to 3.3.4 (master...deps/libp2p/3.3.4) https://git.io/voR0S