<alu>
one day we're all going to live inside the internet
amiller has joined #ipfs
Guest13955 is now known as amiller
ppham has joined #ipfs
ppham has quit [Client Quit]
<zignig>
how goes ipfs-ering ? I've been out of the loop as of late.
<zignig>
there seems to be a bit of an uptick after the npm debarkle ! ;)
<redfish>
re distro packaging (again): i don't mean to troll, just interested in packaging questions: since v1.5-1.6 Go supports building shared libs ("targeted at distro builders"), which is awesome news, will IPFS relax its build framework to support more classic packaging (e.g. not requiring dep code to hash-match)?
<baccenfutter>
jbenet: news from my gentoo dev mentor is that this will not work, whatsoever. if I build ebuilds for the deps, that pull from IPFS, then other ebuilds depending on those ebuilds will suddenly also use IPFS sources. however, another go tool my very well want a non-IPFS version of the dep. so this is a no-go (as they say it in #gentoo.de jolingly...) from our maintainer side, it would be highly desireable, that
<baccenfutter>
your provide two independant ways of installation. 1) go get 2) Makefle with absolutely no dependency magic via go get or other. what is your response to that?
<baccenfutter>
jokingly*
<baccenfutter>
our requirements are: 1) we want to build from source 2) we want to take care of dep management completely on our own with no voodoo `go get` in the background. every software component needs to be fetched, unpacked, configured, made and installed via portage.
<baccenfutter>
and to be completely honest, I think the gonuts community in geneneral has to get more aware of the fact that dependency magic is desireable by precisly one person only. the dev. not the admin, not the distribution, not the user. I understand this may not seem completely logic at first because dep-magic is soooo comfortable. my personal oppinion is: with gx you are trying to solve a problem that you should not
<baccenfutter>
be solving. you should be completely uncopiupled from distribution.
<baccenfutter>
but I do understand that this is something that comes from the nature of go and that they are actually trying to tackle exactly this problem with the entire `go get` thing. its just that their proposed solution sucks.
<baccenfutter>
which you obviously know, otherways you would not have created gx to pin things down.
eaxdigitalhash has joined #ipfs
a1uz10nn has joined #ipfs
a1uz10nn is now known as aluuu
<baccenfutter>
I am starting to think that I should perhaps just start committing to py-ipfs , help gat that done and then package that and still end up with lees hassle than trying to package go-ipfs xD
insanity54 has quit [Quit: No Ping reply in 210 seconds.]
insanity54 has joined #ipfs
<baccenfutter>
I'd hate see ipfs end up with a binary-blob distribution that - and we've seen this many times - will not spread due to strong beliefs and biases in the open-source community regarding binary blobs.
i_am_ss has joined #ipfs
reit has quit [Ping timeout: 265 seconds]
i_am_ss has quit [Ping timeout: 240 seconds]
cryptix has joined #ipfs
disgusting_wall has quit [Quit: Connection closed for inactivity]
reit has joined #ipfs
aluuu has quit [Quit: Bye]
eaxdigitalhash has quit [Remote host closed the connection]
ygrek has joined #ipfs
eaxdigitalhash has joined #ipfs
dignifiedquire has joined #ipfs
thufir has joined #ipfs
reincite has quit [Ping timeout: 260 seconds]
rajabose_ has quit [Remote host closed the connection]
rajabose has joined #ipfs
s_kunk has quit [Ping timeout: 252 seconds]
coremaking has joined #ipfs
Arakela007 has joined #ipfs
jinx_ has joined #ipfs
rendar has joined #ipfs
Arakela0_ has joined #ipfs
Arakela007 has quit [Read error: Connection reset by peer]
ecloud has quit [Ping timeout: 246 seconds]
rajabose has quit [Remote host closed the connection]
rajabose has joined #ipfs
hermanjunge has quit []
zorglub27 has joined #ipfs
hermanjunge has joined #ipfs
ecloud has joined #ipfs
jinx_ has quit [Quit: Hoopdity Shloopin!]
jinx_ has joined #ipfs
curlybracket has joined #ipfs
s_kunk has joined #ipfs
eaxdigitalhash has quit [Remote host closed the connection]
jinx_ has quit [Quit: Hoopdity Shloopin!]
jinx_ has joined #ipfs
rajabose_ has joined #ipfs
rajabos__ has joined #ipfs
i_am_ss has joined #ipfs
rajabose has quit [Ping timeout: 244 seconds]
rajabose_ has quit [Ping timeout: 252 seconds]
i_am_ss has quit [Ping timeout: 276 seconds]
zootella has joined #ipfs
lispmeister has joined #ipfs
zootella has left #ipfs [#ipfs]
ylp has quit [Ping timeout: 260 seconds]
Arakela007 has joined #ipfs
Arakela0_ has quit [Read error: Connection reset by peer]
zz_r04r is now known as r04r
ylp has joined #ipfs
hermanjunge has quit []
rmi7 has joined #ipfs
dignifiedquire has quit [Quit: Connection closed for inactivity]
<ipfsbot>
[js-ipfs-api] diasdavid force-pushed use-interface-ipfs-core from 11a75d1 to d5e4f46: https://git.io/vrq16
<ipfsbot>
js-ipfs-api/use-interface-ipfs-core d5e4f46 David Dias: use object tests from interface-ipfs-core
coremaking has quit [Ping timeout: 240 seconds]
eaxdigitalhash has joined #ipfs
jinx_ has quit [Quit: Hoopdity Shloopin!]
eaxdigitalhash has quit [Remote host closed the connection]
eaxdigitalhash has joined #ipfs
i_am_ss has joined #ipfs
eaxdigit_ has joined #ipfs
rajabose has joined #ipfs
i_am_ss has quit [Ping timeout: 265 seconds]
zootella has joined #ipfs
Encrypt has joined #ipfs
Arakela007 has quit [Read error: Connection reset by peer]
eaxdigitalhash has quit [Ping timeout: 276 seconds]
Arakela007 has joined #ipfs
rajabos__ has quit [Ping timeout: 276 seconds]
eaxdigit_ has quit [Remote host closed the connection]
<ipfsbot>
[js-ipfs-api] diasdavid force-pushed use-interface-ipfs-core from d5e4f46 to a3917f8: https://git.io/vrq16
<ipfsbot>
js-ipfs-api/use-interface-ipfs-core a3917f8 David Dias: use object tests from interface-ipfs-core
eaxdigitalhash has joined #ipfs
eaxdigitalhash has quit [Remote host closed the connection]
zootella has quit [Quit: zootella]
zootella has joined #ipfs
zootella has quit [Client Quit]
ylp has quit [Ping timeout: 260 seconds]
ylp has joined #ipfs
insanity54 has quit [Ping timeout: 252 seconds]
HastaJun_ has joined #ipfs
HastaJun has quit [Ping timeout: 276 seconds]
HastaJun_ is now known as HastaJun
zorglub27 has quit [Ping timeout: 276 seconds]
leer10 has quit [Ping timeout: 260 seconds]
dignifiedquire has joined #ipfs
rajabose has quit [Remote host closed the connection]
lispmeister has quit [Ping timeout: 240 seconds]
zootella has joined #ipfs
lispmeister has joined #ipfs
zootella has quit [Client Quit]
jinx_ has joined #ipfs
lactinate has joined #ipfs
zootella has joined #ipfs
zootella has quit [Client Quit]
zootella has joined #ipfs
eaxdigitalhash has joined #ipfs
zootella has quit [Client Quit]
rajabose has joined #ipfs
<ipfsbot>
[js-ipfs-api] diasdavid force-pushed use-interface-ipfs-core from a3917f8 to 88e923e: https://git.io/vrq16
<ipfsbot>
js-ipfs-api/use-interface-ipfs-core 88e923e David Dias: use object tests from interface-ipfs-core
zorglub27 has joined #ipfs
Oatmeal has quit [Ping timeout: 276 seconds]
maxlath has joined #ipfs
zorglub27 has quit [Ping timeout: 246 seconds]
maxlath is now known as zorglub27
<daviddias>
whyrusleeping: do you have any test to verify that yaml encoding is still working?
<daviddias>
getting an error coming from go-ipfs when I pass a yaml encoded object through js-ipfs api
<daviddias>
this object.api refactoring to the core spec is taking longer than initially expected, requires a lot of tests to make sure everything is ok. Just missing the object.patch methods now :)
zootella has joined #ipfs
<dignifiedquire>
daviddias: :) I'm also here now, really want to ship this in js-ipfs today, so I will start pulling in the tests you already have
<dignifiedquire>
let me know if you need anything from me
<daviddias>
I've one thing to share which is kind of annoying on the http-api itself
zootella has quit [Client Quit]
<daviddias>
essentially the fact that it converts everything to strings
<dignifiedquire>
not sure what you mean
<daviddias>
when you get an object
<daviddias>
the Data field
<daviddias>
is always a string
<daviddias>
so, if you put an object, with a Data field being `new Buffer('hey ho')`
Guest98064 has quit [Changing host]
Guest98064 has joined #ipfs
<daviddias>
it is all good cause Buffer <-> String have one to one translation
<dignifiedquire>
right because http only does strings
<dignifiedquire>
dates as well?
<daviddias>
but complex types inside Buffers, not so sure
Guest98064 is now known as brimstone
computerfreak has joined #ipfs
zorglub27 has quit [Ping timeout: 244 seconds]
zorglub27 has joined #ipfs
rmi7 has quit [Ping timeout: 260 seconds]
<dignifiedquire>
right, but you have to remember object.put is very low level, so people should make sure their serilization format is good enough for them so that they can retrieve their values
Oatmeal has joined #ipfs
<dignifiedquire>
and we just store the raw thing for them, and when they retrieve it they will have their own methods for getting back types if they need it
<daviddias>
got it
<daviddias>
but since the API always expected a serialized json (or yaml)
<daviddias>
in order to let users use Buffer as the Data field
<daviddias>
I need to stringified before sending it
<daviddias>
because a JSON.stringify() of an object with a Buffer
<daviddias>
will convert that Buffer to {type: Buffer, data: [...]}
<daviddias>
And what I'm doing, is when I get the result of put
<daviddias>
I check the multihash to make sure that the stored thing was the correct one
<daviddias>
but I return a DAGNode with the Data field from the one that was the input of the function
<daviddias>
so that everything matches
<daviddias>
it works, it is just annoying and the http-api could make this seemless
<daviddias>
btw, did you had any luck/time oon getting a core api + tests for the add, cat, get and ls ?
Encrypt has quit [Quit: Quitte]
lactinate has quit [Remote host closed the connection]
mdangear has joined #ipfs
dvn has quit [Ping timeout: 264 seconds]
angreifer has quit [Ping timeout: 264 seconds]
dvn has joined #ipfs
mrpoopybuttwhole has quit [Ping timeout: 264 seconds]
Arakela007 has quit [Read error: Connection reset by peer]
mrpoopybuttwhole has joined #ipfs
Arakela007 has joined #ipfs
The_8472 has quit [Ping timeout: 264 seconds]
bauruine has quit [Ping timeout: 264 seconds]
vladimir-sn has quit [Ping timeout: 264 seconds]
vladimir-sn_ has joined #ipfs
The_8472 has joined #ipfs
<dignifiedquire>
not yet, but voxelot made a pr for js-ipfs-api for get
Akaibu has joined #ipfs
eaxdigitalhash has quit [Remote host closed the connection]
zootella has joined #ipfs
chromatogenous has joined #ipfs
M-DrKrieger has joined #ipfs
bauruine has joined #ipfs
M-DrKrieger has left #ipfs [#ipfs]
eaxdigitalhash has joined #ipfs
maxlath has joined #ipfs
chromatogenous has quit [Quit: Disconnecting]
zorglub27 has quit [Ping timeout: 276 seconds]
maxlath is now known as zorglub27
zootella has quit [Quit: zootella]
Arakela0_ has joined #ipfs
Arakela007 has quit [Read error: Connection reset by peer]
eaxdigitalhash has quit [Read error: Connection reset by peer]
eaxdigitalhash has joined #ipfs
rgrinberg has joined #ipfs
eaxdigitalhash has quit [Remote host closed the connection]
underroof has joined #ipfs
zootella has joined #ipfs
angreifer has joined #ipfs
mdangear has quit [Quit: mdangear]
zootella has quit [Quit: zootella]
<dignifiedquire>
daviddias: I got all your object tests passing on js-ipfs core :)
<daviddias>
dignifiedquire: sweet :D
rmi7 has joined #ipfs
<daviddias>
the promise ones as well?
mdangear has joined #ipfs
<dignifiedquire>
yes
PrinceOfPeeves has joined #ipfs
herzmeister has quit [Quit: Leaving]
herzmeister has joined #ipfs
lmatteis has quit [Ping timeout: 260 seconds]
taw00 has joined #ipfs
taw00 has quit [Remote host closed the connection]
<ipfsbot>
[js-ipfs-api] diasdavid force-pushed use-interface-ipfs-core from 50bc13a to 80218b6: https://git.io/vrq16
<ipfsbot>
js-ipfs-api/use-interface-ipfs-core 80218b6 David Dias: use object tests from interface-ipfs-core
<daviddias>
it's the test. Thank you for catching that
cemerick has joined #ipfs
leer10 has joined #ipfs
<dignifiedquire>
daviddias: the fix you made is still not right
<dignifiedquire>
obj.Data is a a String
<dignifiedquire>
but node.data is a Buffer
computerfreak has quit [Quit: Leaving.]
<daviddias>
this is one of the annoying things of not being able to send Buffers as is, to the API
corvinux has joined #ipfs
corvinux has joined #ipfs
<voxelot>
dignifiedquire: daviddias: going to finish up the suggestions on js-ipfs-api get and then try to finish any last small glitches for js-ipfs http add(still can't find right ndjson return format), cat (works), get (still need implementation + tar)
<daviddias>
because indeed that node.data is a String too
<daviddias>
voxelot hey o/
<daviddias>
that sounds like good stuff :)
herzmeister has quit [Quit: Leaving]
<dignifiedquire>
daviddias: hmm
herzmeister has joined #ipfs
<dignifiedquire>
daviddias: it's not for me in js-ipfs
<cehteh>
mhm data on ipfs dont have a ttl .. couldnt that be handy in some cases (by default +inf) .. but that would make streaming and other temporary data possible, and deleting expired data should just handled voluntary by the gc?
mildred has joined #ipfs
ZaZ has quit [Read error: Connection reset by peer]
vijayee has joined #ipfs
s_kunk has quit [Ping timeout: 240 seconds]
<richardlitt>
So, I am satisfied with the current state of the standard-readme specification. Would be great if you all had any feedback on it (@daviddias @dignifiedquire @whyrusleeping @lgierth @noffle @voxelot @haad_
<Kubuxu>
cehteh: smarter garbage collecting would solve it, and it is something I am thinking about working on.
<Kubuxu>
also livestream could turn into VOD without a problem in ipfs world.
<noffle>
richardlitt: cool! I responded to your email btw
<richardlitt>
noffle: Sweet!
<noffle>
richardlitt: you state your goals twice fyi -- once above the TOC and once below (not sure if intentional)
<richardlitt>
noffle: Is intentional.
<richardlitt>
Guess I could change that.
libman has joined #ipfs
<noffle>
overall nice stuff -- good background section, aggressive linkification, very detailed
<noffle>
it'd be nice to see an example on a code-based project. I faced the same thing for common-readme; human document projects tend to structure differently than the code projects that have apis, etc
<richardlitt>
Well, go-ipfs is almost compliant
<richardlitt>
For instance.
<noffle>
nice; that'd be a great high profile project to get fully compliant and show off :)
<noffle>
seems like a lot of work -- aren't most already pretty close?
<richardlitt>
No.
<richardlitt>
Not even close.
<richardlitt>
And that's good work. That's why I am doing this. People shouldn't have to think about this stuff.
<richardlitt>
They should be able to say: My Readme is good. It has what it needs. Ok, what about the code?
<richardlitt>
be back, heading home.
cryptix has joined #ipfs
ashark has quit [Ping timeout: 276 seconds]
nskelsey has quit [Remote host closed the connection]
hosh has quit [Remote host closed the connection]
true_droid has quit [Remote host closed the connection]
cdata has quit [Remote host closed the connection]
zmanian__ has quit [Remote host closed the connection]
ELLIOTTCABLE has quit [Read error: Connection reset by peer]
xicombd has quit [Remote host closed the connection]
bigbluehat has quit [Write error: Broken pipe]
grncdr has quit [Remote host closed the connection]
micxjo has quit [Remote host closed the connection]
<cehteh>
Kubuxu: a ttl (aka with absolute time here) could be a hint for the gc, to make the gc smart it needs some hints
<Kubuxu>
gc should work on basis on how frequently users uses given file.
<Kubuxu>
If you set short TTL on hash but I used that picture as my wallpaper (directly from IPFS) it doesn't matter.
bigbluehat has joined #ipfs
<cehteh>
that would be the non-smart way, but if it can collect things anticipatory it would make it smarter
ehd has joined #ipfs
ELLIOTTCABLE has joined #ipfs
<cehteh>
"even if i used this thing 10 time today, hey i am not going to use it anymore"
<cehteh>
that would improve messaging and streaming
hosh has joined #ipfs
ashark has joined #ipfs
libman has quit [Ping timeout: 250 seconds]
jinx_ has quit [Quit: Hoopdity Shloopin!]
insanity54 has quit [Ping timeout: 276 seconds]
jinx_ has joined #ipfs
<Kubuxu>
But you can't know if user is not going to use it anymore.
Encrypt has joined #ipfs
gaboose_ has joined #ipfs
Gaboose has joined #ipfs
<Kubuxu>
Maybe he will share this stream with his friend as a VOD. User individually could tag something, but tagging at a source would only cause mess.
<Kubuxu>
Also storage is cheap, transfer is expensive. Perfectly the cache should be almost always at 100% of allowed space usage, but it might be hard to pull off.
s_kunk has joined #ipfs
mildred has quit [Read error: Connection reset by peer]
<cehteh>
Kubuxu: yeah its only meant as hint, the user may remove/alter/ignore it on his discretion
libman has joined #ipfs
zmanian__ has joined #ipfs
<cehteh>
technically its not required, the gc will prune things eventually
true_droid has joined #ipfs
Oatmeal has quit [Read error: Connection reset by peer]
r0kk3rz has joined #ipfs
<Kubuxu>
Problem is: metadata is not yet allowed (IPLD might change it). Then something like "expected lifetime" might have place but it would be only really small tip for the GC.
<cehteh>
but if it can prune blocks earlier than they expire by 'normal' policies then there is more cache space availabel for non hinted or rarely used things
mildred has joined #ipfs
grncdr has joined #ipfs
<cehteh>
of course
nskelsey has joined #ipfs
<cehteh>
this is only a idea
<cehteh>
think about one starts some massively used messagins/streaming platform .. its contents will easily displace other infrequently (but meant to be permanent) blocks
<redfish>
hi, is this normal for gateway.ipfs.io: first load in browser of a hash i pinned on my node works just fine, but reloading in the browser almost never works: "waiting for..." indefinitely
<redfish>
no such problem with my node's gateway. Also tried by ip, not via gateway.ipfs.io dns: same thing.
<whyrusleeping>
redfish: i've seen that too, i think i've found the problem and should be pushing a fix soon
<whyrusleeping>
the gateway nodes get hung up talking to eachother
<redfish>
oh, thanks in advance for pushing the fix.
ashark has quit [Read error: Connection reset by peer]
zorglub27 has quit [Ping timeout: 276 seconds]
Encrypt has quit [Quit: Quitte]
Encrypt has joined #ipfs
xicombd has joined #ipfs
ashark has joined #ipfs
<noffle>
whyrusleeping: ha I'm still getting Medium notifications for that one time I linked to gx during the leftpad incident
libman has quit [Remote host closed the connection]
<whyrusleeping>
noffle: niceeeee
<whyrusleeping>
theres still a steady stream of traffic coming in to it
<noffle>
is there a nice way for users who don't have access to pinbot to ensure their packages stay available? it'd be nice to guarantee availability with an ipfs-run automatic mirror
micxjo has joined #ipfs
cdata has joined #ipfs
<whyrusleeping>
mars.i.ipfs.team:9444
<whyrusleeping>
noffle: o/
<noffle>
awesome -- gx publish automatically replicates to this?
<noffle>
ah, it's opt in
<noffle>
I wonder if there's a way to get a github webhook when someone pushes to a gx package repo
<noffle>
automatic passive pinning would be super cool
zootella has quit [Quit: zootella]
jinx_ has quit [Ping timeout: 260 seconds]
<whyrusleeping>
yeah, we could probably do that
<whyrusleeping>
set up that server to accept hooks from github
<whyrusleeping>
then provide text for users to copy/paste and add a hook
<ipfsbot>
[go-ipfs] Kubuxu opened pull request #2687: Increase timeout in t0060 (master...fix/t0060) https://git.io/vrOdK
zootella has joined #ipfs
zootella has quit [Client Quit]
Oatmeal has joined #ipfs
<noffle>
whyrusleeping: it'd be great if this was a passive service that just picked up the gx repos automatically -- no manual opt-in, but free pinning
<whyrusleeping>
eh... completely passive is against a lot of what gx stands for
<whyrusleeping>
that means we're making assumptions about things for people
<whyrusleeping>
but i think some easy opt in would be sensible
<whyrusleeping>
a 'one click' type thing
<whyrusleeping>
add a post publish hook to gx that does an http post to the gexpin server
<whyrusleeping>
except post-publish wouldnt really work because its not on GH yet
<noffle>
one of the best things about ipfs is that you can have passive hosting: others can mirror your content without your knowledge and to your benefit. your data becomes more reliably available without any coupling
<ipfsbot>
[js-ipfs-api] diasdavid force-pushed use-interface-ipfs-core from 7852050 to 88a31d0: https://git.io/vrq16
<ipfsbot>
js-ipfs-api/use-interface-ipfs-core 88a31d0 David Dias: use object tests from interface-ipfs-core
computerfreak has quit [Quit: Leaving.]
computerfreak has joined #ipfs
charley_ has joined #ipfs
zorglub27 has joined #ipfs
zorglub27 has quit [Ping timeout: 260 seconds]
<richardlitt>
in `ipfs name resolve ipfs.io`, what is the technical name for `ipfs.io`?
<richardlitt>
It is a name, still, but it's not a hash.
<daviddias>
dignifiedquire: no more CR on the js-ipfs-api?
<daviddias>
haad_: do you want to check it?
<dignifiedquire>
daviddias not on my computer right now
Encrypt has quit [Quit: Quitte]
Encrypt has joined #ipfs
<whyrusleeping>
richardlitt: its an ipns path
<whyrusleeping>
its actually inferred to be /ipns/ipfs.io
charley_ has quit [Read error: Connection reset by peer]
<whyrusleeping>
since youre using it in the context of an ipns command
<richardlitt>
There's no terminology difference between a mutable and an immutable name?
<whyrusleeping>
nope
<whyrusleeping>
theyre just paths
charley_ has joined #ipfs
<richardlitt>
not really.
<richardlitt>
it's a simlink, isn't it?
<whyrusleeping>
/ab/c/d/e/fg is a path
<whyrusleeping>
whether or not c,d,e are symlinks or directories
<richardlitt>
yeah, but /ipfs/ipfs.io is linked to ipfs/somehash
<whyrusleeping>
and /ab/c/d/e/fg is linked to /z/x/y
<whyrusleeping>
still paths
<richardlitt>
Yeah. What I'm interested is, in the path /a/b/c, what is c called if it is a hash, and what is it called if it isn't.
<richardlitt>
I understand that /a/b/c is a path.
<richardlitt>
Does my question make sense?
<computerfreak>
guys, can u update subfolder of ipns / without updating the hole path?
<richardlitt>
There should be a verbal abstraction for the final node in a path
<whyrusleeping>
i'd call it a reference
<whyrusleeping>
computerfreak: nope, updating a subfolder changes the hash
<computerfreak>
well but ipns isnt a hash that represents a file right?
<computerfreak>
its a pubke
<computerfreak>
key*
<ipfsbot>
[go-ipfs] RichardLitt created feature/2617-update-resolve (+1 new commit): https://git.io/vr3JB
<ipfsbot>
go-ipfs/feature/2617-update-resolve bb4a4ae Richard Littauer: Updated resolve hash to Mars, added reference...
<ipfsbot>
[go-ipfs] RichardLitt force-pushed feature/2617-update-resolve from bb4a4ae to b356089: https://git.io/vr3Jz
<ipfsbot>
go-ipfs/feature/2617-update-resolve b356089 Richard Littauer: Updated resolve hash to Mars, added reference...
<ipfsbot>
[go-ipfs] RichardLitt opened pull request #2688: Updated resolve hash to Mars, added reference (master...feature/2617-update-resolve) https://git.io/vr3JV
reit has quit [Quit: Leaving]
reit has joined #ipfs
charley_ has quit [Remote host closed the connection]
charley_ has joined #ipfs
<dignifiedquire>
daviddias: not sure about the layout stuff, do you thinks it makes sense or not?
<dignifiedquire>
daviddias: can you please review my pr for interface-ipfs-core?
HostFat has quit [Ping timeout: 276 seconds]
<daviddias>
dignifiedquire: I feel it is weird to have unixfs mixed with object api
<daviddias>
and I don't see any particular use case for it
<daviddias>
specially when it has never been fully supported
<daviddias>
on the specific object.new
charley_ has quit [Remote host closed the connection]
jaboja has joined #ipfs
<dignifiedquire>
daviddias: sounds good
<ipfsbot>
[js-ipfs-api] diasdavid force-pushed use-interface-ipfs-core from 88a31d0 to 2000d22: https://git.io/vrq16
<ipfsbot>
js-ipfs-api/use-interface-ipfs-core 2000d22 David Dias: use object tests from interface-ipfs-core
libman has joined #ipfs
charley_ has joined #ipfs
<dignifiedquire>
daviddias: will remove it from core then
<daviddias>
thank you :)
Encrypt has quit [Quit: Quitte]
computerfreak has quit [Quit: Leaving.]
gaboose_ has quit [Quit: Connection closed for inactivity]
reit has quit [Ping timeout: 276 seconds]
insanity54 has joined #ipfs
jaboja has quit [Ping timeout: 240 seconds]
<ipfsbot>
[js-ipfs] dignifiedquire pushed 1 new commit to core/object: https://git.io/vr3qC
<ipfsbot>
js-ipfs/core/object 2c276f7 Friedel Ziegelmayer: remove layout option from object.new
<dignifiedquire>
daviddias: I'm thinking about implementing the protobuf option for object.put and in go-ipfs the proto file for a dagnode is in "merkledag", so would make sense if we put this in js-ipfs-merkle-dag, what do you think?
<dignifiedquire>
daviddias: so close with the object stuff
<dignifiedquire>
we might actually finish it today :)
<daviddias>
we are still on time :D
<daviddias>
did you check my comment on the PR
<dignifiedquire>
yeah just saw it
<dignifiedquire>
how did you check?
<dignifiedquire>
with js-ipfs-api?
<daviddias>
yep
<daviddias>
running it on js-ipfs-api
<daviddias>
I get that error as a response for go-ipfs
Encrypt has joined #ipfs
<dignifiedquire>
master or sth else?
<daviddias>
master
<dignifiedquire>
checking
<dignifiedquire>
daviddias: hmm
<dignifiedquire>
daviddias: there is no test that .marshal() is compatible with go-ipfs, only unMarshal
<dignifiedquire>
daviddias: go-ipfs is broken
<daviddias>
go-ipfs?
<dignifiedquire>
yep
<dignifiedquire>
posting repro now
<daviddias>
(I was going to say that you also have to account the unixfs + unixfs-engine and how if we import a file, it creates the matching merkledag nodes and we can move them around with go-ipfs without a problem)
<daviddias>
dignifiedquire: you mean, the http-api on go-ipfs?
<whyrusleeping>
daviddias: not sure if you saw my message earlier, but go-ipfs doesnt have any kind of yaml encoding support
<sdgathman>
Why is that needed?
<daviddias>
whyrusleeping: sorry, I missed that one, but then later we checked through the CLI (and I don't know where I got the idea it supported yaml)
<whyrusleeping>
sdgathman: its not, lol
<dignifiedquire>
daviddias: nvm it was my mistake
<daviddias>
ok. So what did you learn?
<ipfsbot>
[go-ipfs] whyrusleeping created deps/libp2p/muxers (+1 new commit): https://git.io/vr3Cb
<ipfsbot>
go-ipfs/deps/libp2p/muxers a657ccf Jeromy: update deps to introduce yamux hang fix...
* sdgathman
likes yaml for human editing, but json is prly easier to parse
<Kubuxu>
daviddias: I think jbene.t was using it for some examples in IPLD discussions.
zootella has quit [Quit: zootella]
<whyrusleeping>
sdgathman: yeah, yaml is nicer than json for a lot of usecases. but the part we're talking about is the http api protocol
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #2689: update deps to introduce yamux hang fix (master...deps/libp2p/muxers) https://git.io/vr3Wf
Arakela0_ has joined #ipfs
Arakela007 has quit [Read error: Connection reset by peer]
<noffle>
whyrusleeping: it's not always true that mfs files have 0 links, correct?
<whyrusleeping>
heh?
<noffle>
dag nodes representing mfs files
<daviddias>
noffle: files are Shared
<whyrusleeping>
are just just like normal dag nodes representing normal files?
<noffle>
right
<whyrusleeping>
and mfs file is the same as one created with add
<daviddias>
Only if the file is less than a chunk size
<noffle>
mfs code is assuming that 0 links => file and > 0 links => directory
<noffle>
is why I ask
<whyrusleeping>
ah, well thats wrong
<whyrusleeping>
it should check the type in the data firled
<whyrusleeping>
field*
<noffle>
whyrusleeping: this is for adding a dag node to mfs, which has no type (yet)
<dignifiedquire>
daviddias: it's a bug in js-ipfs-api you are passing the encoding wrong to the api
<whyrusleeping>
noffle: it should
<whyrusleeping>
wheres this code?
<daviddias>
What is the right way ?
<noffle>
hamfs/dir.go
<noffle>
mfs/dir.go
<daviddias>
(I reused that part from the previous code, assumed it was good)
<noffle>
AddChild
<noffle>
pretty sure this is the cause of the 1st bug you sent me
<ipfsbot>
[js-ipfs-api] dignifiedquire created object-proto (+1 new commit): https://git.io/vr3lR
<ipfsbot>
js-ipfs-api/object-proto 19cd3c5 Friedel Ziegelmayer: fix: add support for protobuf in object.put
<noffle>
anyways, I ask because Dag.Node has no inherent notion of file/dir, so I'm trying to deduce the best way to decide if it ought to be interpreted as a file or a dir
<ipfsbot>
[js-ipfs-api] dignifiedquire opened pull request #272: fix: add support for protobuf in object.put (master...object-proto) https://git.io/vr3la
<ipfsbot>
[js-ipfs-api] dignifiedquire pushed 1 new commit to object-proto: https://git.io/vr3lM
<ipfsbot>
[js-ipfs-api] diasdavid tagged v4.0.0 at 41804af: https://git.io/vr3Bn
<noffle>
the difference is that the top namespace exists *outside* of the files concept
<daviddias>
dignifiedquire: haad_ released 4.0.0
<noffle>
so logic can operate on files at / blindly, and the higher level logic considers the namespace
<noffle>
it's a big difference imo
<whyrusleeping>
eh... apps should know to manage their own namespaces
<whyrusleeping>
apps on linux are generally pretty good about coexisting in your homedir, and .config and /etc/
insanity54 has quit [Ping timeout: 260 seconds]
<noffle>
whyrusleeping: I think the problem is in communication -- it's unclear that all apps will share the same file system on that ipfs daemon
<noffle>
e.g. ipfs-blob-store doesn't tell you it's all shared, so users won't assume it is
charley_ has joined #ipfs
<noffle>
exposing that notion at a higher level makes it very explicit
<daviddias>
how is that different from all the apps we run in our machines today?
<daviddias>
nevertheless, I agree with the feature
charley_ has quit [Read error: Connection reset by peer]
<daviddias>
and if I'm not mistaken, that was one of the motivations behind chroot
gmcquillan has quit [Quit: gmcquillan]
charley_ has joined #ipfs
<ipfsbot>
[js-ipfs] diasdavid created greenkeeper-ipfs-api-4.0.0 (+1 new commit): https://git.io/vr3B7
<ipfsbot>
js-ipfs/greenkeeper-ipfs-api-4.0.0 f668690 greenkeeperio-bot: chore(package): update ipfs-api to version 4.0.0...
insanity54 has joined #ipfs
cemerick has quit [Ping timeout: 240 seconds]
<noffle>
daviddias: the problem is communication -- it's not obvious that you're sharing a root dir with every other app. all the files api examples have you making files right in / -- no unix tutorial would have you doing that; you'd make a directory first and not pollute /
<noffle>
and no other part of ipfs exposes a shared mutable area -- it's not a consistent pattern over the tool
<noffle>
so special care must be taken
<ipfsbot>
[js-ipfs-api] dignifiedquire deleted greenkeeper-interface-ipfs-core-0.1.5 at 60694f6: https://git.io/vr3Rc
Arakela0_ has quit [Read error: Connection reset by peer]
Arakela007 has joined #ipfs
<noffle>
it's not obvious that the whole fs is shared state over all daemon users
<whyrusleeping>
sounds like we just need to work on documentation then
<whyrusleeping>
and not introduce weird extra namespacing logic to make everything harder for everyone
<noffle>
whyrusleeping: +1 for trying to increase communication and not jumping into a complex solution /wo verifying it's a problem
<noffle>
but I suspect it will be as more users use it
paralipsis has quit [Remote host closed the connection]
gmcquillan has joined #ipfs
<noffle>
it's kind of weird for the daemon to expose any kind of shared state, actually
<noffle>
everything else is about operating on a global namespace of /ipfs and /ipns
<noffle>
and the daemon is a cache for that
gmcquillan has quit [Read error: Connection reset by peer]
<noffle>
but an ipfs files root is a purely local notion to that repo
Arakela007 has quit [Remote host closed the connection]
thufir has left #ipfs [#ipfs]
<ipfsbot>
[go-ipfs] whyrusleeping closed pull request #2666: Capitalize print statements, add period to `external` (master...feature/external-clean-up) https://git.io/vrLWd
<noffle>
before going into this: do you agree that shared mutable state across the daemon can be problematic? agreeing on the problem matters before discussing potential solutions
<richardlitt>
ahhhhhhhhh
<ipfsbot>
[go-ipfs] whyrusleeping deleted feature/mount_unix_long_help at fedf553: https://git.io/vr3gX
<ipfsbot>
[go-ipfs] whyrusleeping closed pull request #2651: Add bin to gitignore (master...feature/gitignore-bin) https://git.io/vrTq5
<richardlitt>
did not expect all of these notifs
<richardlitt>
This is what I get for coding for a couple of hours on something else.