<Kubuxu>
as I don't know if it is my frakup or what
<whyrusleeping>
is there anything in the output of 'mount'
<Kubuxu>
just that /ipfs /ipns were mounted
<whyrusleeping>
does it show anything after you doa systemctl stop?
<whyrusleeping>
also, what systemctl script are you using?
corvinux has quit [Remote host closed the connection]
<Kubuxu>
From logs it says that it unmounts /ipfs /ipns mount: ERROR: (c *closer) Close(),/ipfs mount.go:91 but from code it is just information about unmounting
<whyrusleeping>
okay
<whyrusleeping>
what ips version?
<whyrusleeping>
ipfs*
anticore has joined #ipfs
trenchmaster has joined #ipfs
<Kubuxu>
build from master branch
<whyrusleeping>
okay
<whyrusleeping>
one sec, gotta stash the code i'm working on and switch to a different daemon
Encrypt has quit [Quit: Quitte]
<Kubuxu>
ok thanks
anticore has quit [Client Quit]
anticore has joined #ipfs
trenchmaster is now known as instancy
anticore has quit [Client Quit]
anticore has joined #ipfs
<whyrusleeping>
Kubuxu: what systemctl script are you using?
<whyrusleeping>
if its just once, it should be fine
instancy has quit [Ping timeout: 256 seconds]
<whyrusleeping>
but if it sends it more than once before the daemon can safely shutdown then it probably isnt able to complete the unmount
<Kubuxu>
SIGTERM+SIGCONT (to wake up suspended processes), waits 90s and sends SIGKILL but it took less than 90s to restart so it did not send SIGKILL .
<Kubuxu>
Now I can't even recover it buy trying to write into it.
<whyrusleeping>
i'm having trouble reproducing
<Kubuxu>
and now it works
<Kubuxu>
woot
<whyrusleeping>
lol, what did you do?
<Kubuxu>
nothing
<Kubuxu>
I published different hash to it.
<Kubuxu>
but previousy it didn't change anything
<whyrusleeping>
you ran publish while the fuse stuff was mounted?
<Kubuxu>
unmounted, run publish, remounted
<whyrusleeping>
ah, good
<Kubuxu>
and it works with has that didn't work
<Kubuxu>
IDK
<Kubuxu>
Ok, got it again
<ion>
whyrusleeping: Have you seen a doctor about that?
<ansuz>
happy new year ipfs community
<Kubuxu>
whyrusleeping: I might have tracked it down and fixed it. rechecking
<Kubuxu>
nope
ralphtheninja has joined #ipfs
<whyrusleeping>
ion: lol, watch it
<whyrusleeping>
coffee in germany is harder to find
<whyrusleeping>
i'm not as awake as i'm used to
<whyrusleeping>
ansuz: happy new year :)
<whyrusleeping>
Kubuxu: you could try enabling the fuse debug logs
grahamperrin has left #ipfs ["Leaving"]
<Kubuxu>
I could if I knew how.
<whyrusleeping>
in fuse/ipns/ipns_unix.go, right under the var log = logging... line
<yangwao>
daviddias: we've been near small bridge with CS folks, and during fireworks there was ambulance trying drive through crowd, was kinda paradox, nobody pay attention during fireworks noise :)
<Kubuxu>
Roadmap: 1. Port hastebin to display content from IPFS and use resources only from IPFS. 2. Integrate it with crypto 3. make CLI from adding files from shell. 4. work on Web interface
<Kubuxu>
Hastebin made it perfect. It is fully static site :D
amiller has quit [Ping timeout: 255 seconds]
voxelot has quit [Read error: No route to host]
Guest97248 has joined #ipfs
maxlath has joined #ipfs
travis__ has joined #ipfs
Looking has joined #ipfs
Matoro_ has quit [Remote host closed the connection]
<brimstone>
That's cool, but can you upload from the ipfs only one?
<Kubuxu>
No, I will work on it next. For now it will require local node and allowing in CORS.
<Kubuxu>
In future gateway will allow for upload and then you have to use some pinning service.
<voxelot>
11:39:11.697 ERROR commands/h: multipart: NextPart: multipart: Part Read: multipart: Part Read: malformed chunked encoding client.go:272
<voxelot>
random recursive add fail, couldn't reproduce
Senji has joined #ipfs
diffalot has quit [Remote host closed the connection]
Senji has quit [Ping timeout: 255 seconds]
diffalot has joined #ipfs
diffalot has quit [Remote host closed the connection]
asyncsrc has joined #ipfs
diffalot has joined #ipfs
Senji has joined #ipfs
diffalot has quit [Remote host closed the connection]
diffalot has joined #ipfs
diffalot has quit [Changing host]
diffalot has joined #ipfs
diffalot has quit [Remote host closed the connection]
jaboja has joined #ipfs
diffalot has joined #ipfs
CarpeDiem1 has left #ipfs [#ipfs]
jaboja has quit [Read error: Connection reset by peer]
silwol2 has joined #ipfs
leer10 has joined #ipfs
jaboja has joined #ipfs
anticore has joined #ipfs
jaboja has quit [Read error: Connection reset by peer]
Senji has quit [Ping timeout: 240 seconds]
<NightRa>
Kubuxu: Is it in some specification yet? what's hashed, what's the object structure and in what enc.
jaboja has joined #ipfs
anticore has quit [Quit: bye]
anticore has joined #ipfs
<Kubuxu>
NightRa: It will be specified when IPLD comes into play, until that, code is the spec in this matter AFAIK.
<NightRa>
Kubuxu: Understood.
<NightRa>
I'm sad it's not directly the content's hash. ehh..
<Kubuxu>
it might be in future
<NightRa>
Because it makes alot of sense to use it as an inverse hash function to get content from the internet based on the content's hash itself
<NightRa>
And it's easier to implement from the outside
<NightRa>
Actually, if it's just some prefix and suffix, it's easy enough to implement from the outside. As SHA256 is merke-damgard, just feed it some content before and after
<NightRa>
Now I'm on to figure out what bytes exactly...
<NightRa>
Wa, it's not that simple
computerfreak1 has joined #ipfs
computerfreak has quit [Ping timeout: 272 seconds]
cemerick has quit [Ping timeout: 240 seconds]
<NightRa>
It seems to be [10,totalLength-2,8,2,24,0] for the empty file and [10,totalLength-2,8,2,18,totalLength-8,<content>,24,totalLength-8]
Senji has joined #ipfs
bren2010_ is now known as bren2010
<NightRa>
Or [10,4,8,2,24,0] for the empty file and [10,len+4,8,2,18,len,<content>,24,len] when len is the size of the content in bytes
<Kubuxu>
NightRa: it is protobuf encoding if it helps any
whyrusleeping has joined #ipfs
<whyrusleeping>
weird... my vps rebooted or something
<NightRa>
Kubuxu: Yeah I guess, but I'd like not to use protobuf and do something minimalistic. Also with protobuf it would be a huge performance overhead creating a new string in memory compared to feeding the hash function some bytes before and after
<NightRa>
Maybe that's why 'ipfs add' is slow currently
<Kubuxu>
Yup, but you can look what it encodes and how protobuf work to not use protobuf
<Kubuxu>
Nah, it is more complicated
<NightRa>
Yeah, sure
<Kubuxu>
s/work/works
<NightRa>
But I seem to have figured it out already by reverse engineering the output of ipfs :)
<NightRa>
I'll document it somewhere
Senji has quit [Read error: Connection reset by peer]
Senji has joined #ipfs
<NightRa>
It took 'ipfs add' 20 minutes to add 6GB of data, compared to about 2 minutes to hash everything to MD5+SHA256 in a high level language
<Kubuxu>
you can read any text file that has no extension
<NightRa>
No, 3.10 and before
<ion>
Try 0.4.0
<whyrusleeping>
Kubuxu: thats nice!!
<Kubuxu>
NightRa: difference is that ipfs add chunks the file, looks for copies of block in database and create higher level structure
<NightRa>
ion: I certainly will! Also, really waiting for mfs
<whyrusleeping>
now we just need gateways to be writeable
jaboja has quit [Ping timeout: 265 seconds]
<whyrusleeping>
NightRa: what mfs stuff are you looking for?
<NightRa>
Kubuxu: Oh right! IPFS chunks too T_T
<Kubuxu>
yup and have open pinning service with timeout and possibility of extension
<NightRa>
whyrusleeping: My application intends to use ipfs as a file system. Content-hashed. Before mfs it's not really possible
<whyrusleeping>
NightRa: ah, cool
Senji has quit [Ping timeout: 240 seconds]
<NightRa>
Trying to make a collaberative fs by manipulating folder-structure-indexes pointing to files by hashes
<whyrusleeping>
NightRa: ah, cool
<whyrusleeping>
i have some nice primitives for merging dags and stuff in merkledag/utils/diff.go
<NightRa>
Also, really need proper ipns w/ many shared mutable hash pointers
<NightRa>
whyrusleeping: nice :)
<NightRa>
whyrusleeping: you're the one working on mfs?
* whyrusleeping
checks git log
<whyrusleeping>
yeah, something like that
<NightRa>
nice
<whyrusleeping>
its pretty slick in 0.4.0
<whyrusleeping>
since it doesnt really exist in 0.3.10
anticore has quit [Quit: bye]
<whyrusleeping>
the code behind the ipnfs fuse mountpoint is the predecessor
<NightRa>
Argh... some of my plans just got ruined because I realized ipfs chunks the data at some point
<NightRa>
I've made a server serving SHA256 -> HTTP Url for when users go offline, but have a file-hosting cdn-like
cemerick has quit [Ping timeout: 256 seconds]
<NightRa>
But now I either store there the chunks seperately and ruin the http experience or not have ipfs... Or store both at twice the cost T_T
<NightRa>
Well, filecoin will solve this problem when it comes
* NightRa
looks far into the bright future
leer10 has quit [Ping timeout: 246 seconds]
<NightRa>
Would there be a way to quickly determine if a file is chunked and has further links without fetching & looking at the data if it isn't chunked?
<NightRa>
I just started thinking that maybe IPFS hashes don't need to have the standard cryptographic hash properties
<NightRa>
Most important is no collisions, preimage is only nessesary for avoiding collisions. Maybe a concatanation homomorphism property (h(A||B) = h(A) |+| h(b)) would be more important than theoretical weakness of the func.
leer10 has joined #ipfs
<whyrusleeping>
NightRa: you don't have to fetch the entire file to do an ipfs object get on it
<whyrusleeping>
thats the whole point of chunking and the dag
mvollrath has quit [Ping timeout: 255 seconds]
amiller has quit [Ping timeout: 256 seconds]
<NightRa>
Throwing an idea: What if we put in all chunked files the hash of the whole file as additional metadata?
cemerick has quit [Ping timeout: 260 seconds]
<ion>
That could make deduplication difficult.
<NightRa>
The inner chunks will remain shared, the hash of the whole file isn't inside the chunks themselves, but at the chunks' aggregate linking dag
<ipfsbot>
[go-ipfs] whyrusleeping created AtnNn-index_content_type (+2 new commits): http://git.io/vuTgc
<ipfsbot>
go-ipfs/AtnNn-index_content_type a31393e Etienne Laurin: use ServeContent for index.html...
<ipfsbot>
go-ipfs/AtnNn-index_content_type 7f06515 Jeromy: add sharness test for index.html content type...
<ipfsbot>
[webui] Dignifiedquire opened pull request #154: tests: Add example test for react component (master...react-sample-test) http://git.io/vuT2K
<ipfsbot>
[go-ipfs] whyrusleeping opened pull request #2143: index content type (dev0.4.0...AtnNn-index_content_type) http://git.io/vuTa0
ilyaigpetrov has quit [Quit: Connection closed for inactivity]
<ipfsbot>
[webui] Dignifiedquire opened pull request #155: feat: Replace console.log and debug with bows (master...bows) http://git.io/vuTwV
<ipfsbot>
[go-ipfs] whyrusleeping closed pull request #2008: use ServeContent for index.html (master...index_content_type) http://git.io/vBRME
<ipfsbot>
[go-ipfs] whyrusleeping pushed 1 new commit to dev0.4.0: http://git.io/vuTwo