<kythyria[m]>
MikeFair: I've never worked out if POSIX actually lets you do that.
<kythyria[m]>
(namely, have a filesystem object where you can treat it as a directory, but if you read it as a file you get something other than dirents)
<kythyria[m]>
You can do it in webdav, and I suspect Windows Explorer can, but in any case being able to read into individual files like that requires your "filesystem" to understand the formats involved.
ol1garchy has joined #ipfs
<ol1garchy>
Can anyone give me a quick rundown on ipfs
<alu>
yeah sure i got you ol1garchy
<ol1garchy>
Many appreciation points sent to you friendo, alu
<ol1garchy>
Kinda reminds me of freenet to an extent - no clue why.
<SchrodingersScat>
it is neat
<ol1garchy>
Too bad I can't run freenode ;(
<SchrodingersScat>
I guess I could see that, it is a tad similar, at least in anything that caches data will act as sort of a repo
<SchrodingersScat>
the tahoe guy was at an IPFS talk :>
yoh has quit [Ping timeout: 260 seconds]
<whyrusleeping>
IPFS talk? where?
<ol1garchy>
It would be really rad if someone could take IPFS and modgepodge it with GNUNet
<ol1garchy>
you could make a really rad public archieving service
alanz has quit [Ping timeout: 240 seconds]
yoh has joined #ipfs
<SchrodingersScat>
whyrusleeping: or it could have coincidentally had those two, it was on the ipfs.io, https://www.youtube.com/watch?v=CAfagNmIeOE think it was this. Note, it's a shameful youtube video and not a file hosted on ipfs
alanz has joined #ipfs
<whyrusleeping>
SchrodingersScat: i've talked with brian warner and zooko about ipfs on many occasions. The tahoe team definitely knows about ipfs :)
<SchrodingersScat>
I like both projects so far, I had a tahoe cluster up until a bit ago.
<whyrusleeping>
Yeah, a lot of their team has been focused more on zcash lately
<M-anomie>
Is there a good IPFS directory? Sorry if this has been asked before, but I haven't found one.
<alu>
im interested in these social media networks poppin up
<alu>
steem / akasha
<whyrusleeping>
M-anomie: you could check out ipfs-search.com
<alu>
I think something like this could really power some missing meme market infrastructure.
<whyrusleeping>
alu: yesss, ipfs was really made for hosting memes
<alu>
yeah janusvr has had site translators for reddit and youtube for awhile but there could be a 3D decentralized version with market place that belongs to the people.
<whyrusleeping>
I think open bazaar + VR could be really cool
<whyrusleeping>
sell real good in an MMO VR marketplace
ianopolous has quit [Read error: Connection reset by peer]
<whyrusleeping>
s/good/goods/
ianopolous has joined #ipfs
<alu>
yeah, the basic direction the web is going is to swallow gaming/film (webassembly!) and turning it into a giant mmo
matoro has quit [Ping timeout: 255 seconds]
<kythyria[m]>
Ew
* kythyria[m]
is firmly in the multiplayer camp
<kythyria[m]>
*single-player
<SchrodingersScat>
make life your game, destroy real lives
<SchrodingersScat>
If you don't know what it's like to crush the dreams of someone, you're playing on the wrong level.
<alu>
how many levels u on mah dude
<SchrodingersScat>
all eleven dimensions brother, I can see through space and time.
<aggelos__>
nevermind the 'federated' aspect, wth does this have to offer?
<aggelos__>
(sorry for the offtopic)
nausea has joined #ipfs
nausea has joined #ipfs
Boomerang has joined #ipfs
ruby-ruby has quit [Ping timeout: 240 seconds]
ashnur_ has joined #ipfs
<ashnur_>
hi
<ashnur_>
what does "merkledag: not found" mean?
<ashnur_>
i am trying go-ipfs on win10
ruby-ruby has joined #ipfs
<MikeFair>
ashnur_: I think it means you haven't issued an ipfs init yet
<ashnur_>
i did
<MikeFair>
Then perhaps ipfs daemon?
<ashnur_>
yeah, it actually means that ipfs developers can't write meaningful error messages
<ashnur_>
also, i had to start the daemon
<MikeFair>
ashnur_: Is that what it was the daemon wasn't there?
<ashnur_>
i am sorry, i don't understand the question
<MikeFair>
ashnur_: It worked after starting the daemon
<MikeFair>
?
<ashnur_>
yes
<MikeFair>
ok; got it; thanks
<ashnur_>
not sure why i had to start the daemon for only getting a file though
<MikeFair>
Because you need a peer id to go hopping through the p2p net
<ashnur_>
well, if that's how it works
cemerick has joined #ipfs
<MikeFair>
ashnur_: Your local daemon makes requests, issues provides, and collects all the disparate responses for all the pieces of the file your collecting for you
<MikeFair>
but I've found that's not as good for downloading file as it is just for testing it works
reit has quit [Remote host closed the connection]
colpoplasty has joined #ipfs
espadrine has joined #ipfs
<MikeFair>
I don't suppose getComponent<ClassName> is smart enough to get either the provided class or a derived child class of ClassName?
<MikeFair>
wrong channel :)
<ashnur_>
yeah, so seems like files get corrupted
<ashnur_>
says it finishes 99.98% byte size on disk is same, but md5 hash gives different results
<ashnur_>
and when i try to use the file, virtualbox says it got corrupted
damongant has joined #ipfs
<ashnur_>
so, is this uploaded in a corrupted way if i can't download only 99.98% of it?
<ashnur_>
or what can be the issue
<MikeFair>
I'd suspect; given the earlier errors either something got written wrong (unlikely) or the file was corrupted on upload
<MikeFair>
ashnur_: what command uploaded and what command downloaded it
<johan__>
@ashnur - Check the file and try to add it again.
<MikeFair>
ashnur_: I've download a directory but didn't realize it at first once
<ashnur_>
we tried to add it again and it gave the same hash, is that a problem?
tmg has quit [Ping timeout: 240 seconds]
<ashnur_>
i don't think there is any problem regarding the content
<ashnur_>
not sure what you mean by check the file @johan__
<ashnur_>
the original file works as intended
<ashnur_>
the one i download is corrupted, even though the byte size is the same, ipfs reports on 99.98% downloaded before finishing and virtualbox is complaining that it can't import the file
<johan__>
I meant that you should check the original file and the try to add it again. (the 99.98% I think is just cosmetic issue, happens me all the time but never anything wrong with the files)
<MikeFair>
ashnur_: And the file works on teh source box obviously
<ashnur_>
the original file does, yes
<ashnur_>
i think ipfs broke the file
colpoplasty has quit [Remote host closed the connection]
ShalokShalom has quit [Ping timeout: 260 seconds]
ShalokShalom_ has joined #ipfs
<ashnur_>
so now we are using the traditional method of moving large files
<ashnur_>
pigeons!
dimitarvp has joined #ipfs
maxlath has quit [Ping timeout: 268 seconds]
<johan__>
@ashnur, try to run ipfs pin rm <your hash> and then do a garbage run with ipfs repo gc and finally ipfs add <your file> again
aquentson has joined #ipfs
chriscool has left #ipfs [#ipfs]
aquentson1 has quit [Ping timeout: 260 seconds]
<ashnur_>
ooook, i think i will have to read up first on what all those commands do :)
<MikeFair>
ashnur_: Yes; 'ipfs pin filehash' on the source file is a good idea
konubinix has quit [Remote host closed the connection]
aquentson1 has joined #ipfs
ttk2[m] has quit [Ping timeout: 240 seconds]
ttk2[m] has joined #ipfs
pcre has quit [Ping timeout: 260 seconds]
konubinix has joined #ipfs
aquentson has quit [Ping timeout: 240 seconds]
aquentson has joined #ipfs
aquentson1 has quit [Ping timeout: 240 seconds]
maxlath has joined #ipfs
pcre has joined #ipfs
aquentson1 has joined #ipfs
aquentson has quit [Ping timeout: 240 seconds]
aquentson has joined #ipfs
aquentson1 has quit [Ping timeout: 240 seconds]
anaptrix has quit [Ping timeout: 240 seconds]
krig[m] has joined #ipfs
anaptrix has joined #ipfs
cwahlers has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ShalokShalom_>
What are your thoughts about this?
<ShalokShalom_>
I think its possible that the current technology is able to provide a Wikipedia fork based on distributed internet software.
<ShalokShalom_>
Wikipedia is one of the most visited sites in the world, while it takes donations to provide their service, since there is no advertising.
ShalokShalom_ is now known as ShalokShalom
cehteh has quit [Ping timeout: 252 seconds]
<MikeFair>
ShalokShalom: I'd be more inclined to partner with Wikipedia then attempt to fork them
<MikeFair>
ShalokShalom: Do you have something against them?
<ShalokShalom>
you think they are interested?
cehteh has joined #ipfs
The_8472 has quit [Excess Flood]
<ShalokShalom>
well, i think they know well, how much it costs for them
The_8472 has joined #ipfs
<ShalokShalom>
so you suggest that they are not aware about distributed systems like IPFS?
<ShalokShalom>
and no, i have nothing against them
<MikeFair>
I think if you showed them a technology that repesented an open CDN where they could still provide their editors content authority / voting / and management systems; of course they would
<MikeFair>
ShalokShalom: I think these distributed systems don't solve all their needs
<ShalokShalom>
and why do they not develop that by them self?
<ShalokShalom>
would cost much less money
<ShalokShalom>
not all of course and not now
<ShalokShalom>
they get millions each year..
<MikeFair>
ShalokShalom: It's an entirely different kind of engineering mindset; it's like "Why don't all the Soccer Player's taek on mastering Rugby"
<ShalokShalom>
i dont think so
<MikeFair>
ShalokShalom: p2p distributed systems using content based addressing require an entirely different way to manage things; you can't even "Log in" to ipfs
<ShalokShalom>
you can create that
<ShalokShalom>
so you think its undoable?
<MikeFair>
ShalokShalom: That's the challenge; you have to invent the how
<ShalokShalom>
cost much more money as their current server setup?
<ShalokShalom>
ey come on
<ShalokShalom>
thats easy, compared to what they do know
<ShalokShalom>
imho
<MikeFair>
ShalokShalom: They control their current setup; p2p puts them out of control; now they have a whole new set up of problems
<ShalokShalom>
yeah, maybe
<ShalokShalom>
true
<MikeFair>
ShalokShalom: Can someone poison the data using their local node?
<ShalokShalom>
thats probably a key question, yes
<MikeFair>
ShalokShalom: Can someone bypasss the authentication and post as someone else?
<ShalokShalom>
yep
<ShalokShalom>
is all that still unsolved in ipfs?
<MikeFair>
ShalokShalom: How do you handle it if someone locks themselves out of an account they created and no one else can access?
<ShalokShalom>
plus all the other distributed systems, like Tox, morphis and so on?
<MikeFair>
ShalokShalom: How do we answer people if they want to do a security audit
<MikeFair>
ShalokShalom: There are ideas; but not "the solution"
<ShalokShalom>
aha, ok
<MikeFair>
ShalokShalom: There's still no good implementation of a search function
<MikeFair>
ShalokShalom: So I think an system like IPFS can help them with content distribution; but it's not a replacement
<MikeFair>
yet
<MikeFair>
Search and the concept of an Authenticated User are the two areas I'm focused on
<MikeFair>
Some systems do authentication pretty well; but for example, in IPFS, every time you drop a file in your home directory, your home directory changes address
<MikeFair>
Anyone and everyone can read your home directory
<MikeFair>
(So then you encrypt it -- and now the fact the bytes change every write is enhanced)
shizy has joined #ipfs
<MikeFair>
ShalokShalom: That's not to say these aren't "solveable" problems; but they are different than making sure you filesystems stay synchronized across your distributed data centers
<ShalokShalom>
you can combine different systems?
<ShalokShalom>
which solutions handle authetification well in your opinion?
matoro has quit [Ping timeout: 240 seconds]
<ShalokShalom>
i can think of one main server or more, with mirrors
<ShalokShalom>
plus the actual pages get shared via ipfs
<ShalokShalom>
how many users call the page without being editors?
<ShalokShalom>
so you can manage all that security related stuff on the main server and its mirrors?
<MikeFair>
centralization :)
<MikeFair>
If you have to build a large centralization infrastructure to handle that load; it's not that much harder to serve the content too
<MikeFair>
and you have a simpler system overall
<MikeFair>
err simpler to understand for those doing the engineering of it
<MikeFair>
I've only recently come up with an auth solution I like; and it starts by eliminating the keys from user's view (which is something I've never seen applied to a system before) -- mostly because the user's need the key to authenticate their updates
<MikeFair>
What happens instead is a passphrase encrypted version of the key is uploaded instead; and then you have to send in an encrypted version of the passphrase along with your update
<MikeFair>
It doesn't quite work in native ipfs because anyone/everyone can see everything and there's no real "txn approval" concept; but I've been working with the Stellar consensus system; and there it seems to make more sense; because the ledger has this "txn approval" phase before it acknowledges a change to an object
<MikeFair>
Ther is txn approval in ipfs for something like ipns updates ; but that only checks to ensure you had the right private key
<MikeFair>
It's interesting working through some of these things when you have to assume that anyone on the internet can plug their system straight into your data
<ShalokShalom>
why does it cost so much performance to handle the load of the edits?
<MikeFair>
ShalokShalom: It's the integrity of the edits more so than the load
<ShalokShalom>
So, the integrity of the edits cost most of Wikipedias load?
<MikeFair>
ShalokShalom: No; keeping their integrity does
<ShalokShalom>
why?
Silke[m] has quit [Ping timeout: 240 seconds]
<MikeFair>
Because you want to ensure that thoe who made the change were authorized to do so; that edits weren't made without comments; that discussion boards aren't messed with; that spammers don't put ads into the middle of all your content
Silke[m] has joined #ipfs
<ShalokShalom>
i mean why does it cost so much performance
<MikeFair>
It doesn't, the read load does to ensure the data gets served up
<MikeFair>
ther eare a load of people editing; and the writes do cost more than the reads; but distributing the read load to where the demand is creates the "performance" requirements
<MikeFair>
Fast sites are typically always fast because they have a lot of extra unused capacity
<MikeFair>
Data that's in a server's RAM cache is fast than on disk; even if that server is across a local network
<MikeFair>
s/fast/faster
zeus1 has joined #ipfs
<MikeFair>
So making many servers see the same data quickly is a challenging engineering task
<MikeFair>
(At the WikiPedia Scale)
<MikeFair>
The integrity portion of it is so they can earn our trust
<ShalokShalom>
so what is the solution?
<MikeFair>
Which largely translates to having control over the data sources
<MikeFair>
ShalokShalom: We create breakthroughs in how we understand and solve the problem; the successful execution of BitCoin stands on at least a dozen different solutions to different problems combined in just the right way; people saw what BitCoin did with their ledger and have done even better things with it
<ShalokShalom>
ok, fine?
<MikeFair>
ShalokShalom: In the WikiPedia case; we need to put data on your hard drive that you can't edit even if you knew how
<ShalokShalom>
ok
<MikeFair>
that's hard
<ShalokShalom>
kk
<ShalokShalom>
because its against the very nature of ipfs?
<MikeFair>
security in computers has largely meant "physical security" as the base axion
<ShalokShalom>
so, when i share a static page, can everybody edit that one?
<Reventlov>
No, because if you have physical access to something, you can pretty much do what you want with this something
<MikeFair>
ipfs isn't ipfs if it's not "distributed nodes"
<ShalokShalom>
what i mean
<MikeFair>
ShalokShalom: not in ipfs; that's one of the "breakthroughs" the static page you published has a unique page ; to change the page is to change the address
<MikeFair>
err unique address
<MikeFair>
ShalokShalom: which means that in ipfs; the filename changes every time you update and you have no way to predict what it's going to be before you make the changes
<MikeFair>
ShalokShalom: so if you want to ensure everyone gets your versions of the page; then this works great
<ShalokShalom>
hnn
<MikeFair>
ShalokShalom: until you make a change; no you have to give everyone the new address
<MikeFair>
nwo
<ShalokShalom>
so, i can share a Wikipedia which is just editable by me?
<MikeFair>
ShalokShalom: Sure
<MikeFair>
ShalokShalom: but we'd all have to agree to your edits
<MikeFair>
ShalokShalom: I mean in the sense that we likely won't visit your pages
<ShalokShalom>
sure
<MikeFair>
ShalokShalom: What we want is "the latest/best answer the community has created so far"
<ShalokShalom>
i see
<ShalokShalom>
which is not always the case, anyway :)
<MikeFair>
ShalokShalom: Fewer of us, but enough, also want to see the history of the page and why the edits that have been made were made and the discussion about all that
<ShalokShalom>
is that possible now with distributed systems?
<MikeFair>
ShalokShalom: So ipfs gives a great way to "store" all the data; retrieving and assembling it has presented a bit of a challenge; mostly because all the addresses keep changing every time anything is changed
<ShalokShalom>
and redirect the old address to the new one?
<MikeFair>
ShalokShalom: so you have to create some "non-changing things" aka "immutable addresses" where you can hook things; and ipfs does this with their ipns system
<ShalokShalom>
ah
<ShalokShalom>
so?
<ShalokShalom>
whats the down side of it?
<MikeFair>
ShalokShalom: Yes; though you can't "redirect" an old address; the address is the address; you make a new address that redirects to the "latest answer"
<MikeFair>
who gets the permission/privilege of updaating the central address?
<MikeFair>
it's not so much a "downside" as a consequence
cwahlers has joined #ipfs
ashark has joined #ipfs
<MikeFair>
ShalokShalom: How can the system tell that the one's we wanted to control that address were the people who actually did it; and what are the security weak points of independent systems "hijacking" the address
<ShalokShalom>
i see
<ShalokShalom>
so its all about the authentification now
<MikeFair>
when it's all under your physical control ; these answers are "easy" but "content distribution" is hard ; ipfs inverts the situation ; and has exactly the opposite problems :)
<MikeFair>
ShalokShalom: It's not the only one; however it's one of the "linchpins" in my book
<ShalokShalom>
and a hybrid system is not suitable?
<MikeFair>
ShalokShalom: It could be; but it mixes "thought methodologies" it's putting rugby players and soccer players on the same field
<ShalokShalom>
yeah, great
<ShalokShalom>
a new game
<ShalokShalom>
;)
<MikeFair>
Who's going to coach them?
<ShalokShalom>
both, soccer and rugby trainers?
<MikeFair>
ShalokShalom: Now you have to get those trainers to agree with each other
<ShalokShalom>
yep
<MikeFair>
ShalokShalom: After you've found them in the first place :)
<ShalokShalom>
do you know about such a hybrid system?
<MikeFair>
ShalokShalom: Raising more money and buying a bunch of equipment that doesn't talk back is "easier" :)
<ShalokShalom>
decentraled systems are something like that?
<ShalokShalom>
doesnt talk back?
<ShalokShalom>
what does this mean in this example?
<MikeFair>
ShalokShalom: The distributed ledger systems are the closest examples of this
<MikeFair>
"doesn't talk back" means a computer won't disagree with how you intend to approach a solution; it'll do what it's been programmed to do without having an opinion about its correctness
<ShalokShalom>
yep
<MikeFair>
it always agrees with what you told it to do
<MikeFair>
(which is not saying that you always told it do what you meant to say)
ruby-ruby has quit [Ping timeout: 240 seconds]
<ShalokShalom>
where is the difference between distributed ledger systems and stuff like ipfs and ZeroNet?
<MikeFair>
don't know about zeronet; but in distributed ledger there's a general systemwide agreement process on what happened that each individual computer can recognize
<MikeFair>
in ipfs it's much more "trusting" and "local"
<MikeFair>
For instance; if you ask me for a certain address; in ipfs I can give you the data; and you check to see if I messed with it
ruby-ruby has joined #ipfs
<ShalokShalom>
how is that auth solution called, which eliminate the keys from users view?
firemound[m] has quit [Ping timeout: 240 seconds]
firemound[m] has joined #ipfs
<MikeFair>
ShalokShalom: I don't have a real name for it yet; but I've been calling it a "signing service"
<ShalokShalom>
ah, its from you?
<MikeFair>
ShalokShalom: There is some "server" or "service" out there; that can get or recreate the required key; if I give it the right information
<ShalokShalom>
and Stellar is suitable to handle auth?
<ShalokShalom>
is there a source code from you?
<ShalokShalom>
did you already code it?
<MikeFair>
ShalokShalom: it's only task is receive this information; stamp or encrypt the data I gave it; and send it back to me
<ShalokShalom>
kk
<ShalokShalom>
and Stellar is suitable for distributed systems, including ipfs?
<MikeFair>
parts of it; but no Stellar does not do this yet; but I see that it can
<ShalokShalom>
ah ok i see
<ShalokShalom>
:)
<ShalokShalom>
hnn, do you know about a central place of all this stuff?
<MikeFair>
Stellar is a distributed "ledger" which I can see being the place to store the "fixed addresses" that we want to protect
<ShalokShalom>
i really think it can be benefiting, when tools like ipfs, morphis and so on join forces
<MikeFair>
ShalokShalom: I don't know of any place that centralizes it; and I'm sure that if you were talking to someone else instead of me; you'd get an entirely different opinion. (Well maybe not "entirely different" but definitely different :)
<MikeFair>
I gotta go now; but it's been fun; if you hang out; there'll be more folks from the project around here who can fill you in
<ShalokShalom>
fine, thanks
<ShalokShalom>
good day
<MikeFair>
And yes! many people agree that as we solve these challenges; the benefits are tremendous
<ShalokShalom>
ok, so together ;)
reit has joined #ipfs
ruby-ruby has quit [Ping timeout: 240 seconds]
ulrichard has quit [Remote host closed the connection]
Boomerang has quit [Quit: Lost terminal]
<ashnur_>
btw, the pin rm, repo gc, readding seems to fixed the file and now it works
<ashnur_>
thanks
ylp1 has quit [Quit: Leaving.]
<ashnur_>
although it seems that using an usb flash drive is still the easiest method for sharing stuff :)
pcre has quit [Ping timeout: 260 seconds]
vapd is now known as \\\\\\\\\\\\\\\\
<johan__>
ashnur. Good to hear. And yes, usb flash drives still fill a good purpose.
<johan__>
But I personally think that ipfs sees a great future for the web. Might be as a complement or as a new standard for how we store and share information.
Mizzu has joined #ipfs
<MikeFair>
ashnur_ / johan__ : fwiw, I plan on making an ipfs backed virtual usb "flash drive" mountable device :)
<dryajov>
lgierth: you around?
<MikeFair>
So you can put in a single hash that represents the virtual drive; and then format / use it like a normal USB drive
matoro has joined #ipfs
Boomerang has joined #ipfs
caiogondim has quit [Quit: caiogondim]
caiogondim has joined #ipfs
Boomerang has quit [Quit: Lost terminal]
maciejh has joined #ipfs
mguentner has joined #ipfs
cemerick has joined #ipfs
konubinix has quit [Quit: ZNC 1.6.4+deb1 - http://znc.in]
<ShalokShalom>
and why not simply take one snapshot from wikipedia, provide that via a distributed system and update regular?
<ShalokShalom>
so you cant edit directly on that snapshot release
<ShalokShalom>
MikeFair: ^
yosafbridge has quit [Quit: Leaving]
Foxcool has quit [Ping timeout: 240 seconds]
seharder has joined #ipfs
konubinix has joined #ipfs
Guest40893 has quit [Ping timeout: 240 seconds]
yosafbridge has joined #ipfs
<lgierth>
dryajov: yeah in a few -- didn't have time to respond to your relay comment yet
<dryajov>
lgierth np.. also, might be good to hop on hangouts/skype/whatever to discuss relay in general, let me know what you think
bvcosta has joined #ipfs
Guest44283 has joined #ipfs
<lgierth>
dryajov: ok we can do that -- can you do 30min at 9p utc?
ruby-ruby has joined #ipfs
maciejh has quit [Ping timeout: 260 seconds]
kenshyx has quit [Quit: Leaving]
Encrypt_ has joined #ipfs
<dryajov>
lgierth: yeah I think I can make it at 9pm UTC. What time zone are you in, for future reference? I'm in Pacific.
<lgierth>
UTC+1 / CET
<dryajov>
kk
maciejh has joined #ipfs
<lgierth>
but my daily life is in fact mostly eastcoast time :)
<dryajov>
ha! :)
<dryajov>
yeah, can imagine
cemerick has quit [Ping timeout: 260 seconds]
PrinceOfPeeves has joined #ipfs
s_kunk has quit [Ping timeout: 268 seconds]
matoro has quit [Ping timeout: 268 seconds]
Mossfeldt has joined #ipfs
<zippy314_>
hi folks libp2p question re travis-ci: Some of my tests run connections over localhost, and they work just fine on my machine. The travis run fails with: 'dial attempt failed: context deadline exceeded' (https://travis-ci.org/metacurrency/holochain/jobs/204630289) anybody had to deal with this before?
mildred4 is now known as mildred
cemerick has joined #ipfs
Kane` has joined #ipfs
<Kubuxu>
zippy314_: yeah, travis can be slow-ish sometimes
<Kubuxu>
hmm
<Kubuxu>
I would suggest rebuilding.
<zippy314_>
Kubuxu: well, I just committed something else, so we'll see if it works this time...
<zippy314_>
hmmm, looks like it worked. That's interesting.
absdog18[m] has quit [Ping timeout: 240 seconds]
cemerick has quit [Ping timeout: 240 seconds]
absdog18[m] has joined #ipfs
cemerick has joined #ipfs
Captain_Beezay is now known as TurboC
TurboC is now known as Binome
gmoro has quit [Ping timeout: 255 seconds]
jedahan has joined #ipfs
tilgovi has joined #ipfs
gmoro has joined #ipfs
gmoro has quit [Remote host closed the connection]
<Mossfeldt>
whyrusleeping: Got the wiki to work. Had configure CORS for the daemon and restart. It's in the readme but as I understand it shouldn't be nessecary.
<whyrusleeping>
Mossfeldt: what version of ipfs did you first run 'ipfs init' with?
<Mossfeldt>
0.4.5
<whyrusleeping>
Hrm, odd. That should have those commands as the default.
<whyrusleeping>
Kubuxu: any idea there?
<Kubuxu>
it is
<Kubuxu>
they require access to the full API
<Kubuxu>
Mossfeldt: I wouldn't recommend running it like that, now every page you open in your browser can mess with your daemon.
<Mossfeldt>
Yeah, I know. It's running on a machine just for testing. Will wipe it clean after som more playing around.
<Mossfeldt>
Kubuxu: should be enough to clean the config of API Access-Control and restart the daemon right?
<Kubuxu>
yeah
<Mossfeldt>
Great.
<ShalokShalom>
Mossfeldt: can i use content from Wikipedia?
jedahan has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<ShalokShalom>
so, download/fork Wikipedia and share its content via that software?
Mizzu has quit [Ping timeout: 255 seconds]
<Mossfeldt>
I guess its doable in some way but as far I understand it not out of the box. Don't think its intended to do that. Contact the creator jamescarlyle and discuss the idea, its interesting at least.
<zippy314_>
yah but did you see that is took the equivalent of 6500 years of a GPU computation time... I'd love to see the comment you'd have to insert into a real world git commit to get it to hash correctly and still compile..
__uguu__ has joined #ipfs
__uguu__ has joined #ipfs
__uguu__ has quit [Changing host]
<lgierth>
in 3 years you can do it on any smartphone in a few seconds
Mossfeldt has quit [Quit: Konversation terminated!]
__uguu__ has left #ipfs [#ipfs]
<MikeFair>
zippy314_: I saw that; and the equivalent CPU time was a number I don't even a name for the highest decimal place ;)
ianopolous has quit [Ping timeout: 255 seconds]
\\\\\\\\\\\\\\\\ is now known as t2deep
jeraldv has quit [Quit: leaving]
<zippy314_>
But here's the important point, taken from a the ycombinator thread: "they can't find a collision for an arbitrary file, they can only craft two particular files that happen to collide."
<zippy314_>
still great fodder for multihash!
<whyrusleeping>
history has shown us that collision attacks lead to preimage attacks
<whyrusleeping>
its only a matter of time before someone is able to use this work to figure out how comping a preimage works
ashark has quit [Ping timeout: 260 seconds]
wallacoloo_____ has quit [Ping timeout: 240 seconds]
ruby-ruby has joined #ipfs
wallacoloo_____ has joined #ipfs
tg has quit [Quit: Leaving]
tg has joined #ipfs
AndromedaGamarue is now known as Captain_Beezay
n0z has quit [Ping timeout: 260 seconds]
Aranjedeath has quit [Ping timeout: 240 seconds]
bwn has quit [Ping timeout: 268 seconds]
<mildred>
Kubuxu: hello
<Kubuxu>
\o
<mildred>
shouldn't bitswap be put to its own repo?
maxlath has quit [Quit: maxlath]
<mildred>
in libp2p orga?
<Kubuxu>
we are considering it, but it is what makes ipfs the ipfs. Also it is very rough and not library like iirc.
ashark has joined #ipfs
<lgierth>
yeah it's more ipfs than anything else
<mildred>
in fact, I'd like a version of the bitswap codebase with no gx dependencies
<lgierth>
go extract it :)
<lgierth>
ipfs/go-bitswap
bwn has joined #ipfs
<Kubuxu>
problem is, non-gx version will break frequently
<lgierth>
it can't depend on anything in go-ipfs.git though
<Kubuxu>
we use gx as otherwise developing with go in modular fashion is not possible
<mildred>
in fact, I'd like to use a custom DHT, but I can't give it to the bitswap code because the import paths are wrong
<mildred>
basically, I'd like to customize go-libp2p-kad-dht
<mildred>
and I get tons of errors because things like peer ids have different import paths
<Kubuxu>
we have go-libp2p-routing interface
<Kubuxu>
aah
s_kunk has quit [Ping timeout: 240 seconds]
<mildred>
something like "github.com/libp2p/go-libp2p-routing".IpfsRouting does not implement "gx/ipfs/QmZghcVHwXQC3Zvnvn24LgTmSPkEn2o3PDyKb6nrtPRzRh/go-libp2p-routing".ContentRouting
<mildred>
if the interfaces used plain []byte, that may work
<mildred>
or I can use a proxy object to do some type conversion of identical objects
<lgierth>
either 1) extract bitswap, or 2) try gx-go rewrite --undo but good luck with that :)
Aranjedeath has joined #ipfs
Aranjedeath has quit [Changing host]
Aranjedeath has joined #ipfs
<whyrusleeping>
mildred: i'm going to have tooling to make this much easier soon, but in the meantime you can try `gx-go rewrite --undo go-libp2p-kad-dht`
<whyrusleeping>
and then in the dht repo, run `gx-go rewrite`
<whyrusleeping>
or for go-libp2p-routing
<mildred>
whyrusleeping: wouldn't it be better to have import paths that don't change each time (using IPNS?)
<whyrusleeping>
mildred: no, thats the whole point of using gx
<whyrusleeping>
we want our dependencies to always point to what we expect them to
<whyrusleeping>
you can revert back to the github imports using the gx-go rewrite --undo
tilgovi has quit [Ping timeout: 260 seconds]
<whyrusleeping>
(though i don't think go-ipfs builds currently without its gx deps, the libp2p packages all do)
<whyrusleeping>
and yes, extracting bitswap is on the todo list
ashark has quit [Ping timeout: 240 seconds]
<tmg>
what's the downside to merging these gxified "dependencies" back into the go-ipfs repo?---do any other projects actually use the go-libp2p*?
<lgierth>
lidel: about "faking fs: in the address bar", we can make an fs: uri into a fs:// url with proper origin, by deriving the host part from the path. do you think we can still keep the fs: uri without host in the address bar and other UI, while under the hood using the fs:// url variant of it?
s_kunk has joined #ipfs
<lgierth>
lidel: i have the fs: uri => fs:// url figured out, i'm just wondering how feasible it currently is to keep the fs: uri in the UI
tilgovi has joined #ipfs
Caterpillar has quit [Quit: You were not made to live as brutes, but to follow virtue and knowledge.]
<lgierth>
lidel: if it's not feasible at the moment, the fs:// url in the UI would be very verbose (fs://$base32hash/ipfs/$originalhash/path/within) and we should then just go with ipfs://$base32hash/path/within and ipns://same for now
<lidel>
lgierth, I am not aware of WebExtension API for manipulating address bar, if that is what you ask. so the original protocol and URI will be there
<lgierth>
lidel: ipfs:// and ipns:// are clearly inferior to a properly "faked" fs: uri with fs:// url
<lgierth>
and in addons, at the moment?
<lidel>
in addons you can do whatever you want, as long as you have some good painkillers
<lidel>
;)
<lidel>
s/addons/legacy-sdk/
redlizard has joined #ipfs
redlizard has joined #ipfs
redlizard has quit [Changing host]
<lgierth>
btw, i tried it out in the addon v1.x last night, no fs: faking didn't work with /ipns
<lgierth>
what's the general issue with faking code (apart from unclear support in webextensions), is it that it depends very tightly on firefox internals?
palkeo has quit [Quit: Konversation terminated!]
<lidel>
lgierth, yes, you need to know internals basically
<lidel>
and from what I've seen people read firefox sources to come up with XUL code when it comes to advanced stuff
<lidel>
this is one of reasons why they deprecated all XUL-based APIS
<lidel>
(that, and the fact they broke between firefox versions)
<lgierth>
it's still most awesome to have it working, even if it's just very hackish. just to make a case for this kind of functionality
shizy has quit [Ping timeout: 268 seconds]
<lidel>
legacy sdk == have it working till the end of 2017
<lidel>
:)
<lgierth>
there could be restrictions for it, e.g. you can keep the origin-less uri in the address bar, IF your protocol handler provides a way to derive an origin (host) from the path. ipfs would meet that
<lgierth>
yeah i read that announcement the other day