<daviddias>
interesting, git-sha1 perf kicks on crypto perf for randomBytes
<daviddias>
and also that git-sha1 was 5x faster in Node.js than Chrome
<daviddias>
I'm not sure even if I understand why such delta
<dignifiedquire>
yeah those perf results confused me as well
ylp has quit [Ping timeout: 260 seconds]
<dignifiedquire>
this might be not due to git-sha, but also due to the Date + bitwise ops being done
<dignifiedquire>
the benchmark tested the whole generation
ylp has joined #ipfs
<dignifiedquire>
daviddias: updates the PR on js-ipfs now
<dignifiedquire>
daviddias: where can I write down that I am not a big fan of the stream interface of "subscribe" using a simple event emitter like we do on the floodsub impl itself would make much more sense
<dignifiedquire>
and have less over head, and could be transformed into whatever people need
<ansuz>
I'll be deploying three new translations tomorrow
<Kubuxu>
you might want to switch line of work
<Kubuxu>
:)
<ansuz>
polish, german,, and brazilian portuguese
<ansuz>
Kubuxu: why not both?
<ansuz>
I like writing code every now and then
<Kubuxu>
hmm, right
<zaden>
great
rendar has joined #ipfs
slothbag has quit [Quit: Leaving.]
reit has joined #ipfs
<kebyn[m]>
Sweet
ivo_ has joined #ipfs
infinity0 has quit [Remote host closed the connection]
infinity0 has joined #ipfs
Encrypt has joined #ipfs
infinity0 has quit [Remote host closed the connection]
infinity0 has joined #ipfs
<Mateon1>
ansuz: I hold my belief that "guys" is coloquially gender neutral. It is also a fact that there are fewer women in software engineering, so it wouldn't double the responses. Also, nothing is stopping women from responding to a question that starts with "guys".
<ansuz>
suppose I hold up a picture of a woman and say "look at this guy"...
<ansuz>
if you don't feel even a little bit of cognitive dissonance, maybe I'm wrong
<ansuz>
I think most people would
<ansuz>
maybe ask some of the women you know if they identify as 'guys'
* ansuz
shrugs
<Mateon1>
ansuz: That's because "guy" singular is different from "guys" plural.
<ansuz>
if you say so
<ansuz>
I'd still suggest you ask some women how they feel
<zaden>
how'd you feel?
<Mateon1>
Well, I believe it isn't rare for a group of women to refer to themselves as "guys" (as a group) in certain situations
* ansuz
can't tell if implying ansuz is a woman is supposed to be an insult
<mmuller>
FWIW, I decided "guys" is gender neutral when I started hearing girls applying it to groups of girls.
<ansuz>
that's a good argument for me to address groups of men as 'ladies'
<alterego>
Usually it's "lads" not ladies.
<ansuz>
lads actually refers to biological males, though
<ansuz>
lads and lasses
<alterego>
I do often meet some friends and say: "Hello, ladies" but that is considered a joke :)
<ansuz>
well there you go
<ansuz>
ladies is gender neutral
<ansuz>
anyway, just something to think about
<ansuz>
I've made y'all uncomfortable enough, feel free to talk about ipfs
<Mateon1>
I wouldn't say that, as "guys" gender neutral usage is more widespread. Even common. As alterego said, his use of "ladies" is considered a joke.
<ansuz>
k
<Kubuxu>
alterego: yeah, I do something similar as a joke/sarcasm.
<Kubuxu>
Mateon1: the guys thing on IRC for me isn't about the gender but about the etiquette of "don't ask about asking, just ask".
<Kubuxu>
Guys, anyone here? == "Can I ask a question?" == "Is someone willing to commit to answering the question I want to ask?"
<Mateon1>
Well, sometimes I like to ask whether said topic is appropriate to the channel I'm asking on.
<Mateon1>
But people misinterpret that and say "don't ask just ask" or equivalent.
<ansuz>
I fail to see why "don't ask to ask" implies "guys"
<Kubuxu>
ansuz: it doesn't, but joining channel and "guys, anyone around" falls under don't ask to ask.
<Kubuxu>
Mateon1: you loose nothing by just asking the question, if it is wrong channel someone will tell you that, and you can copy the question into another
<ansuz>
seems like "anyone around" would work just as well
<ansuz>
or yea, just the question
bastianilso_____ has quit [Read error: Connection reset by peer]
bastianilso_____ has joined #ipfs
josselin has quit [Remote host closed the connection]
zanadar has joined #ipfs
maxlath has joined #ipfs
sprint-helper has quit [Remote host closed the connection]
zanadar has quit [Client Quit]
sprint-helper has joined #ipfs
<cblgh>
hm, can i somehow see which peers are sharing a piece of content through use of the ipfs hash somehow?
chris613 has joined #ipfs
<alterego>
cblgh: ipfs dht findprovs Qm...
<cblgh>
thanks alterego!
<cblgh>
findprovs short for find provisioners or similar?
<alterego>
providers
<cblgh>
oic
<cblgh>
anyone know of any registries of ipfs hashes?
<herzmeister>
regarding ipns, on an example page it says "Now, there are a few things to note; first, right now, you can only publish a single entry per ipfs node. This will change fairly soon."
<herzmeister>
Has "fairly soon" already happened?
<alterego>
cblgh: you can run: ipfs pin ls
<alterego>
And that returns all the pinned hashes on your local node ;)
<cblgh>
woah
<cblgh>
are all things i add with ipfs add <file> regarded as being indirectly pinned?
<cblgh>
seeing as ipfs pin ls only returns hashes which are regarded either as indirect or recursive
<alterego>
Yes, when you add something locally you are pinning it.
<cblgh>
yes, but when is a hash regarded as directly (rather than indirectly) pinned?
<cblgh>
only if i use ipfs pin <hash>?
<alterego>
Hrm, I think indirect are due to recursive adding maybe.
<alterego>
As in, children of a directory.
<cblgh>
huh strange
maxlath has quit [Ping timeout: 258 seconds]
<cblgh>
i've pinned several tar-files but no direct entries
<cblgh>
doesn't really matter other than to sate my curiosity
palkeo has joined #ipfs
<alterego>
Oh, if a file is large then it gets split into multiple blocks.
<alterego>
So indirect could be blocks of that file.
<cblgh>
wouldn't there be one block that is the direct one?
<cblgh>
maybe not tho if it just pins all the components or something
<zaden>
what if one of them goes offline?
infinity0 has quit [Remote host closed the connection]
infinity0 has joined #ipfs
robattila256 has quit [Ping timeout: 252 seconds]
UgJkA has quit [Quit: ByE ByE]
<kewde[m]>
When you browse something on ipfs will your computer host that content indefinitely?
<alterego>
kewde[m]: only if you pin it.
<kewde[m]>
Ahh, how does the network ensure that the file stays online? And does the initial uploader automatically pin the file?
UgJkA has joined #ipfs
<Mateon1>
The person who adds the file pins it on their computer, then anyone who requests that file caches it for a while (until the repo is garbage collected)
<Mateon1>
IPFS doesn't push content to other nodes, so it's possible for content to be lost if the people pinning the file go offline
DiCE1904 has quit [Remote host closed the connection]
Magik6k has quit [Quit: Bye!]
<Kubuxu>
herzmeister: yes and no, keystore is in place, it wasn't released yet
<zaden>
thats not a convincing feature
<kewde[m]>
This cache, is it visible to other nodes or local only?
DiCE1904 has joined #ipfs
<Mateon1>
IPFS's sister project Filecoin will aim to incentivize storage, so you can pay other people to pin your content
<cblgh>
so excited about filecoin tbh
<cblgh>
retailers will be very grateful for all the hdd sales
<cblgh>
is there any kind of ETA when development of filecoin will start?
<Mateon1>
kewde[m]: No, the cache can't be seen from other nodes. You can only tell if a node provides a hash, but not if it's pinned or in cache.
<Kubuxu>
zaden: otherwise you expose yourself to 1. legal issues 2. game theory says that it is very hard
Magik6k has joined #ipfs
<Kubuxu>
cblgh: Possibly 2017, can't tell anything more.
<zaden>
so everytime i use the service, i've to pay?
<cblgh>
Kubuxu: cool
<Mateon1>
Actually, that's not 100% true, if you use `ipfs dht findprovs <hash>` you only see the nodes providing content that you can find in the DHT. That potentially doesn't include all nodes, but every node in that list is a provider.
<cblgh>
zaden: every time you want to guarantee something is up ya (i think?)
<kewde[m]>
A node also provide unpinned files? Is there are certain percentage to pinned vs unpinned?
<Mateon1>
kewde[m]: Yep, a node provides unpinned hashes
<cblgh>
if you pay for x content to be up, what are you actually paying for? that it will be up for the amount of time you paid for?
<Kubuxu>
kewde[m]: pinning is manual decision made by you, the node ownder
<Mateon1>
There is no percentage, there is a soft limit on the size of the storage.
<Kubuxu>
cblgh: depends on how contract is made, it will be integrated with Eth
<kewde[m]>
Mateon1: is the file retrieve in a direct manner (you get the IP of the node storing the hash) and then connect to it to retrieve it?
<Kubuxu>
kewde[m]: yes
<cblgh>
Kubuxu: ah, neat
<cblgh>
anyone have any resources for reading up on eth contracts?
<kewde[m]>
Is there a ratio of pinned to unpinned files?
<Mateon1>
No, there is no ratio
<Kubuxu>
pinning is manual
<Mateon1>
The owner of a node decides whether to pin content or not
<Kubuxu>
you decide what you pin, seen some content you liked, pin it.
<cblgh>
are there nodes that traverse the net and pin everything as a way of helping the network?
<Kubuxu>
not really, but there are few projects that will be working in this regard
<kewde[m]>
Yes but what about empty storage space is that always filled with unpinned items (stuff you host as charity or use to build up rep?)
<zaden>
so what if the pinning is done random, and no one knows what they're pinning, will that resolve the legal obstacles?
<Kubuxu>
only if you explicitly request this file from the network, for example by visiting a this file
<Kubuxu>
zaden: you don't end up with random content on your node/disk
<Kubuxu>
only stuff that you visit
<zaden>
no no, im asking what if i program to be it so, like rewriting ipfs basics.
<zaden>
Imaginary
<kewde[m]>
Ohh but wouldn't that basically tell everyone this node browsed this content?
<zaden>
so u r telling that only popular contents will survive
<Kubuxu>
kewde[m]: you have to request this content to see it either way
<kewde[m]>
I'm looking into creating more privacy conscious DHT networks, and I have a few ideas, might fork IPFS and add it up there..
<Kubuxu>
zaden: popular content will have higher chance of survival
<zaden>
so ipfs have a limited application, unlike internet
<zaden>
I thought this has a potential to replace Internet, my bad
<Kubuxu>
for non popular content it works like normal HTTP but with the difference that 3rd party might keep the content alive if the 1st party (the original publisher) is not longer interested or doesn't have resources to keep it up
<Kubuxu>
I disagree with your statement
<Kubuxu>
Worst case scenario IPFS works like normal HTTP based internet
<Kubuxu>
unpopular content won't die if there is at least one party interested in keeping it alive
<zaden>
One who visits the ipfs a lot will have to buy large space on his computer , won't he?
<kewde[m]>
Kubuxu well I had something in mind (inefficient but more private) where the blocks are retrieved in a more indirect manner, so no direct connects but a whole series of swaps a bit like Tor I suppose.. But the node you're asking wouldn't know if it was you searching for it or of you're just passing it
<Kubuxu>
zaden: no, as if he is not really interested in the content he seen in the past (he hasn't pin it) it will be removed from his node
<zaden>
kewde[m]: Thats convincable
<Kubuxu>
kewde[m]: there is already TOR integration that is being worked by some people but it will require more work to say that it is not leaking any metadata
<kewde[m]>
It's terribly inefficient and does not scale very well but it works okay up until..
<Kubuxu>
^^ yup, and also it is very DoS vulnerable
<zaden>
couldn't it be just rerouted?
<kewde[m]>
The thing is, Tor does not provide a foolproof way, as you said the metadata, all the unique blocks make up an identity..
<kewde[m]>
A unique combination of what you've searched for, Tor nodes would have to erase all content before switching to a new onion address..
<kewde[m]>
A reboot for example^
<kewde[m]>
Anyways DHTs are a very interesting field, I'm following IPFS closely :)
<kewde[m]>
Thanks for answering my Q's!
<zaden>
Is this a promising field?
<zaden>
I'am a grad student, and i'am currently following AI.
<zaden>
Is this going to get in the mainstream like, in 2 years or so?
<kewde[m]>
I think it has great potential to revolutionize the hosting business, where there is a true free market and perfect competition between cloud providers. Businesses dealing with large networks can mitigate data rapidly to wherever it is cheaper.. It will make the CDN's bleed lol
<lgierth>
PSA: the default bootstrap nodes now all speak ipv6 -- have seen a nice bump in the number of ipv6 peerings last night. the ipv6 multiaddrs will be added to the default config soon
<zaden>
what caused that dim in ipv4 and rise in ipv6
<lgierth>
i restarted all the bootstrap nodes so their peers would try the now-working ipv6 addrs
<zaden>
oh, I see
zaden has quit [Ping timeout: 260 seconds]
ivo_ has quit [Remote host closed the connection]
ivo_ has joined #ipfs
<cblgh>
hot graph
testtesting has joined #ipfs
pfrazee has joined #ipfs
testtesting has quit [Client Quit]
infinity0 has quit [Ping timeout: 240 seconds]
infinity0 has joined #ipfs
gigq has quit [Ping timeout: 258 seconds]
* alterego
ponders a browser plugin that detects when you're doing a download from a site and checks to see if theres a matching ipfs hash.
gigq has joined #ipfs
<alterego>
For instance, I'm downloading an ubuntu image now. I think you can safely presume that someone has mirrored it on IPFS, having a plugin that could find the hash for me would be cool.
<_mak>
is it possible to download a file from a specific seed? can I specify the seedID when downloading a file?
infinity0 has quit [Remote host closed the connection]
infinity0 has joined #ipfs
<alterego>
_mak: there isn't such thing as a seed
<_mak>
alterego: by seedID I meant node ID
<_mak>
sorry
<alterego>
_mak: you'd need to implement it yourself, in go-ipfs, but it is possible.
<alterego>
There's not much point though, you want to get it from the most efficient source possible.
<_mak>
I see.. ok then, thanks
infinity0 has quit [Remote host closed the connection]
infinity0 has joined #ipfs
Pharyngeal has joined #ipfs
Encrypt has quit [Quit: Quit]
ulrichard has joined #ipfs
<pjz>
It would be cool if the plugin maybe knew about overloaded etags
ulrichard has quit [Read error: Connection reset by peer]
<pjz>
it'd require content providers to make their servers provide etags though
Boomerang has quit [Ping timeout: 264 seconds]
ianopolous has joined #ipfs
M-sol56 has joined #ipfs
muvlon has joined #ipfs
Mizzu has quit [Quit: WeeChat 1.6]
<r0kk3rz>
alterego: how would you find the multihash of a file you dont have?
Encrypt has joined #ipfs
<pjz>
r0kk3rz: known-format ETag
<pjz>
r0kk3rz: what if web content were to understand ipfs URLs as ETags
<r0kk3rz>
pjz: yeah, but then you might as well have an ipfs link anyway
<dignifiedquire>
daviddias: are you around?
<daviddias>
+-
<dignifiedquire>
I see pubsub in go and js land are not identical
<dignifiedquire>
in this proto file "topicIDs" is used
<dignifiedquire>
but in libp2p-floodsub "topicCIDs" is used
<dignifiedquire>
also go-ipfs sends via the http-api the full msg object with data, seqno, etc
<dignifiedquire>
but libp2p-floodsub currently only emits the data field, and the interface for pubsub in interface-ipfs-core is not testing anything except the data field
<dignifiedquire>
so I have data missing when trying to communicate with the ipfs-api
<pjz>
r0kk3rz: depends; ETags let you do 'If-Not-Match' in requests, so it falls back to http if ipfs is unavailable