i think some of my personal data on p2p concerns can be alleviated with encrypting the contents inside the dat archives, and maybe some nice 2FA way of managing keys. nice private key management still hasn't really been done nice yet
i wouldn't trust myself to design a key management scheme
iml_ joined the channel
ogd: compiled my own hyperkit from git ... let's see if that helps
iml_ has quit
ogd
jimpick: that's what i'm using
jimpick
hmm, still seeing the same problem
iml_ joined the channel
ogd: not 100% sure, but i think you need to pass something like '-s 8,virtio-rnd' to hyperkit to emulate a hardware entropy source
and run a kernel driver
CONFIG_HW_RANDOM_VIRTIO perhaps?
i'll see if i can rebuild hyperos with that
it's been so long since i've built a kernel ... which is funny because i used to run kernel.org :-)
dat-gitter1
(serapath) @peterVG
(serapath) > Fortunately, Dat has a rich ecosystem of components that is addressing this so I can focus on app-level collections management functionality (which I see a layer above Beaker, e.g. as a Beaker app or fork).
(serapath) Just followed another discussion on patchwork where @luxxor wants to take time to fork beaker so he can add ssb support. Yet someone else mentioned the need for an `ad blocker` in beaker and i'd like to have a `bookmarkbar`. so rather than a fork - what do you think of: https://github.com/beakerbrowser/beaker/issues/712
(e-e-e) mafintosh: jhand: I am noticing strange behaviour using dat to clone from my computer to a remote server. The connections on the server are ossilating between 0-6 while my computer goes from 0-1 every few seconds. I am guessing that this means that the connections are being interupted and reestablished constantly. Any thoughts?
i'm sort of dogfooding the blogging platform ... i'm surprising myself ... it's actually really easy to blog this way
i'm realizing the reason i never blogged was because i only had one blog ... if i have the freedom to create an unlimited number of blogs basically for free, there's really no excuse not to write stuff down
i haven't dropped an api into it yet, but a single linux vm should be enough to support thousands of blogs
mafintosh
@e-e-e update to latest both places. Bunch of important fixes that could be related
dat-gitter1
(e-e-e) ah cool - I was thinking that might be the case mafintosh:
mafintosh
@e-e-e let me know if that works better :)
dat-gitter1
(e-e-e) will do
ZerataX has quit
ZerataX joined the channel
(e-e-e) mafintosh: faster but still seemingly eratic with the connections oscilating
(e-e-e) but much quicker
mafintosh
@e-e-e still between 0-6?
dat-gitter1
(e-e-e) less 0-3
(e-e-e) with it more often at 2
(e-e-e) actually watching it now it looks fine - stable at 2
(e-e-e) not sure who the second connection is though :/
jimpick
hmmm, this buildroot build has been going for 3.5 hours ... all so i could change one kernel config setting
i think i'll have to let it run and check in on it when i wake up
dat-gitter1
(e-e-e) Oh also mafintosh: and anyone else interest - I have been working on a little http proxy for serving static dats from http addresses. Kind of like hypercored with our the writing. https://github.com/e-e-e/dat-http-proxy
(e-e-e) its a bit fragile at the moment, malformed urls seriously break.
(e-e-e) but it works
mafintosh
@e-e-e the other connection is yourself - a duplicate. we should move to show the connections per default
@e-e-e very good!!
dat-gitter1
(e-e-e) the idea is to have a small server that can server as redundancy for many many dats, at a low cost overhead
(e-e-e) its definitely way way slower though
red-tomato has quit
(e-e-e) mafintosh: could you offer any insight into when/if raf or ram .read() call returns error to the callback it would fail hard. I am noticing that using http most of the time ’timeout’ errors are non critical, but then once every few minutes if a timeout occurs it is critical - killing the process entirely
(e-e-e) any thoughts?
kobkunkrab joined the channel
kobkunkrab has quit
(e-e-e) its ok - looks like I can capture the errors by adding an event listener to hyperdrive archive. Still a bit confused why not having a listener would mean the process crashes - but the problem appears to be fixed by it.
i'll test some more, and if it works, i'll send some PRs ... but i'm moving back into my condo today after having the flooring replaced, so it might be later
G-Ray joined the channel
ogd
jimpick: going to a hot springs in the mountains today
jimpick
nice!
i've been looking at some of the hot springs in oregon ... might do that on my next drive down there
my wife is japanese ... she's into hot springs and onsen :-)
i grew up in a hot springs area of british columbia too...
G-Ray has quit
G-Ray joined the channel
G-Ray has quit
G-Ray joined the channel
Makeitfun joined the channel
Makeitfun
Can apps be Co created to built roleplay app
Nothing yet is adequate
G-Ray has quit
G-Ray joined the channel
Makeitfun has quit
G-Ray has quit
G-Ray joined the channel
millette has quit
millette joined the channel
G-Ray has quit
G-Ray joined the channel
G-Ray has quit
G-Ray joined the channel
G-Ray joined the channel
G-Ray has quit
mafintosh
cblgh: i'm in ireland until thur unfortunately
cblgh
ohh nodeconf rite?
mafintosh
yuh
but another time :)
cblgh
for sure! :3
ilyaigpetrov joined the channel
samm joined the channel
goldbuick joined the channel
laduke-13 joined the channel
samm
Hi everyone, I've been looking into dat protocol recently for an application I'm creating. I want to use it to create social links to users and to coordinate signaling of WebRTC connections. I've written down my implementation ideas and was hoping to receive some feedback.
I'm mostly unsure about how I could coordinate WebRTC signaling using dat. I think it would require multi-user writing to a dat archive which doesn't seem well supported at the moment.