jackdk has quit [Quit: Connection closed for inactivity]
pie_ has quit [Ping timeout: 245 seconds]
<Church->
Heya folks
<ldlework>
hi Church-
jackdk has joined #nixos-chat
<drakonis1>
hey its church
pie_ has joined #nixos-chat
endformationage has quit [Ping timeout: 268 seconds]
drakonis2 has quit [Ping timeout: 245 seconds]
drakonis1 has quit [Ping timeout: 252 seconds]
drakonis has joined #nixos-chat
drakonis_ has joined #nixos-chat
drakonis has quit [Ping timeout: 250 seconds]
jackdk has quit [Quit: Connection closed for inactivity]
<MichaelRaskin>
Replacing products with significantly different ones under the same name should lead to loss of trademark. It doesn't fulfill the function of specifying the goods being sold, so why bother. I guess Apple would keep the iPhone line (it is numbered) but just lose anything related to Macbooks…
<MichaelRaskin>
(my girlfriend is trying to install Ubuntu on an _old_ Macbook Air and I am trying to check if there was ever a NixOS configuration for that specific generation with working WiFi and of course it is impossible)
<sphalerite>
MichaelRaskin: hmm, bcm4331?
<MichaelRaskin>
No, Broadcom 4321
<sphalerite>
The broadcom chips in apple laptops were the bane of my existence for a while :p
<sphalerite>
I wasn't using nixos at the time, but all the options for using it with linux were awful
<MichaelRaskin>
I guess if I had the device physically near me I would already have installed ndiswrapper
<MichaelRaskin>
Well, old Broadcom in non-Apple hardware was no better!
<sphalerite>
heh, fair enough
<MichaelRaskin>
ndiswrapper was more of a one-time setup though
<MichaelRaskin>
«Why are we opposed to Broadcom-Qualcomm merger? Because they will manufacture horrible crap like Broadcom and patent-troll like Qualcomm.»
<sphalerite>
oh yeah, I was wondering what you think about surf? I thought it might fit into your setup better than firefox
<MichaelRaskin>
I was using Uzbl with a custom event manager until an event happen
<MichaelRaskin>
Which was unrelated to surf or Uzbl, technically speaking…
<MichaelRaskin>
Can you guess what the event was?
<MichaelRaskin>
gchristensen: re: notifications from Packet: subscribe to all GitHub notifications for Nixpkgs and learn to skim-and-filter these!
<sphalerite>
MichaelRaskin: hm, no?
<MichaelRaskin>
Opera abandoning their own rendering engine
<MichaelRaskin>
At that point I decided that I do not touch WebKit/Blink «IE6 2.0» if I can avoid it
<sphalerite>
aah ok
<MichaelRaskin>
Actually, I don't use Firefox _that_ much
<MichaelRaskin>
Most of the stuff gets grabbed as HTML-only and converted to txt and viewed in GVim
<MichaelRaskin>
_Some_ gets grabbed by JS-enabled FF, then converted to txt…
<MichaelRaskin>
And probably most of the stuff gets autofetched in advance and without FF
<MichaelRaskin>
The stuff that gets treated as applications… well, there is not too much that is needed from a browser UI there anyway, and my FF is styled to be a big rectangle of web content with browser UI hidden.
<sphalerite>
ah, sounds good :D
jackdk has joined #nixos-chat
pie__ has joined #nixos-chat
pie__ has quit [Remote host closed the connection]
pie_ has quit [Read error: Connection reset by peer]
pie__ has joined #nixos-chat
pie__ has quit [Remote host closed the connection]
pie__ has joined #nixos-chat
pie__ has quit [Remote host closed the connection]
pie__ has joined #nixos-chat
<sphalerite>
I seem to recall there being something wrong with the Brave browser, joepie91 I suspect you would know things?
* joepie91
rolls in
<joepie91>
tl;dr it is a thoroughly unethical company
pie__ has quit [Remote host closed the connection]
<joepie91>
their two business models so far have been 1) replace other people's ads on sites with their own, and 2) solicit "donations" on behalf of people without their consent, and then pocketing them if gone unclaimed for X time
<joepie91>
and now they are getting involved in the cryptocurrency nonsense too
pie__ has joined #nixos-chat
<joepie91>
sphalerite: ^
<sphalerite>
joepie91: perfect, thanks.
<joepie91>
on that note, I would generally recommend treating a given project getting involved in cryptocurrency stuff, as a red flag :) it highly correlates with other general shitbaggery
<sphalerite>
Thanks in general for always knowing and caring about this kind of stuff <3
<joepie91>
not *every* such project is unethical, but *almost* every one is
<joepie91>
:)
<joepie91>
sphalerite: I appreciate that somebody values it :P
<joepie91>
nice change from the usual people trying to automatically trivialize everything / throwing snark at somebody caring about this / etc.
pie__ has quit [Ping timeout: 258 seconds]
<infinisil>
joepie91: What do you think of Filecoin?
<joepie91>
infinisil: I have no reason to take it seriously - far as I can tell, they're still not actually addressing the perennial problem of decentralized storage accounting, namely the ability to reliably prove that a peer actually *stores* data, without needing a local copy or having to transfer the entirety of the data for every check
<joepie91>
to my knowledge, this is more generally an unsolved problem
<joepie91>
and all of the various blockchain-storage-token-y things are just rearranging deck chairs on the titanic and conveniently ignoring that problem
<MichaelRaskin>
joepie91: I think you are shifting a problem one step
<MichaelRaskin>
It is easy to prove that a counterparty stores at least one copy of data
<joepie91>
and considering that IPFS *still* hasn't solved their spec/docs/reference-implementation issues, and all their attention now seems to go towards Filecoin... well.
<joepie91>
MichaelRaskin: not without a local copy.
<joepie91>
(and bandwidth usage constraints)
<MichaelRaskin>
Even without a full local copy
<joepie91>
okay, how?
<MichaelRaskin>
Pretty standard Merkle tree
<joepie91>
no, I'm asking concretely: how? what is the full process?
<joepie91>
the verification process, that is
<MichaelRaskin>
I know the root and I ask to walk a random branch
<joepie91>
because with the obvious implementation using a merkle tree, you have two options: a) ask the peer for a hash and trust it (which means they can just store the hash and not the data) or b) ask the peer for the data and hash it against your local tree (which means you have to fetch every bit of data every time you want to verify it)
<MichaelRaskin>
I store the root.
<joepie91>
yes, but the question is what the *peer* stores.
<MichaelRaskin>
At each step I ask for all the data hashing to it
<joepie91>
okay, so you're asking for the full data, which is B
<MichaelRaskin>
I go down the entire branch to real data blocks
<MichaelRaskin>
I do not ask _entire_ data
<joepie91>
then what *do* you ask for?
<MichaelRaskin>
I ask for a single block of data and a branch towards it
<MichaelRaskin>
Which is logarithmic instead of linear
<joepie91>
this would still be too high-bandwidth to do granular accounting with, like blockchain-y storage-token-y things claim to do
<MichaelRaskin>
Actually no, there is a harder problem there
<joepie91>
anyhow, before this becomes an extensive discussion: I was not planning on a long discussion about this today, as it is my weekend and I am trying to relax :)
<MichaelRaskin>
This is high bandwidth, but it also does nothing to detect collusion between peers who pretend to store data independently
<MichaelRaskin>
(well, you could store encrypted data, but this kills the desired application of storing publically valuable content in a replicated way)
<sphalerite>
hm, thought on that problem: could the "client" hash the data with various "salts", then later request that the "server" return the data hashed with one of the salts?
<sphalerite>
It only allows a limited number of verifications, but it should work, no?
<MichaelRaskin>
Doesn't help, the server can rehash
<MichaelRaskin>
If the server has one copy of entire file, it can pretend to have more
<sphalerite>
yeah not for the multiple-copies bit
<MichaelRaskin>
Maybe storing the top parts of Merkel trees, which are useless
<sphalerite>
I meant for the first problem
<MichaelRaskin>
Single copy is easy
<MichaelRaskin>
I mean, you can fit around ten levels into a single MTU
<MichaelRaskin>
Then a kilobyte block in a second one
<MichaelRaskin>
Of course, in a blockchain scenario you need to _also_ verify if the prover and the verifier collude…
<sphalerite>
for multiple copies my guess would be that it's provably impossible
<sphalerite>
(just a guess though)
<sphalerite>
(you never know what crypto[graphy] experts will come up with next!)
<MichaelRaskin>
I do know for a fact they had not yet in 2015
<MichaelRaskin>
I mean, if you care for a file and nobody else does, you can just encrypt it differently for each server
* manveru
just assumes smarter people already bashed their head against this :)
yorick has quit [Quit: brb]
<MichaelRaskin>
Rest assured they did it in a way with completely messed-up incentives
<manveru>
anytime there's VC money involved, i basically ignore it
<manveru>
it's not worth waiting for the eventual implosion
<MichaelRaskin>
I actually _also_ include academia incentives
<manveru>
heh
<MichaelRaskin>
Which are messed up in a slightly different way, but also often messed up
pie_ has joined #nixos-chat
pie_ has quit [Remote host closed the connection]
pie_ has joined #nixos-chat
<colemickens>
I'm so curious about GuixSD but so unfamiliar with scheme. Just saw GNU Mes linked from a podcast, seems really neat. I'd love to read a "GuixSD for NixOS users" type article
<joepie91>
[15:17] <manveru> anytime there's VC money involved, i basically ignore it
<joepie91>
++
<joepie91>
(applies to ~everything, not just decentralized tech stuff)
endformationage has joined #nixos-chat
<eyJhb>
Seems like my little project from yesterday is nearing completion!
<MichaelRaskin>
Oh?
<joepie91>
eyJhb: did you see my over-engineered parsing code? :P
<eyJhb>
joepie91: yeah :p That was quite some haha :D
<eyJhb>
MichaelRaskin: basically parsing a bad website in Aalborg, and showing the data in a better and easier way
<MichaelRaskin>
Ah, I remember that
<joepie91>
eyJhb: feel free to reuse, if it's of any use :P WTFPL/CC0 license as usual
<joepie91>
hrmmm. I think I need to replace the foam in my elbow pads...
<joepie91>
(of my desk)
Guanin has joined #nixos-chat
<ajs124>
sphalerite: there's actually also a patch against coreutils... not sure if the README mentions _that_ though
cjpbirkbeck has joined #nixos-chat
<Guanin>
Hi, is someone here who knows how to setup a yesod project on NixOS? Most tutorials uses yesod init, which is outdated, as stack should be used, but stack2nix fails with NoHaddockTarball and other approaches fail with a missing zlib (complete newbie regarding haskell tooling and somewhat newbie regarding haskell, but I struggle to keep exercising if it is not some usable project)
<joepie91>
eyJhb: I would recommend moving the 'search' button to the top rather than the bottom, and calling it 'apply filters' or something - I was very confused for a bit why my filters didn't change anything :)
pie_ has quit [Ping timeout: 258 seconds]
pie_ has joined #nixos-chat
<gchristensen>
I realized the mbuffer path to fix znapzend and it immediately used all my bandwidth annd the airport kicked me off wifi :')
<joepie91>
lol
pie__ has joined #nixos-chat
pie_ has quit [Ping timeout: 248 seconds]
cjpbirkbeck has quit [Quit: Quitting now.]
Guanin has quit [Remote host closed the connection]
pie___ has joined #nixos-chat
pie__ has quit [Read error: Connection reset by peer]
<eyJhb>
joepie91: will do!
<eyJhb>
Or, creating a issue so I will remember it
<eyJhb>
Anybody here have experience with Gitlab Pages? Holy hell it is SLOW
<Ralith>
that is my experience
<Ralith>
slow and prone to idiosyncratic bugs
<eyJhb>
Really not impressed by that at least... :/ Seems like it is WIP, but still, what...
<eyJhb>
Anyone up to coding a PR for them? ;)
drakonis_ has quit [Remote host closed the connection]
<MichaelRaskin>
Is it code or deployment issue, though…
pie___ has quit [Ping timeout: 272 seconds]
drakonis has joined #nixos-chat
<eyJhb>
It is somewhat both, and also me not naming the folder `public/` specifically.......
<eyJhb>
Still does not work though :( Might take some time for it to be active, but.. Shouldn't
<joepie91>
eyJhb: I would recommend a second line just for those numbers, color-coded for easy scannability, small text, brief label for each number
<joepie91>
that is:
<joepie91>
[__] label name goes here
<joepie91>
second small print line with numbers goes here
<joepie91>
[__] next label name goes here
<joepie91>
and I don't know what the numbers mean :D but the second line might look something like: _112_ foo, _0_ bar, _11_ baz, _4_ qux
<joepie91>
with __ denoting bolding
<eyJhb>
Adding a `about` pane, the explains it. But it is basically.. Confirmed to have facility/Confirmed NOT to have facility/Have description about facility/Have nothing about facility
<joepie91>
eyJhb: an explanation pane just requires people to cross-reference more stuff, it's more usable to have it inline
<joepie91>
maybe something like "112 have this, 0 do not, 11 have description, 4 do not"
<eyJhb>
It fills quite some to do that too
<eyJhb>
Even though I agree
<joepie91>
eyJhb: that's why I suggested the second small-print line! :)
<joepie91>
if you reduce the margin between the two lines (which is fine for related info like this), it won't take up much extra space
<eyJhb>
You are pushing me to do UX design joepie91 ! You should know by now I hate doing anything design related :D
<eyJhb>
I will look at it, but god do I hate anything with that
<joepie91>
eyJhb: hey, I already handed you the solution! :P
<joepie91>
now it's just a matter of implementing it!
<joepie91>
eyJhb: yeah, though I'd increase the indentation of the second line to match the text of the first line (easiest to do by floating the checkbox with a margin that makes it span two lines), and color-coding each item in that second line for easy scanning; like light green, light red, light blue, light orange
<joepie91>
just enough color tint to make it easily recognizable
<joepie91>
then you don't have to mess with aligning the items themselves on that line
<eyJhb>
*inserts 4
<eyJhb>
:D
<MichaelRaskin>
joepie91: implementing design _is_ related in design, and therefore covered in the Hatred Claim
<joepie91>
NOT GOOD ENOUGH!
<joepie91>
:p
<joepie91>
MichaelRaskin: lol
<eyJhb>
Color tint?
drakonis has quit [Quit: WeeChat 2.4]
<joepie91>
eyJhb: do you have the current version online somewhere?
<joepie91>
I can just show it to you :P
<joepie91>
(ideally with a different class name for each item in that second line)
<eyJhb>
Nope, only locally :p
<eyJhb>
And I do not trust Gitlab Pages to push within 10 hours
<MichaelRaskin>
I bet if the four descriptions were in Danish, they would be shorter
<pie_>
so i got some used cheap samsung 850 evos
<pie_>
total LBA written: 5358945088 , power on hours: 11000
<pie_>
hows that?
<pie_>
sector size 512 so...
<eyJhb>
MichaelRaskin: 67 har det her, 59 har ikke det her, 0 har beskrivelse, 1 har ikke
<eyJhb>
Not quite sure
<pie_>
so 3 terabytes of traffic on it
<pie_>
lassulus, stats on one ^
<joepie91>
eyJhb: yes, I did say that :D
<joepie91>
but, nice :)
<MichaelRaskin>
eyJhb: isn't that «her» optional, though?
jasongrossman has quit [Ping timeout: 258 seconds]
<eyJhb>
MichaelRaskin: well, I guess, but it seems more incomplete in danish when omitting it :/
<eyJhb>
Also, everything except the checkboxes are english. The only reason for danish checkboxes is basically.. The site I am scraping from has no english option
<MichaelRaskin>
Pretty typical for rental condition options to have standard local name and no standard translation
* joepie91
repeats: rental sites are awful
<joepie91>
:P
<pie_>
lassulus, behaving hella weird when i plug it in tho..
<eyJhb>
MichaelRaskin: yeah.. but most be quite annoying for none-danes... Loads of Google Translate I guess
<joepie91>
eyJhb, hey, hey eyJhb, I have a suggestion...
* joepie91
runs
<eyJhb>
Haha, you can create issues!
<joepie91>
:P
<eyJhb>
I should not have said that... Like a trillion issues tomorrow :p
<joepie91>
hehehe
<MichaelRaskin>
eyJhb: learning to _read_ Danish wasn't too painful.
<joepie91>
eyJhb: I assume that this just needs to be Good Enough, not perfect? :P
<eyJhb>
MichaelRaskin: speaking it? Wait.. Do you live in Denmark?
<eyJhb>
joepie91: pretty much... It is already better than the original site, so :p
<MichaelRaskin>
I had a postdoc there for year and a bit (Aarhus)
<joepie91>
eyJhb: alright, I won't bother you with more issues then :P
<MichaelRaskin>
Now I am in Germany
<eyJhb>
joepie91: original site: https://vl.aku-aalborg.dk/Forside.aspx , but I am open for any suggestions. My next thing is basically changing the marker icons, and fixing the about. And then it's good-to-go
<eyJhb>
AHh I remember! Sorry MichaelRaskin :D
<joepie91>
"but I am open for any suggestions" -- are you sure about that? :P
<MichaelRaskin>
Denmark is definitely a nice place
<MichaelRaskin>
joepie91: being open to suggestions and promising to actually react to them is different!
<eyJhb>
Yeah it is, and Aarhus too! Originally from Aarhus
<eyJhb>
*issues that will never get closed :D
<joepie91>
eyJhb: 'number of rooms' is ordered wrong, price selection could be improved by having a slider in addition to the input boxes (ideally bounding to the min-max prices in existence in your dataset), instead of an 'apply filters' button you could have them apply automatically after not changing the filters for 500ms or so
<MichaelRaskin>
* never, not even if actually fixed/implemented
<joepie91>
eyJhb: if you have price data, you could display it in the detail popup boxes
<joepie91>
eyJhb: also, you could color-code the markers by how cheap <-> expensive a given location is
<joepie91>
(again, on the scale of prices in your dataset)
<MichaelRaskin>
It's also cool how in Copenhagen they patiently write on the «do not feed ducks» sign all the undesirable consequences of feeding the ducks…
<lassulus>
pie_: weird in what way?
<joepie91>
eyJhb: bonus points for actually marking the buildings as rectangles or w/e based on their address, although I don't know whether google maps has the necessary data for that - I know that openstreetmap does, though
<pie_>
lassulus, my machine probably being dumb, automounting, throwing weird software errors in kde, smartctl very slo
<pie_>
joepie91, for some reason it shows up as the first result for haskell documentation
<eyJhb>
MichaelRaskin: not uncommon for us to do that. We should have more signs at Aalborg saying not to throw bicycles into the Harbor, consequence, uni students who are procrastinating might spend the day fishing them out of the harbor... :D
<pie_>
for some things
<eyJhb>
joepie91: color coded prices are somewhat off, as I want to use the color codes for how hard it lives up to the requirements :D
<joepie91>
pie_: their search feature is surprisingly pleasant for such a weird design
<eyJhb>
But night night ;)
<pie_>
joepie91, i cant find an about page or anything
<joepie91>
eyJhb: you could do two color codes! bulk of the marker used to indicate how well it matches requirements, and a bar on the side or whatever for price :P
<MichaelRaskin>
Looking at the name — so the site name is _supposed_ to mean «ringing (as in, ringing of the bell)»?
<joepie91>
pie_: I've run across a few of these sort of sites throughout the years - experimental ways of presenting/indexing information, often with really weird/old designs like this, but decades ahead in terms of information presentation and such
<joepie91>
they're super interesting
<joepie91>
sec, lemme try to find this other weird thing
<MichaelRaskin>
I would say that being good at information preservation is nowadays decades _behind_
<joepie91>
grmbl, I can't find it. it was some weird multi-column interface, each column showed one text, and a sort of 'overlay slider' that you could use to magnify a particular part, and follow references
<joepie91>
was called hypertext something, I believe
<MichaelRaskin>
Multicolumn?
<joepie91>
I'm going off memory here, so it's difficult to describe
<MichaelRaskin>
Sounds like (among other things) smallest federated wiki
<joepie91>
:p
<joepie91>
wasn't a wiki iirc
<MichaelRaskin>
Might also be something from Xanadu?
<pie_>
joepie91, wanted to try treesheets at some point, didnt get around to it
<pie_>
joepie91, clicked links in tiddlywiki, nothing interesting happened_
<pie_>
?
<joepie91>
pie_: it should load all wiki links as new 'cards'
<pie_>
oh looks like i had to enable js and reload
<joepie91>
inbetween existing cards
<joepie91>
oh, yes
<pie_>
ended up on the static site by accident
<joepie91>
anyway, it is scarily good at placing the new card in a logical position, and in moving the viewport when you open/close cards in a sensible way
<pie_>
joepie91, ok this seems rather appealing
<joepie91>
and curiously, it can work entirely as a self-contained HTML file
<joepie91>
without a server
<joepie91>
when you 'save' it, it generates a new version of itself
<pie_>
i need to look at tis again later
<joepie91>
(though it can also work with a server iirc)
<joepie91>
hehe
<makefu>
joepie91: i really like tiddlywiki, i am using it as my personal wiki, also stored on http://wiki.euer.krebsco.de
<pie_>
made friends with someone with some interesting window manager ideas
<joepie91>
makefu: neat :)
<pie_>
(basically save moar context tho)
<joepie91>
pie_: you have my attention
<pie_>
joepie91, uhh maybe i shouldnt have said anything xD
<joepie91>
hehe
<joepie91>
makefu: you running it as a self-contained file, or as a server?
<pie_>
joepie91, if you see a guy named ilmu show up ask him
<joepie91>
name doesn't ring a bell
<joepie91>
but anybody who is unhappy with the current state of window managers, is a friend of mine :P
<joepie91>
in unrelated news
<joepie91>
the rust-embedded WG will be moving over to Matrix
<pie_>
why do people keep wanting to move away from irc
<joepie91>
it's a surprisingly decisive victory
<pie_>
isnt it like the best thing ever? :P
<pie_>
i mean irc
<joepie91>
pie_: well, IRC is... not great
<joepie91>
lol
<joepie91>
and unlike previous attempts at replacing it, Matrix actually has the potential to do so, I think
<MichaelRaskin>
joepie91: why are you unhappy with modern window managers? You just install StumpWM or XMonad (whatever your language preference is), then try to codify your «I wish it would now do …» i to configuration
<joepie91>
MichaelRaskin: that is a whole topic unto itself
<MichaelRaskin>
(default settings are indeed kind of silly)
* joepie91
is off to bed
<elvishjerricco>
Anyone know of a dumb, usb connected, 5 drive enclosure that just shows the computer five drives; i.e. doesn't do any RAID?