<joepie91>
stuff like "comment in a thread, when displayed in the PR's main conversation view, gives no visual indication whatsoever that it is a reply to anything and just looks like a stand-alone comment on the line itself"
rajivr has joined #nixos-chat
kalbasit has joined #nixos-chat
tilpner has quit [Remote host closed the connection]
tilpner has joined #nixos-chat
kalbasit has quit [Ping timeout: 240 seconds]
kalbasit has joined #nixos-chat
tilpner_ has joined #nixos-chat
tilpner has quit [Ping timeout: 272 seconds]
tilpner_ is now known as tilpner
kalbasit has quit [Ping timeout: 246 seconds]
kalbasit has joined #nixos-chat
Dotz0cat has joined #nixos-chat
Dotz0cat_ has quit [Ping timeout: 256 seconds]
andi- has quit [Remote host closed the connection]
andi- has joined #nixos-chat
Dotz0cat_ has joined #nixos-chat
Dotz0cat has quit [Ping timeout: 272 seconds]
lunc has quit [Ping timeout: 264 seconds]
maljub01 has quit [Quit: maljub01]
lunc has joined #nixos-chat
maljub01 has joined #nixos-chat
Dotz0cat_ has quit [Ping timeout: 260 seconds]
Dotz0cat has joined #nixos-chat
waleee-cl has quit [Quit: Connection closed for inactivity]
<supersandro2000>
lovesegfault: I found me in the nixos tree video
<lovesegfault>
supersandro2000: I've been trying to find myself :P
<lovesegfault>
I saw bbigras' name fly by
cole-h has quit [Quit: Goodbye]
<eyJhb>
Well that esculated quickly!
<samueldr>
just think about how approximately half of issues/PRs have been made in the last year
<__monty__>
Can a single USB-C cable cover all standards-compliant use cases?
<__monty__>
I.e., is the magical future USB-C promised simply not a thing because of cost-savings or are there fundamental issues with the standard that make it impossible?
thibm has joined #nixos-chat
<sphalerite>
supersandro2000: it's so pretty!
<sphalerite>
__monty__: the former, I think.
<joepie91>
__monty__: that seems like a question for Peetz0r
<gchristensen>
you may not like it but that is the ideal let/in body
<patagonicus>
__monty__: https://www.delock.de/produkte/G_84845/merkmale.html?setLanguage=en this one for example claims to support 20 Gbps, which according to wikipedia is the fastest USB 3.2 goes and also does power delivery up to 100W. I'm not really sure what else there is for USB standards.
<Peetz0r>
Yes, those cables exist but they're expensive
<Peetz0r>
patagonicus: thunderbolt 3 or usb 4
<Peetz0r>
look for 40 gbit/s
elvishjerricco has quit [Ping timeout: 260 seconds]
<sphalerite>
gchristensen: what do you mean? It doesn't even have { body = …; }
<sphalerite>
__monty__: that said, I'm fairly pleased with the current reality of USB-C — being able to charge my laptop and connect USB devices and displays via one cable, and most modern laptops working with this, is so much better than anything we had in the past
aranea has joined #nixos-chat
Emantor_ has joined #nixos-chat
<sphalerite>
And I can charge my laptop with a regular phone charger! It won't charge fast, and may only slow draining down if the laptop is powered on while charging, but that's saved me considerable annoyance already.
<gchristensen>
interesting, none of my attached /sys/class/typec/* devices support_power_delivery but I am clearly charging over one of them
<sphalerite>
TIL /sys/class/typec
<gchristensen>
me too
<gchristensen>
in that link from Peetz0r
<sphalerite>
I have nothing in /sys/class/typec even though my laptop has type-C
<patagonicus>
sphalerite: Totally agree! I have a personal laptop + two for work at home. Also two phones, wireless headphones and a tablet. All can be charged with the same charger.
<Peetz0r>
gchristensen: Many usb-c controllers do all these things in firmware and never tell the OS about these things. Especially those in intel thunderbolt laptops
<patagonicus>
Haven't quite gotten the docking for the laptops down to one cable - I want 4k@60Hz display port and I found a KVM that does that directly via one cable, but it doesn't charge. So now I'm using two cables, one for power, one for display + keyboard + mouse + audio interface + webcam.
<Peetz0r>
others can do these things in drivers, and then the OS does (have to) know about it
cosimone has quit [Quit: cosimone]
<sphalerite>
(in case anyone is considering buying such equipment: I'm very happy with my Dell P2720DC monitor, which does charging, USB, and DisplayPort daisy-chaining
<sphalerite>
)
waleee-cl has joined #nixos-chat
thibm has quit [Ping timeout: 256 seconds]
evanjs has quit [Ping timeout: 256 seconds]
evanjs has joined #nixos-chat
<__monty__>
What does "phy" mean in "This is due to advancements in the underlying phy on the host and client-side?"
<lunc>
woot big discount on .tech domains at the moment
<gchristensen>
phy means the physical interface, the silicon
<sphalerite>
__monty__: it's the bit that implements the physical layer of the protocol
<sphalerite>
gchristensen: well, not all of it :p
<gchristensen>
aye
<sphalerite>
optical USB4 when??
<__monty__>
So it's a controller between the port and the CPU?
<gchristensen>
I think the phy might involve the physical routing and connection too? not sure
<sphalerite>
routing?
<__monty__>
The copper tracks connecting the port's pins to the controller?
<__monty__>
What are _Cable-SOP'_ and _Cable-SOP"_? So many hard to search for abbreviations >.<
<sphalerite>
(disclaimer: not much of an expert) at (multi-)gigabit speeds, the communication isn't just a matter of switching a voltage on the wire on and off, but weird magic modulation in high-frequency signals. Like radio but over a wire. The PHY does that modulation stuff, so that the signal can travel over long(ish) distances without completely breaking down
<__monty__>
Analogous to a "modem?"
<sphalerite>
yes
<gchristensen>
perhaps even the same thing? (mo(dulator)dem(odulator))
<sphalerite>
yeah, not sure what the difference between a PHY and a modem is
<gchristensen>
a bunch of people here talking out of their butt :P
<__monty__>
One more thing, how does Power Delivery help with getting info from a cable?
<sphalerite>
Cables need to have identification information to be used for power delivery, so that your cable doesn't catch fire because you shove too much current through it. Devices that don't use PD don't need to care as much about the capabilities of a cable
<sphalerite>
Also, the radio thing was a lie.
<sphalerite>
I think.
<AMG>
bah slack having issues... that complicates work :/
<philipp[m]>
Everything is great! (Except messaging and channels might not like work at all, but we are succesfully shipping all static content! Yaaay us!)
<AMG>
Hehe ageism in IT at 29?! wtf. I have yet to experience that
<joepie91>
oh, there's a lot of ageism in IT
<joepie91>
especially in silicon valley esque circles I can imagine "your thirties" to be a point where you're considered old
<srhb>
Alternatively you can just use Nix everywhere instead of whatever FOTM and you'll be considered old no matter your age!
<AMG>
heh
<siraben>
srhb: FOTM?
<srhb>
Flavour of the month, sorry.
<siraben>
ah
<siraben>
Using Nix would be considered old? haha
<srhb>
Oh, am I the only one that gets a retro vibe for Nix?
<siraben>
I get a futuristic vibe
<siraben>
Imperative package management is the retro way :P
<srhb>
Now I want to make a poll correlating person age and "felt Nix age"
<siraben>
Nix is objectively old, dating to 2003 IIRC
<siraben>
But in that regard, so is Haskell, Lisp etc. which modern languages are playing catchup to
<srhb>
siraben: ouch. ;)
<siraben>
I guess it's because academia? mm
<gchristensen>
old languages are playing catch-up to new languages, too
omnd has quit [Ping timeout: 264 seconds]
<gchristensen>
new languages' compilers/evaluators have different architecture and ideas about parsing, tracking errors, developer tooling, and they're able to integrate these newer ideas in to their earlier development, vs. an approach of massive refactor/integration work or bolt-on
lunc has joined #nixos-chat
<gchristensen>
there was a great video by Microsoft about the old model of compilers vs. the newer model, it was really good -- I wish I could find it
<srhb>
gchristensen: Completely off topic, how much do you remember still about docker layers and their tar representation? Specifically, do you know if there's something akin to a specification of unpack behaviour for them?
<gchristensen>
I'm sure there is in the OCI Image Spec ...
<siraben>
gchristensen: if you manage to find that video I'd be interested
<srhb>
gchristensen: Looks very helpful, thanks!
<gchristensen>
you're welcome!
<gchristensen>
understand though that since all my work was on the /nix/store, I literally never needed to consider anything but additions
<gchristensen>
I worked with a team whose sys people decided to implement their own runtime though, which was a hilarious exploration of spec deviation
<srhb>
gchristensen: I'm trying to determine what to do with a buildLayeredImage bug, (or rather, streamLayered...) and need some spec guidance to decide :)
<gchristensen>
ah ha ...
<srhb>
gchristensen: Custom runtime sounds fun though :D
<gchristensen>
fun for the team to implement, yes
<gchristensen>
as a user .......
<srhb>
Yes, sure :P
<gchristensen>
as a user I felt like everyone forced in to using nix: why can't you just use something normal???
<srhb>
Relatable. :D
<joepie91>
<gchristensen> there was a great video by Microsoft about the old model of compilers vs. the newer model, it was really good -- I wish I could find it
<joepie91>
please let me know if you do...
<srhb>
It's weird being the crazy minority. I mean, you get used to it, but... Sure get used to not invoking wisdom of the masses.
<joepie91>
srhb: a lesson I am still trying to more generally teach people over in #Node.js..
<joepie91>
many years later
<eyJhb>
srhb: Also note, you cannot be 100% sure that Docker follows the spec
<eyJhb>
They are weird.
<srhb>
I think this is just a case that isn't mentioned explicitly, but I can't think of a reasonable way it could be legal.
<srhb>
eyJhb: Note that the store path is both a symlink and has entries
<srhb>
As if it were a directory
<srhb>
But not sure yet :)
<srhb>
basically dir is dangling with no parent.
<srhb>
philipp[m]: Everything is marked outage now at least :P
<srhb>
I wonder where the Slack team discusses solutions...
<eyJhb>
404 srhb
<srhb>
eyJhb: ...9 :P
<eyJhb>
:D
<eyJhb>
Is slack down? Also, the Docker implementation we have in Nix, is nice but it functions in WEIRD ways, and will override a bunch of things if you use outside images.
<srhb>
Yes, it is.
<srhb>
(slack down, that is)
<eyJhb>
:( Wanted to use it for my platform... But sadly I can't at this stage. Hope I can have some time to spent on it soon. Then I can write it off at LEGO :p
<eyJhb>
I just need to pass all my exams first, before I can dream of such things. But yeah, the gist does look weird
<adisbladis>
eyJhb: Seems like slack is working again
<andi->
noes /o\
<andi->
status page is still red and I'm getting "service unavailable" on the website..
<eyJhb>
Honestly, it would be a little nice if Slack just stayed down
<eyJhb>
My Bitlbee integration with Slack still seems to be up
<eyJhb>
Still was a little nickname(Richard) to take down the IRC bridge...
<cransom>
my weechat slack plugin was still working, but i'm sure i'm the only one using it, so no one else to talk to there...
<srhb>
Seems up again here, though severely overloaded. :)
<eyJhb>
cransom: But shouldn't that also mean that the desktop apps should work?
<eyJhb>
Or are they just a browser?
<cransom>
they are all electron apps, afaik. i don' tknow how they operate differently.
rajivr has quit [Quit: Connection closed for inactivity]
<__monty__>
Hmm, had a thumbdrive where I couldn't write any files or move files on it around unless I used sudo. Reading was fine. Permissions all looked fine.
thibm has joined #nixos-chat
cosimone has joined #nixos-chat
<bbigras>
I think I had something similar the other day
<__monty__>
bbigras: As in a device you couldn't write to?
<bbigras>
__monty__: yeah and I had to use sudo. If I remember correctly.
cole-h has joined #nixos-chat
<__monty__>
I'm starting to doubt myself, maybe I took away write permissions? But there was some data corruption too, so I doubt it.
<__monty__>
Doubt-ception, is this how insanity starts?
<bbigras>
yeah probably
endformationage has joined #nixos-chat
<energizer>
does anybody sandbox cli apps, so e.g. grep doesn't have write access to filesystem?
<andi->
I would buy into that.
<bbigras>
me too. Is apparmor good enough for that? Ubuntu uses it, right?
<energizer>
there are a bunch of options
cosimone has quit [Remote host closed the connection]
cosimone has joined #nixos-chat
cosimone has quit [Quit: cosimone]
cosimone has joined #nixos-chat
<infinisil>
Are there hosting providers where you only pay for the CPU time used?
<gchristensen>
AWS has something adjacent to that with its t* series, where you get burstableCPU credits
<eyJhb>
*sigh* no more space on my VPS
<eyJhb>
Just freed up 43 MB, and then it ate it all - `error: executing SQLite statement 'delete from ValidPaths where path = ?;': database or disk is full (in '/nix/var/nix/db/db.sqlite')`
<lukegb>
I have a, heh, exciting Unifi controller derivation
<lukegb>
(takes the nixpkgs.unifi, deobfuscates it a bit, applies some aspectj stuff to inject my own code into it, reobfuscates it...)
<ma27[m]>
eyJhb: yeah, nix-collect-garbage IIRC writes stuff into its db. Can you remove some other stuff temporarily?
<eyJhb>
ma27[m]: I tried... I could only get 43 MB , and it all it all
<ma27[m]>
I'm talking about stuff not related to Nix to gather some free space.
<sphalerite>
lukegb: that sounds very legal :p
<lukegb>
sphalerite: I have no qualms with doing this to a company that can't respect the GPL :p
<eyJhb>
ma27[m]: yup, I have nothing on it :p It is a VPS with 12 GB of space