gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
<gchristensen> sigh.
<gchristensen> I put the wrong mount options in and now my system is most unhappy
<cole-h> :(
<infinisil> Alright I think I understand flakes a bit more now, reading the RFC also helped hah
<infinisil> Pretty neat, I'm liking it
<infinisil> Currently haskell.nix does many levels deep of IFD, and I think some of it could be alleviated with flakes
<energizer> it would be nice to have a standardized test suite that would try to boot, start display manager, log in, check networking
<energizer> i dont know how to do it
<cole-h> I think the gnome3 test does that. Could draw inspiration from it.
<bqv> i love that nix can have tests as comprehensive as that
<bqv> infinisil: sounds great, when i briefly used haskell.nix it was tiring to download gcc861 repeatedly
<cole-h> omg new feature preview for commit design looks so much better. I can actually see the commit message
<cole-h> Actual improvement
<gchristensen> oh thank you
<bqv> oh yeah, i activated that and didn't even notice
<infinisil> bqv: Using the iohk cachix is really necessary with haskell.nix unless you want to build everything yourself
<infinisil> Possibly that was causing it
<bqv> huh, i thought i was using it
<bqv> maybe i have a dodgy overlay causing rebuilds
<neeasade> bqv: at the least, our conversation earlier today helped remind me to use dired more for navigation, and it is indeed a pleasure
<bqv> ^_^
* gchristensen edits /nix/store
* bqv edits /gchristensen/brain
<gchristensen> that is how I feel
<cole-h> gchristensen: Is your system still fscked? :^(
<gchristensen> yea
<gchristensen> hopefully editing the nix store to not mount one of these disks will fix it
<gchristensen> I *think* it works to just edit the fstab ...
<cole-h> gchristensen: Ouch :(
<gchristensen> my own fault for not having any spare drives around
<gchristensen> ...and not buying a replacement for like 3 weeks .....
<gchristensen> ......and then rebooting the system...
<cole-h> What actually happened?
<gchristensen> a disk went corrupt
<cole-h> Ouch
<gchristensen> it came up!
<gchristensen> I guess it just had to get over itself
slack1256 has quit [Remote host closed the connection]
iqubic has joined #nixos-chat
iqubic` has joined #nixos-chat
iqubic` has quit [Client Quit]
iqubic has quit [Ping timeout: 260 seconds]
iqubic has joined #nixos-chat
<energizer> is there a way to hide a file so it doesn't appear in dir listings?
<energizer> (without writing my own fuse)
<samueldr> prefix with a dot (probably not what you want) or install a rootkit
<samueldr> rootkit-like thing that can be used https://gobolinux.org/doc/articles/gobohide.html
<energizer> cool, never seen that before. thanks
<samueldr> it irks some people apparently
<samueldr> this is pretty bad packaging work that I should re-do and review https://gitlab.com/samueldr/nixos-configuration/-/tree/master/overlays/gobohide
kalbasit has quit [Ping timeout: 272 seconds]
drakonis has quit [Read error: Connection reset by peer]
cole-h has quit [Quit: Goodbye]
parsley936 has joined #nixos-chat
* eyJhb *shoo shoo samueldr *
<samueldr> tbf, it's one of the first things I packaged and... as nix is quite amazing, I never really had to touch it ever again
<samueldr> so, there you have it
<samueldr> a major downside of nix
<samueldr> you don't ever have to fix your bad early work
<eyJhb> Did you come directly from a Mac?
<samueldr> nope
<eyJhb> Why would you then do something like that to yourself? :(
<samueldr> never really used a mac
<samueldr> why not?
<eyJhb> Especially the uppercase!
<eyJhb> Damn
<samueldr> why not??
<eyJhb> You got me there
<samueldr> it's perfectly cromulent
<samueldr> now, if I had it as "/Documents and Settings"
<samueldr> I'll tell you one reason to do it: you can make sure there's no bad assumptions that /home/$USER is a user's home
<eyJhb> Simpsons fan?
<eyJhb> True
<samueldr> not anymore, but I'm still sticking to the naming scheme
<eyJhb> Have you experienced that before?
<samueldr> I don't think I ever had
<samueldr> (experienced that)
<samueldr> but the main reason: it irks some people
<ar> samueldr: no no, it should be /Ðóćümęńẗś äńd Śęẗẗïńgś
<eyJhb> I am still more scared of programs that run `sudo ${foo}/${bar}`
<eyJhb> :(
<eyJhb> ar: don't give ideas! :(
<ar> to properly trigger all the non-unicode-clean programs
<samueldr> and the naming scheme, it's been one I'm using since ~2004, so it's kinda neat I think that I'm still using it
<worldofpeace> at long last, I've cleared 400 notifications. soon, there will be peace
<ar> worldofpeace: "mark all as read"?
<eyJhb> worldofpeace: my bank just released a new UI, where you cannot go back in you "old" messages to see them. This includes the 46 unread messages I have. Now I have 6 unread I cannot read
<worldofpeace> ar: hehe, I marked as read the one's that looked very no op. But the rest I actually read
<worldofpeace> eyJhb: so that means any messages made in the old ui are just unreadable?
<eyJhb> Pretty much. Couldn't approve payments either, load accounts list, etc...
<eyJhb> Luckily they still provide the old UI as fallback. But holy hell the new one is straight up shit *sorry for the language*
<worldofpeace> I don't even know how someone would write off on that even being made available to the user for all things banking
Arahael has quit [Ping timeout: 256 seconds]
<eyJhb> No clue either, it seems so so bad... Worst part is, it is my bank for my company... I wold expect them to be a lot better, when dealing with such things
Arahael has joined #nixos-chat
<eyJhb> Btw. worldofpeace up for a little PR?
<worldofpeace> lol, maybe just one eyJhb 😸
<bqv> samueldr: but if there are bad assumptions, you end up with a broken app
<bqv> i do often want to shake up what remains of the fhs on nixos though, too
<worldofpeace> I mean, might as well before I fade into some mist
<eyJhb> worldofpeace: #90307 just anything on it would be great. Currently run a setup where I have to apply this patch to my nixpkgs :p
<{^_^}> https://github.com/NixOS/nixpkgs/pull/90307 (by eyJhb, 3 weeks ago, open): boot.initrd.luks.devices: add preOpenCommands and postOpenCommands
<eyJhb> And I know how much you love all my PRs <3
<worldofpeace> eyJhb: thanks, I merged. I saw no problem just having that now
<eyJhb> Perfect! Thanks! I wish for more people like you. I have had a PR open for Docker since May... :(
<eyJhb> <3 worldofpeace
<{^_^}> worldofpeace's karma got increased to 175
<worldofpeace> thanks eyJhb 💓 Maybe try to hang out in the nixos (unofficial discord if you can) jonringer is probably another one to look out for
<worldofpeace> and I think in big projects, people mind getting pinged less
<worldofpeace> erm, they don't mind getting pinged really
<eyJhb> I hope so! I just hate continuously pinging people
<eyJhb> Just sent you a PM worldofpeace :)
<bqv> argh, how do you ltrace a setuid program
<bqv> wow that went spectacularly badly
<bqv> i was going to say, solution was to run it as root
<bqv> but i think that was a mistake
<bqv> my display crashed, and now i have three mice and lots of screen corruption
<eyJhb> bqv: three times the productivity?
<eyJhb> I think I have the worst rebase atm..
<eyJhb> This is not smart or fun
<eyJhb> Looks like it has just eaten parts of the code
<bqv> :D
<bqv> when the rebases get that hairy, i just settle for a merge
<bqv> ok this is starting to get silly, i'm recompiling some code that gets used in a library, and seeing none of the edits at runtime
arahael1 has joined #nixos-chat
Arahael has quit [Ping timeout: 272 seconds]
<eyJhb> bqv: this is such a mess either way
<eyJhb> Sad that I have not kept it up-to-date
<eyJhb> I think making a new branch a literally just copying over the files/packages is more efficient
<eyJhb> I have a BUNCH of untracked files now. damn it
<bqv> Haha, I've done that before
<bqv> Basically a soft reset
<eyJhb> Damn it
<eyJhb> Merged with the wrong branch
<eyJhb> :(
<bqv> Oh!
<bqv> Wait, nevermind
<eyJhb> Now everything is OK, but... There is sooo much rewrite as always
<eyJhb> bqv: does stuff somewhat work?
<bqv> No I almost thought what I was trying was impossible
__monty__ has joined #nixos-chat
<bqv> Maybe it is, I have a crazy error :D
<eyJhb> That is at least something!
<eyJhb> I just have a lot to refactor...
<eyJhb> :(
polezaivsani has joined #nixos-chat
<polezaivsani> i've been snowflake-blind for a while yet still lust for some critical thinking - can somebody here shed some light on the shortcommings of the nix approach in general? any ailments that don't get enough attention?
<__monty__> UX
<energizer> what does snowflake blind mean?
<energizer> the main principles of nix are good and make very usable systems. some facts of the 2020 implementation are annoying
<polezaivsani> energizer: that's my ref to the nix logo and being generally positively affected by the approach, so much so that i might become less objective
<energizer> ah
<polezaivsani> curious about devil's advocate take on it ;)
<energizer> are you asking for "what is bad about nix" or more "what is bad about immutable package management and system definition"?
<polezaivsani> energizer: i'd curious in any of it frankly
<polezaivsani> since i don't know of much 'bad' aspects about any of it
<energizer> the type system for example
<energizer> security/privacy (world-readable nix store)
<__monty__> Some software is hard to package and FHSUserEnv doesn't always help.
<energizer> rebuilding/activating nixos takes >30s which is way too long for interactive configuration-change cycles
<energizer> PRs often take quite some time to get merged (but i dont have numbers on that, itself a problem)
<energizer> docs are ..that way
<polezaivsani> energizer: i'm new here, but what's up with type system - is it about typed/untyped langs in general, or something else?
<bqv> There isn't one
<energizer> polezaivsani: it's a pretty unsophisticated dynamic type system
<polezaivsani> thanks! i guess i'd better study that aspect some more
<sphalerite> polezaivsani: depending on your balance of pragmatism vs idealism, you could consider nix's approach as having the shortcoming that it breaks a lot of assumptions that software that people need to use makes.
<polezaivsani> sphalerite: you mean the ones that come with reproducible builds / hardcoded assumptions about the env / etc. ?
<energizer> the main problem with nix is that it requires users to do a lot of work
<__monty__> It's not even a dynamic type system. Nix is untyped. There's lib.types but that's more like manually inserted run-time checks rather than dynamic typing.
<__monty__> Error messages are often useless ime. Which makes debugging a little frustrating.
<polezaivsani> __monty__: guess it's a good thing that i haven't been spoiled by langs with good ones :)
<energizer> polezaivsani: what languages are you coming from?
<polezaivsani> energizer: both statically and dynamically typed; leaning to former ones (halfway int haskell2010 now)
<polezaivsani> energizer: that burden on the users - would that be a learning fee or more of a sustained extra?
<eyJhb> __monty__: I had a jq error once when doing some module stuff, because the underlying package I used (which I guess is not often used, but a big part of Nix) threw that error. I got NO trace, at all
<eyJhb> Was fun
<__monty__> polezaivsani: Are you saying you think haskell's compiler errors are bad?
<__monty__> You need to get used to them but they're absolutely stellar compared to what you get with nix.
<polezaivsani> __monty__: not bad, though i have some anecdotes
<eyJhb> bqv: IT IS ALIVE! :D
FRidh has joined #nixos-chat
aszlig_ is now known as aszlig
arahael1 has quit [Ping timeout: 272 seconds]
arahael1 has joined #nixos-chat
<eyJhb> Anybody experienced malware on NixOS?
<eyJhb> Curious of how it would be
<eyJhb> adisbladis: on any of the clients at Tweag, do you use tmpfs/zfs rollbacks there?
<__monty__> eyJhb: Why would it be different from other distros?
<eyJhb> __monty__: not having /bin /sbin, etc. available, not really having a good way to install backdoors, except if you grep the users path, etc.
<eyJhb> Like, you can
<eyJhb> But it gets more annoying I would assume
<eyJhb> And a lot of the crappy malware will not work
cjpbirkbeck has joined #nixos-chat
<immae> I'm not expert bu I expect that usually malware stay mostly in memory and they're reinitiated via repetition of the user "bad" action no? (possibly facilitated with an ad-hoc config)
<__monty__> Aren't most serious malwares self-contained? Like all the ransomware lately?
<immae> ah yes using paths that are common, didn't think of it
<eyJhb> Well, most I would assume expolit libc as fucking always if we talk overflows, rop, etc. and from there you try to start a program you know of, which is most likely /bin/sh. From there you can do fun stuff! But basically you want command execution to do anything malware fun
<__monty__> But /bin/sh is still present.
<eyJhb> And you cannot use the paths, because most programs will not have that path. e.g. Python
<eyJhb> Yes! But from there on out
<eyJhb> You need to know about /nix/store, or be lucky the program you are attacking have a sweet sweet path
<__monty__> There's /usr/bin/env if you want to find python.
<eyJhb> But I wonder if most would do that
<eyJhb> Also, python, etc. is not default on NixOS
<eyJhb> Also, doens't env just use your PATH to give you a abs path?
clever has quit [Ping timeout: 246 seconds]
<__monty__> eyJhb: If you know you're targeting NixOS you have a world-readable store of software at your disposal though.
clever has joined #nixos-chat
<eyJhb> __monty__: yes :D And that is what is nice as well, but it adds a bit more work
<eyJhb> But attacks like changing a bin etc. won't work as well on NixOS
<sphalerite> I wonder if there have been any code quality surveys on malware
<sphalerite> I imagine a lot of it will be just hacked together to the minimal standard of "it works", but I wonder if some malware is developed with high engineering standards as well?
<pie_> i think there have been but i dont hav a real basis for this
<pie_> s/real/
<pie_> well for high engineering standards its usually nation state stuff, but even there people say things like the chinese just throw something together
<pie_> and hey, if it works for them, *shrug*
<pie_> im also not sure about that "usually"
<pie_> like, who has a professional engineering budget for these things
<pie_> not really sure where the typical professional "cybercriminal" falls on that spectrum
* pie_ wishes for references to cite
<pie_> i dont have any ideas offhand than to peek at some malware analysis blogs
FRidh has quit [Quit: Konversation terminated!]
<pie_> energizer: polezaivsani: how do you handle integration while keeping the benefits of the nix approach, by which i mean things like task switchers not being able to access the necessary informations (like icons) from running programs to work properly, because they dont know how to find the information. arguably this is due to bad infrmation discovery APIs in things unrelated to nix, but i think the dead(data at rest)/live (runtime state
<pie_> stuff) structures distinction is a problem that i havent seen anyone writing about, and which i think needs consideration - or likely im looking at it wrong
<pie_> runtime structures arent purely functional and compositional. it may be illustrtive to take a look at how capabilities based OSs deal with this (which i havent done), but at the end of the day if youre doing some sort of integration, you have to pull all your threads into one place, and so its also a question how that should be done (windows registry? dbus?) (not sure if dbus belongs in that enumeration)
* pie_ noob
<pie_> if you start something like a kde shell, you expect things to just work?
<adisbladis> pie_: I think when you mention dbus what you're really thinking of is dconf
<pie_> mm i guess. im not really familiar with the d ecosystem
<pie_> dbus does seem to fit into the "integration" vein tho?
<adisbladis> dconf is not really tied to dbus in any way, it's the gnomy configuration management tool
<pie_> *nod*
<adisbladis> (the interface is dbus, but we can forget that for a moment)
<pie_> i think another way to pose it might be "how do you do integration in a purely functional system", but some expansion would be needed on whats meant by "integration"
<pie_> my perceived restriction of nix to the file system has been bothering me for a long time, thats all I got. *shrug*
<adisbladis> I think the short answer is that you don't
<adisbladis> And the slightly longer answer is that you do it in a best-effort manner
<pie_> currently, yes, but i dont know either how this could/should be cone
<pie_> *done
<adisbladis> pie_: home-manager has some interesting dconf shenanigans
<pie_> its all just a big questionmark for me
<adisbladis> pie_: I think being strictly purely functional means that you have to have an environment that's conducive to that
<adisbladis> Which Gnome/KDE is not
<pie_> yeah
<adisbladis> Sadly I don't think we'll ever quite "get there" on DEs
<pie_> well, thats part of the question isnt it
<pie_> id probably start looking to haskell for solutions but that would require posing a problem i could do that with :P maybe i just havent thought about it much
<gchristensen> I should learn the mapped keys for weechat better, instead of using alt-arrow and alt-a to navigate between any buffer > 10
* etu has alt-<letter on the row below the number> for 10+<number above letter> mapped
<etu> Works well up to 20 at least
<etu> Of course had to adapt them for dvorak
<gchristensen> nice
julm has quit [Quit: switch]
julm has joined #nixos-chat
<sphalerite> gchristensen: I use the mouse navigation and have managed to avoid learning any more keys through that :D
<sphalerite> it has scrolling and everything!
<gchristensen> :o
polezaivsani has quit [Quit: ERC (IRC client for Emacs 27.0.91)]
polezaivsani has joined #nixos-chat
rummik has joined #nixos-chat
<adisbladis> How are people using nixpkgs as a flake with custom patches? Or are you always using an exact upstream nixpkgs?
<LnL> I think it would be really nice if you could specify patches to apply ontop of sources in a builtin way
cole-h has joined #nixos-chat
parsley936 has quit [Remote host closed the connection]
parsley936 has joined #nixos-chat
<gchristensen> I've been using ConditionACPower on a bunch of units, and I realized today that if that condition fails and then becomes true, systemd doesn't go starting those services
<gchristensen> just like it doesn't go stopping them when the condition stops being true
<gchristensen> I think I probably want a udev rule and have like, ac-power.target and battery.target, and use WantedBy and Conflicts to setup these rules the way I actually intend
<colemickens> "SUSE to Acquire Rancher Labs" huh.
kalbasit has joined #nixos-chat
drakonis has joined #nixos-chat
kalbasit has quit [Ping timeout: 246 seconds]
kalbasit has joined #nixos-chat
<infinisil> Neat
<gchristensen> though the actual implementation detail is not up to date
<cole-h> Do I smell another blog post in the making? :D
<gchristensen> ENV{SYSTEMD_WANTS}="webcam.service"
<__monty__> Re window switching for irc. I've taken to using /wg NN and /msg #channel My important opinions.
<gchristensen> I try to use /query over /msg as often as possible
<adisbladis> Emacs zealot time: I use `C-x b` ;)
drakonis has quit [Ping timeout: 272 seconds]
<gchristensen> because if I accidentally do ` /query adisbladis` (leading /) I haven't typed anything which I meant to send privately, only have leaked the intent to message adisbladis privately
drakonis has joined #nixos-chat
<__monty__> Yeah but this is /msg for public messages to channels : ) So you don't have to switch at all.
<__monty__> I mostly use it with channels in the same window.
<gchristensen> whoa
<drakonis> wow the shell completion needs some love
<drakonis> THIS IS AN ABOMINATION
<gchristensen> that did not need a sequel
<drakonis> this should not exist
cole-h has quit [Quit: Goodbye]
<drakonis> basalt looks pretty cool
<drakonis> no flakes support yet, sadly.
<__monty__> Elevator pitch for basalt?
<gchristensen> what is basalt?
<samueldr> that new gray-dark block in the nether
<drakonis> its a git tool for managing nix configurations
<gchristensen> neat
<__monty__> Similar to niv?
<drakonis> no
<drakonis> niv does dependency management
<drakonis> this makes it easier to manage your system configuration
<drakonis> a replacement for nixos-rebuild so to say
<gchristensen> oh I think LnL built something like this
<drakonis> push the change to a branch and the system will automatically advance
<drakonis> rollback the change to rollback the system
<drakonis> its pretty interesting
<__monty__> Sounds a bit like lorri for configuration.nix.
<__monty__> (I am a bird, I can only remember about 3 concepts so all new things have to fit in one of three boxes.)
<drakonis> a magpie you mean
<__monty__> Are you calling me a thief of opportunity?
<drakonis> im calling you a bird brained person!
<drakonis> its basically tooling to replace having to manage your own git repo
<drakonis> for your configs
<LnL> gchristensen: kind of, built something is a big word for it :)
<energizer> LnL: basalt seemed too heavy for me, do you have a lighter version of the same idea?
<energizer> pie_: i didnt really understand what you're asking about icons
<LnL> energizer: this is a basic vereision of the git hook https://gist.github.com/LnL7/c2ff93e635a0779c839277d730829c12
<jtojnar> energizer probably this: https://github.com/Schneegans/Gnome-Pie/issues/170
<{^_^}> Schneegans/Gnome-Pie#170 (by deliciouslytyped, 48 weeks ago, open): failure to load icons
<energizer> LnL: that's a post-receive hook?
<LnL> update hook
<LnL> that one can block commits from being pushed if the build fails
<energizer> LnL: "nixos-rebuild switch --show-trace" doesn't need sudo?
<LnL> I have it setup with a specific key that can push to root@localhost
tazjin has joined #nixos-chat
<energizer> but this isn't being executed by root@localhost, so how does that help?
<LnL> it is, the hook runs on the receiving end
<energizer> oh. then why is there a `sudo nix-env`?
<LnL> hm, not sure :D
<energizer> jtojnar: lots of details in there, what's the eli5?
cjpbirkbeck has quit [Quit: cjpbirkbeck]
<jtojnar> energizer modern desktop linux stack is designed to rely on global state, Nix cannot provide that
<jtojnar> in this case, to find an icon for a window, there needs to be .desktop file installed whose name matches a property set on the window object
<jtojnar> which obviously fails when you are running a program that is not installed (e.g. from Nix store)
<__monty__> jtojnar: Nixos can install .desktop files though?
<jtojnar> monty yes, NixOS can, Nix cannot
<jtojnar> (though it relies on environment variables being properly routed, which can sometimes fail)
<joepie91> jtojnar: are you using RiotX by any chance?
<jtojnar> joepie91 on the phone, yes
<joepie91> RiotX formatting bug :P
<energizer> nix installs out files and man files, why not desktop files?
<jtojnar> joepie91 I sent that from Fractal, which has somewhat buggy formatting
<joepie91> oh, huh
<joepie91> guess there's two clients with this bug then!
<jtojnar> I saw similar thing in Riot Web when mentioning IRC user whose name contained backtick
<joepie91> nah that's a different bug and bridges correctly
<joepie91> just looks weird on the Matrix side
<joepie91> your message will actually bridge incorrectly
<joepie91> (the bridge will use the `body`)
<__monty__> Oh, huh, the bridge uses mIRC formatting codes or something?
<joepie91> __monty__: no, but Matrix clients typically support Markdown. regular Riot converts this to HTML for `formatted_body` and uses the original Markdown for `body`, but some other clients generate HTML for `formatted_body` and then strip out the HTML and store the plaintext in `body`
<joepie91> the IRC bridge is written with the assumption that the `body` contains the message in plaintext, as it is to be bridged
<joepie91> so if the client blindly interprets monty as Markdown... it converts it into bold monty on the Matrix side, and regular monty on the plaintext body ie. IRC side
<joepie91> (in fact, in the above message it still renders as bold monty for me, but Riot does include the underscores in the plaintext body)
<__monty__> Yeah but jtojnar's mention of me is bold too.
<joepie91> this is a bit of a rough edge in the Matrix ecosystem right now, it's not been defined quite well-enough what the body and formatted_body are expected to contain
<joepie91> oh, huh, really?
<__monty__> Yeah, it's definitely bridged as bold.
<joepie91> maybe it started bridging markup recently
Cynthia_ has joined #nixos-chat
<joepie91> it's done so in the other direction for a while
<joepie91> ie. mIRC color codes -> HTML markup
<joepie91> but HTML markup -> mIRC color codes is new to me :)
<joepie91> (I use a different bridge, which may be an older version)
taktoa[c] has joined #nixos-chat
<__monty__> But it doesn't have the *bold* or _bold_ syntax that irc clients typically bold. Hence my assuming it's using mIRC format codes or something.
manveru has joined #nixos-chat
<__monty__> Oh, _underline_ rather, which is interpreted as bold in markdown.
<joepie91> yeah it would be in that case
rizary has joined #nixos-chat
srhb has joined #nixos-chat
waleee-cl has joined #nixos-chat
jackdk has joined #nixos-chat
<manveru> oh hey, irccloud is back
<manveru> that was like a day of outage :P
<gchristensen> wow.
<gchristensen> :o
<aleph-> Alright seems my day is filled with helm
<aleph-> What a joy
<aleph-> manveru: Oh irccloud was down? Explains why my mate wasn't on
<waleee-cl> it was a repeat of their ISP issue from the beginning of the year
<waleee-cl> (the ISP apparently can't get technicians to their datacentre until the next day for reasons unknown)
Cynthia_ is now known as Cynthia
emilazy has joined #nixos-chat
endformationage has joined #nixos-chat
feepo has joined #nixos-chat
<joepie91> waleee-cl: doubt on "their datacentre"
<waleee-cl> * the ISP's datacentre
raboof has joined #nixos-chat
pasukon has joined #nixos-chat
steveeJ has joined #nixos-chat
johanot has joined #nixos-chat
<eyJhb> worldofpeace and ashkitten what Windows managers did you use, e.g. i3
<ashkitten> did i?
<eyJhb> Noo, but you used something?
<eyJhb> /use
<ashkitten> like, past tense?
<ashkitten> at what point in time are you asking about
<ashkitten> i have used i3, i currently use sway
<eyJhb> Like current WM :)
<eyJhb> Ahh, damn, thanks :)
savanni_ has joined #nixos-chat
sdier has joined #nixos-chat
aria has joined #nixos-chat
jw358 has joined #nixos-chat
<gchristensen> anyone a big fan of game of thrones?
liszt has joined #nixos-chat
elvishjerricco has joined #nixos-chat
davidtwco has joined #nixos-chat
zimbatm has joined #nixos-chat
betawaffle has joined #nixos-chat
cbarrett has joined #nixos-chat
kalbasit has quit [Ping timeout: 272 seconds]
mudri has joined #nixos-chat
peel has joined #nixos-chat
kalbasit has joined #nixos-chat
<cransom> i enjoy it
<bqv> I liked it
kalbasit has quit [Ping timeout: 256 seconds]
kalbasit has joined #nixos-chat
kalbasit has quit [Ping timeout: 246 seconds]
kalbasit has joined #nixos-chat
kalbasit has quit [Ping timeout: 240 seconds]
kalbasit has joined #nixos-chat
<eyJhb> gchristensen: I feel asleep at the season three season finale, and couldn't watch it anymore after that
<eyJhb> To fish or not to fish, that is the questoin
<eyJhb> question*
<eyJhb> Might run it for a while I guess
<__monty__> Do eeeeeet!
<eyJhb> I got close to being hit today for talking about NixOS bqv (in regards to the kicking)
<__monty__> lol
<samueldr> I don't follow, is a community kicking people that are NixOS users?
<eyJhb> No no, friends who hate their friend talking about NixOS are samueldr
<bqv> eyJhb: oh yikes
<bqv> Haha
<eyJhb> :D
<samueldr> this almost sounded like the enemy of my enemy is my friend
<bqv> I get the same whenever I mention emacs…
<eyJhb> bqv: I will kick your .... :p
<eyJhb> I wonder if you find enough "try this online in a shell", how much computing power one could have?
<colemickens> gchristensen: what was the thing that auto-fixup'd?
<__monty__> colemickens: git-absorb
<samueldr> I thought thunderbolt 3 was part of usb4, but it seems I must had misunderstood something in the past
<polezaivsani> so, Emacs as desktop env on a NixOS would mean twice the troubles or double negation? :)
<samueldr> so I'm way confused by the separation of USB4 and Thunderbolt 3 in the news about Thunderbolt 4
<bqv> polezaivsani: that was my actual setup for a long while :p
<polezaivsani> bqv: i wasn't joking either B)
<bqv> :D
<bqv> __monty__: colemickens: woah, that looks insane, i want to try it
<gchristensen> "making today’s simplest and most versatile port available to everyone" "simple"
<samueldr> I think it's techncally the truth
<samueldr> if all you have is a thunderbolt 4 port
<gchristensen> me too actually
<gchristensen> usb-c makes basically no promises about what it is, but thunderbolt is very clear
<samueldr> but if you think about a type-c port, then it's not simple
<gchristensen> :D
<samueldr> is it even usb?
<samueldr> can you plug a usb flash drive into that virtuallink port? https://www.google.com/search?q=nvidia+virtuallink&tbm=isch
<__monty__> One thing that botters me about Type C is the fragile nub in the port. Why didn't they do lightning but with springs on the plug?
<gchristensen> btw is it safe to plug in 2 power adapters?
<gchristensen> or two devices capable of providing power?
<samueldr> gchristensen: that's part of the PD spec, which is separate from type-c
<samueldr> so at least there's that
<gchristensen> I'm not sure if that answered my question :P
<samueldr> that's my point :D
<ar> thunderbolt promises to give random external devices access to ram ;)
<samueldr> ooh, according to a GPU manufacturer YES YOU CAN http://blog.pny.com/blogpnycom/pny-pro-tips9-virtuallink-part2
<gchristensen> ar: doesn't usb-c do that without thunderbolt, anyway?
<samueldr> nope, that's TB
<samueldr> usb3 doesn't have DMA
<samueldr> I don't know about usb4
<ar> apparently you can also connect virtualink to monitors that serve as usb "docking stations"
<samueldr> that's actually neat then!
<samueldr> is there a cable long enough that will reach?
<samueldr> and will you be able to buy it?
<ar> i have yet to see a gpu that has the virtualink port
<ar> (or a laptop for that matter)
<samueldr> yeah, looks like USB4 took the DMA part of TB3 https://www.kernel.org/doc/html/latest/admin-guide/thunderbolt.html
<samueldr> BOOOO
<eyJhb> How can I escape a alias in fish? In bash you can \cat
<ar> also, kind of sad my next laptop won't come with a thunderbolt port for egpus.
<bqv> eyJhb: "command" mught do it
<ar> if it did, it would be - with an egpu - a reasonably decent gaming machine
<ashkitten> hmph
<ashkitten> takes so long to build wine
<eyJhb> bqv: no dice :(
<ashkitten> what if i had a 64-core cpu
<__monty__> eyJhb: `command myalias` should ignore builtins and functions (fish doesn't really have aliases).
<__monty__> But also, it's good practice not to have your aliases shadow commands you might need : )
__monty__ has quit [Quit: leaving]
polezaivsani has quit [Ping timeout: 256 seconds]
cole-h has joined #nixos-chat
<abathur> speaking of bash aliases
<ashkitten> my gpu keeps crashing :<
<gchristensen> ouch
<ashkitten> i think it's the mesa 20.0.8 update
<abathur> speaking of bash aliases
<abathur> oops
<abathur> lulz
<ashkitten> unfortunately i can't really just... update mesa again
<ashkitten> because it's a huge rebuild
<ashkitten> and it already took forever to build wine
parsley936 has quit [Remote host closed the connection]
<abathur> bash aliases are some fun stuff
<cole-h> wat
<samueldr> no → abathur
<abathur> could probably make a pretty mind-numbing bash obfuscator with aliases
<samueldr> it's simple
<cole-h> wat
<samueldr> we just need abathur to stop :)
<abathur> stahp
<cole-h> You mean "stahp"
<abathur> <3 cole-h
<{^_^}> cole-h's karma got increased to 80