gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
BaughnLogBot has joined #nixos-chat
BaughnLogBot has quit [Client Quit]
BaughnLogBot has joined #nixos-chat
ece34 has joined #nixos-chat
ece3 has quit [Read error: Connection reset by peer]
ece34 is now known as ece3
BaughnLogBot has quit [Quit: ZNC 1.6.2+deb1 - http://znc.in]
BaughnLogBot has joined #nixos-chat
BaughnLogBot has quit [Client Quit]
Dotz0cat has joined #nixos-chat
Dotz0cat_ has quit [Ping timeout: 256 seconds]
supersandro20005 has joined #nixos-chat
supersandro2000 has quit [Ping timeout: 265 seconds]
supersandro20005 is now known as supersandro2000
ky0ko has joined #nixos-chat
<hexa-> explaining functional programming with factorio examples
<hexa-> s/with/using/
tilpner_ has joined #nixos-chat
rajivr has joined #nixos-chat
cole-h has quit [Ping timeout: 260 seconds]
tilpner has quit [Ping timeout: 260 seconds]
tilpner_ is now known as tilpner
<bbigras> nice
Jackneill has quit [Ping timeout: 256 seconds]
<siraben> hexa-: heh I was about to post that!
<siraben> cc supersandro20005
Jackneill has joined #nixos-chat
<supersandro2000> siraben: why did you ping supersandro20005?
<siraben> oops my element client autocompleted the wrong person, heh
<siraben> meant to ping you obviously
<supersandro2000> it should auto change the nick
<supersandro2000> matrix being matrix 😂
<supersandro2000> but I know
<supersandro2000> already looked at the article
slack1256 has joined #nixos-chat
<siraben> though myself I think category theory is more motivating once you see groups, rings, vector spaces, topological spaces, etc. then the abstractions and their laws do appear quite natural
blueberrypie has quit [Quit: leaving]
blueberrypie has joined #nixos-chat
<supersandro2000> bless you
lunc has quit [Ping timeout: 264 seconds]
<siraben> who sneezed?
<slack1256> me
endformationage has quit [Quit: WeeChat 2.9]
slack1256 has quit [Ping timeout: 240 seconds]
rajivr has quit [Quit: Connection closed for inactivity]
waleee-cl has quit [Quit: Connection closed for inactivity]
rajivr has joined #nixos-chat
supersandro2000 has quit [Quit: Ping timeout (120 seconds)]
supersandro2000 has joined #nixos-chat
srk has quit [Remote host closed the connection]
srk has joined #nixos-chat
<lovesegfault> I'm in nix jail
<aleph-> Is that like super jail?
<aleph-> :p
<lovesegfault> pretty much :P
<lovesegfault> it's when you get absolutely intractable errors :D
<aleph-> Heheheh
<aleph-> Okay wonder how much downloading a channel would run me...
Dotz0cat_ has joined #nixos-chat
Dotz0cat has quit [Ping timeout: 272 seconds]
rj has quit [Remote host closed the connection]
rj has joined #nixos-chat
aleph- has quit [Ping timeout: 265 seconds]
<siraben> lovesegfault: what's a nix jail?
cole-h has joined #nixos-chat
aleph- has joined #nixos-chat
dingenskirchen has quit [Quit: ZNC 1.8.1 - https://znc.in]
dingenskirchen has joined #nixos-chat
skrzyp has quit [Ping timeout: 272 seconds]
skrzyp has joined #nixos-chat
lunc has joined #nixos-chat
rj has quit [Remote host closed the connection]
rj has joined #nixos-chat
cole-h has quit [Ping timeout: 260 seconds]
<DigitalKiwi> probably something on freebsd
<supersandro2000> I just noticed that a haskel program is using 1T of virtual ram
<Taneb> That's how GHC's runtime does things
<leons> yeah, it's a common occasion when using things such as git-annex
<Taneb> Nothing really to worry about
<leons> yes. The linux kernel is very efficient in granting virtual memory without actually allocating too many pages
<supersandro2000> I thought so
<supersandro2000> I don't have near one TB of RAM
__monty__ has joined #nixos-chat
<DigitalKiwi> turns out it only takes 1TB of virtual ram to hold infinite lists
<__monty__> That's just GHC oversubscribing (not sure about the terminology anymore) virtual memory to avoid any penalties when growing it.
<V> overcommitting is the word
<__monty__> Ah, yes, thanks. V++
<{^_^}> V's karma got increased to 12
<V> it just means they're shifting the obligation to add memory maps to the kernel, which means that some (not entirely) arbitrary process gets murdered when the system runs out of memory, instead of them having to deal with ENOMEM
<V> (yes, the kernel would add memory maps without overcommit, but in that case it guarantees the process will have sufficient memory for every allocation it makes, as opposed to pretending it does under the hope that it won't use so much that the kernel has to kill something to free up space)
<__monty__> Thanks, didn't know the details any more.
<V> I'm somewhat conflicted as to whether it's generally a good or bad thing; I can definitely see the use, but it also reduces predictability/robustness
<__monty__> In case you actually go over the 1 TB threshold?
<V> well, it's more like. browsers do this stuff as well.
<V> fork() is also totally broken without overcommit, as the child process needs exactly as much memory as the parent has (there's vfork for this, but not everyone knows/uses it)
<V> I used to have overcommit disabled on a desktop system and browsers would just outright crash, even when they could have presumably freed cached stuff
tilpner_ has joined #nixos-chat
tilpner has quit [Ping timeout: 264 seconds]
tilpner_ is now known as tilpner
<supersandro2000> I don't know what people have about chrome eating ram
<supersandro2000> it has 8GB and now it crashes rather than using more
<supersandro2000> 🙃
<eyJhb> Emantor: ohh cool! What kind of devices do you make? Do you have anything to do with embedded systems that use SIM Cards for communication?
<Emantor> eyJhb: We do not build devices ourselves, but provide consulting services for people building devices based on Linux. Embedded Linux can be found in a lot of places, from industrial climate control systems to embedded entertainment devices.
<Emantor> And yes, some of our customers have devices with SIM cards. Lots of LTE modems are supported natively or only require a single kernel change to be detected. Thanks to Microsoft and their MBIM requirement.
<AMG> Microsoft is trying a new solution for the persistent “would you like to allow notifications from this website” requests that you see across the internet: crowdsourcing data on which ones people block and which ones they allow. According to a blog post today, Microsoft is calling this feature adaptive notification requests, and the company is rolling it out in Edge 88 after it received
<AMG> positive feedback from testers.
<AMG> sweet idea^
<gchristensen> what does that mean?
<eyJhb> Emantor: Nice! We are currently doing a semester project, where we have to work with "somewhat" embedded devices, that gather powergrid information, which is sent using a modem :D Or rather, need to make some diagnostic tools for debugging why the stop sending stuff...
<eyJhb> Unsure how much can be done. But embedded devices are cool
waleee-cl has joined #nixos-chat
<Emantor> The devices are running Linux as well? Could be a lot of stuff all over the stack. Remember to check your assumptions ;-)
<eyJhb> We are looking at their dev setups, so it is a Raspberry Pi with a modem attached to it + a meassuring device. But the "real" systems are just the meassuring device and a sim card plobbed into it, but I would assume they are Linux!
<tilpner> Identifying the right assumptions to check tends to be the crucial part
<eyJhb> But yeah, currently we are doing the system definition, and then we need to do FMEA on that system (or the scope)
<eyJhb> *anyone that has some nice things for FMEA are welcome* :D
* Emantor hadn't heard of FMEA until reading the wikipedia article.
<joepie91> published a new library again: https://twitter.com/joepie91/status/1362046635333873678
pie_ has joined #nixos-chat
<gchristensen> so check out this issue, and click through to the linked comment at the end https://github.com/openzfs/zfs/issues/11610
<{^_^}> openzfs/zfs#11610 (by grahamc, 2 hours ago, open): libnvpair_json tickles incompatibilities between JSON parsers by representing 64 bit numbers as numbers
<Taneb> :D
<supersandro2000> typical told you moment?
<gchristensen> a little bit of a sad/crying/laughing moment
<siraben> lol
<FireFly> haha
lunc has quit [Ping timeout: 272 seconds]
<eyJhb> Emantor: I actually thought that FMEA was what you hinted to :D
lunc has joined #nixos-chat
<joepie91> gchristensen: probably not relevant in this case, but bitwise operations will also truncate integers further to 32 bits
<Emantor> eyJhb: The company I'm working at is (fortunately?) not big enough for fancy business practices :D
<cransom> for any other datadog users, i'm looking around and don't see it. does datadog not collect pressure stall info stats?
<veleiro> i really hate mailservers
<srk> ,nixos-mailserver
<{^_^}> nixos-mailserver defined
<gchristensen> veleiro: I charge hazard pay if the job involves a mail server
<eyJhb> Emantor: cool non the less :D I found this nice little thing from NASA where they describe the process https://ntrs.nasa.gov/api/citations/19700076494/downloads/19700076494.pdf?attachment=true but it needs to be viewed together with other materials to get a full picture of it.
<eyJhb> Is any of the stuff you ( Emantor ) do opensource?
<eyJhb> gchristensen: I really feel that. I have not tried sitting up a mail server since I was 12, until I tried nixos-mailserver some time ago! It is WONDERFUL!
<gchristensen> nice
<eyJhb> However Hotmail + Gmail do not like me...
srk has quit [Remote host closed the connection]
srk has joined #nixos-chat
<eyJhb> Also, if anyone finds the actual FMEA procedures for Apollo feel free to ping me
lunc has quit []
slack1256 has joined #nixos-chat
<hexa-> eyJhb: labgrid is :)
<Emantor> eyJhb: I work for Pengutronix, https://www.pengutronix.de/en/index.html has an open source tab with all our projects. I am one of the maintainers for labgrid and otherwise I have been contributing to OP-TEE.
slack1256 has quit [Ping timeout: 265 seconds]
<infinisil> > timeTo mars
<{^_^}> "1 day, 1 hour, 30 minutes, 16 seconds"
<gchristensen> whew
<infinisil> OH damn, I have a meeting then
<gchristensen> cancel it
<infinisil> Time to move that :)
<pie_> lol
<pie_> just make watching it the meeting
<aleph-> lol
<aleph-> pie_: Good idea
rajivr has quit [Quit: Connection closed for inactivity]
Mic92 has quit [Quit: WeeChat 3.0]
Mic92 has joined #nixos-chat
<veleiro> srk: thanks that module actually made me understand mailservers much more, especially in the docs they list the way to get a perfect score
<infinisil> I do have this perfect score, but unfortunately my mails still get rejected by some google mail servers
<veleiro> that's the story with using nix for me too. it simplifies complex systems in a way that I can better understand it all
<veleiro> at least abstractly
<infinisil> There's a good guide out there which I want to go through at some point to have my mails not be rejected anymore, trying to find it..
<veleiro> cant you just be marked as spam at some point, by google mails and that's it?
<infinisil> I doubt there's nothing you can do then
<veleiro> thanks bookmarked
<infinisil> It's very recent too
<infinisil> Now that nixus has a dns module, I want to implement all of this using it
mcint has joined #nixos-chat
<infinisil> So that people won't have to manually do that in the future
<veleiro> protonmail does everything different so they have their own bridge/api to interact with their mail
<veleiro> i recently found https://github.com/emersion/hydroxide so I could more easily have it listening to smtp connections
<veleiro> ... but it doesnt do STARTTLS, and sourcehut expects it to
<veleiro> so gotta change something somewhere or do a smpt tls proxy
<aleph-> infinisil: Yeah on the one hand I really need to change mail hosts because the viability of the hosting company is... iffy
<aleph-> On the other hand I don't get blocked by anyoneeeee
<infinisil> By the hosting company you mean the ones that own the domain of your mail?
<infinisil> aleph-: ^
<aleph-> infinisil: Nope, the vps hosting company I run my mail server on
<gchristensen> DigitalKiwi: I had the occasion to share https://gsc.io/snaps/a972076a-956e-4d3d-ab81-5d00555d0b5b.png with someone today, what a delight it is <3 thank you
<DigitalKiwi> gchristensen: lol that's funny this was in #hledger earlier 07:10 DigitalKiwi: https://twitter.com/grhmc/status/1269294198836076546?s=20
<gchristensen> nice
slack1256 has joined #nixos-chat
<veleiro> who made the nix logo
<veleiro> any meaning of it I dont see?
<DigitalKiwi> it's a snowflake of lambdas
<mcint> the ciiiiircle of lambdas
<DigitalKiwi> because it's unique and special
<veleiro> ohhh the lambda yeah
<veleiro> now its better for me
<samueldr> veleiro: tim bertson, details are in the nixos-artwork repo
<samueldr> uh, tim cuthbertson*
cole-h has joined #nixos-chat
<gchristensen> anyone mind checking some algebra of mine? I flunked math
<eyJhb> gchristensen: It really depends on what scale it is :D
<gchristensen> it is not tricky
<gchristensen> I think
<gchristensen> woohoo I got it right :')
<gchristensen> it is a shame math is taught the way it is
<cole-h> hehe
<samueldr> by rote learning?
<gchristensen> by maximizing pain I guess
<gchristensen> I just saw a video on this ...
<samueldr> yeah, rote learning :)
<samueldr> the better teachers/profressors I had were those that used out of the box methods to present some of the concepts
<samueldr> but in the end it was still about getting to a simple formula you had to remember by heart
<samueldr> I would have rather seen harder tests, with open books, than being punished by having too much to remember already
<samueldr> since anyway, most of the time open book won't help if you haven't learned the thing
<gchristensen> right
slack1256 has quit [Ping timeout: 264 seconds]
<eyJhb> Luckily all mine are open book...
<eyJhb> But not all have the help of calculators, etc.
<eyJhb> And some do, and you are allowed to use your computer + CAD stuff, but most of us used woldframalpha. BUt that was until I found out about Sympy <3
<__monty__> I far preferred the ones that banned calculators.
<__monty__> Just meant the calculations weren't too hard to do in your head.
<eyJhb> What my professors can do in their head, and what we could do was... not the same to say the least
<__monty__> Ime usually students can do more mental arithmetic.
<samueldr> I'm totally terrible at simple arithmetics in my mind
<gchristensen> yeah, I can'at do it
<samueldr> which, after having listened to someone's experience with dyslexia, is one of the forms of how it can manifest (and in their locale, being diagnosed with dyslexia one thing like that is not enough)
<samueldr> I wouldn't think I have dyslexia, in these conditions, as pretty much it's the only thing I experience, and even then, I'm not even sure it's at a level it would "count" (hah)
<samueldr> (btw, great podcast)
<samueldr> an uh, there's also a link in the show notes about a talk he did about it
<gchristensen> dyscalculia?
<samueldr> it may be something about being in the UK, and something that dyslexia is not _just_ about reading
<samueldr> I'd have to fish for the right minute in the podcast
<mcint> for mental arithmetic, i find helpful that i have a bag of tricks, like rules for some numbers (e.g. 9), and divide-and-conquer sub-algorithms
<samueldr> around 12:41, good thing it was at the beginning of steve's bit
Jackneill has quit [Ping timeout: 240 seconds]
<__monty__> Hmm, I didn't learn the tables by rote either.
<__monty__> Seems to have always been an advantage with mental arithmetic though.
<__monty__> Because it helps with all of it. Rather than just the few pairs that fit the tables.
Jackneill has joined #nixos-chat
Jackneill has quit [Ping timeout: 246 seconds]
Jackneill has joined #nixos-chat
<veleiro> when did NixOS become really usable? it seems its popularity is growing quickly
<veleiro> just nix systems in general really
<__monty__> Quite a few years back I'd say.
<__monty__> Though it's not ready for prime time yet, imo.
<gchristensen> 2014 is when its growth really shifted iirc
<__monty__> UX still needs work.
<joepie91> I've noticed a shift in the past 2 years
<joepie91> where it starts being mentioned fairly reliably in eg. comment threads about deployment etc
<veleiro> yeah well i think flakes are incredibly useful and this intensional store stuff, flood gates will open when this type of upgrades settle
<veleiro> just cant believe nixos is more than a decade old
<veleiro> i dont imagine running my own hydra instance without the intensional store model
<gchristensen> nix's growth has really been a slow burn thing
<gchristensen> kinda cool
<veleiro> hard to believe it was discussed in the original whitepaper!
<__monty__> Avoid success at all costs!
<gchristensen> not here!
<__monty__> You're probably not bracketing correctly.
<gchristensen> I'm likely just delusional
<__monty__> Avoid (success at all costs)!
<veleiro> yep, because once a piece of technology gets so much weight behind it, it will never change, too much at stake, too many huge entities involved in it.. like bitcoin
<veleiro> that or control will get sucked away by some megacorp redhat microsoft google etc like linux
<__monty__> Rather than "(Avoid success) at all costs!"
<gchristensen> infinisil: want to see something fun about taht?
<infinisil> A Nixcon talk?
<gchristensen> no
<gchristensen> try to nix-store -r that path
<gchristensen> in the commit message
<infinisil> Hehe
<gchristensen> I didn't know there was really a land before those rules
<infinisil> Oh it's hex
<gchristensen> yeah
<veleiro> wow @ the original nixos script! haha, 17 years old
<infinisil> I initially thought it was base32 but with the naughty chars allowed
<veleiro> incredible
<infinisil> Oh yeah, if we're going by that commit, 3 days ago was NixOS' 17th birthday
<veleiro> nice
<supersandro2000> I want a time estiomation like llvm has on everything in nixos
<veleiro> have you seen the updates from tweag about the new nix ui with the bars?
<infinisil> > timeSince = dateTime: let now = date.epochToDateTime builtins.currentTime; in date.prettyDuration (date.diffDateTime dateTime now)
<{^_^}> timeSince defined
<{^_^}> nix#4296 (by edolstra, 11 weeks ago, open): Progress indicator improvements
<supersandro2000> if nom could collect time statistics and then estimate...
<cole-h> Only reason I'm not running that branch is because nix repl is still broken
<veleiro> oh that was infinisil haha
<infinisil> > firstNixOSCommit = date.epochToDateTime 1076799411
<{^_^}> firstNixOSCommit defined
<infinisil> > timeSince firstNixOSCommit
<{^_^}> "17 years, 2 days, 21 hours, 51 minutes, 28 seconds"
<infinisil> Nice
<veleiro> thank's cole-h
<veleiro> i dont use nix repl but i need to learn to use it in development daily
<infinisil> (btw, you can get unix epoch git times with --date=unix)
<supersandro2000> does it work outside of repl?
<supersandro2000> I have the power of aliases and nix to work around that
<veleiro> and wow the on demand menu for increasing verbosity, logs etc with the progress bars
<veleiro> i'm going to have to check this out
<cole-h> supersandro2000: when I tried it a month or so ago, it worked fine except for that
<supersandro2000> sounds nice
<supersandro2000> if it now works as a drop in for nixpkgs-reviews
<cole-h> what
<cole-h> how does the progress-bar relate to nixpkgs-reviews?
<veleiro> I cant easily find what dependencies nix needs for building/running, i was interested in seeing them for that progress bar PR
<veleiro> i guess nix abstracts these things so well the idea of listing build dependencies is silly lol
ky0ko has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
__monty__ has quit [Quit: leaving]
<samueldr> gchristensen: (finally came back to what I was talking about) the talk from steve mould also goes in the same direction AFAICT https://youtu.be/iAcvYqDKznE
<supersandro2000> cole-h: I want a better progress bar that does not suck
<bbigras> uh tailscale has a "service monitoring" tab now. https://login.tailscale.com/admin/services
<gchristensen> oh cool
<gchristensen> samueldr: oh cool
slack1256 has joined #nixos-chat
<aleph-> bbigras: Oh interesting.
<aleph-> Probably won't use it but neat.
ece3 is now known as ece
ky0ko has joined #nixos-chat