gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
drakonis_ has joined #nixos-chat
drakonis has joined #nixos-chat
drakonis_ has quit [Ping timeout: 250 seconds]
drakonis_ has joined #nixos-chat
drakonis1 has joined #nixos-chat
drakonis has quit [Ping timeout: 240 seconds]
<samueldr> I just realised I'm terrible at networking :|
<samueldr> (with people, and with computers)
<samueldr> I don't remember who I met during NixCon who had contacts at FairPhone...
<pie__> sounds like a taking notes situation :P
<gchristensen> samueldr: those are people you write down and email as soon as you get back to the hotel :P
<samueldr> yeaaaaaaah
<samueldr> that's what I said: I'm terrible at it!
<gchristensen> you'll get better :)
<colemickens> If a company distributes. as a binary, an OSS software with their own patches, and also distributes the source code and/or patches, what is the legality of someone simply submitting those changes as a patch upstream from a disposable email address?
worldofpeace has joined #nixos-chat
<colemickens> I'm actually rather impressed at the landing page for SAPMachine, for what was running through my head when I clicked on it.
<gchristensen> colemickens: I think the customer is allowed to do that
<gchristensen> colemickens: at the same time, the company is allowed to stop providing future patches to that customer
<gchristensen> like this could happen to grsec I think
<colemickens> gchristensen: does this imply that I must be able to prove (or must prove at time of code submission) that I was a (licensed?) customer entitled to said source?
<colemickens> Can I not as a customer redistribute that to a third party entity who then (unattached to me) might submit such code?
<gchristensen> with GPL patches to the software are also GPL
* colemickens nods
<gchristensen> so no
<gchristensen> there are no restrictions on you sharing the GPL'd patches
<gchristensen> oh I see, how would they know you did it?
<colemickens> I suppose, yeah. I'm trying to imagine how it could play out. They see their patches on an ML, look at the email address, and either (1) realize it's a disposable email that is likely to not have been a licensed customer, or (2) that it's ME and try to come after me or cut me off, or whatever for the sake of argument.
<gchristensen> obviously I'm not a lawyer and this is not legal advice
<pie__> also something something fingerprinting
<gchristensen> maybe they somehow encode a customer ID in to the patch with steganography =)
<pie__> samueldr: what makes the android kernel different from mainline anyway, besides device drivers
<colemickens> it's actually nicely publicly linked from a licensing document included with the download (a link which is public and contained in many linux distributions' package sets).
<samueldr> pie__: in the pictures, there's a graph with arrows
<samueldr> mainline releases a stable version... THEN google adds their crapload of patches over top, THEN the SOC vendor forks a *static* version of that, THEN the OEM vendor gets a *static* version of that
<samueldr> so you get like a year to two year old release with a crapload of patches
<samueldr> at best
<samueldr> so porting updates is not trivial
<pie__> i mean whats in the crapload of google patches
<pie__> why is android not linix
<pie__> linuy
<samueldr> you can do it
<pie__> whatever you know what i mean
<samueldr> :D
<pie__> x'D
<samueldr> most of the time it's features from newer kernel, backported
<samueldr> or misfeatures specific for android
<samueldr> which can't easily be mainlined
<samueldr> that's the android team's fork
<pie__> i think i might be asking about the latter
<samueldr> then, the SOC vendor adds their drivers
<samueldr> they write them without even considering mainlining
<samueldr> it's pump and dump
<samueldr> then get to the next SOC
<samueldr> (SOC being the generation of CPU)
<samueldr> then, the OEM adds different magic sauces on top of the SOC drivers... sometimes drivers for specific hardware, again not written with mainlining in min
<samueldr> mind*
<pie__> samueldr: how's stuff going btw
<samueldr> android misfeatures... I don't have a list
<pie__> i cant imagine how much work an avid nixoser is getting done full time :P
<samueldr> since the talk, https://mobile.nixos.org/
<samueldr> got this going
<samueldr> pixel 2 port was apparently working in the end
<samueldr> now I'm working on pinephone stuff
<pie__> spacekookie: i just remembered our group talk about the whole hardware testing automation stuff, cant remember where we got besides "maybe at some point" but i think we were talking about mobile stuff?
<samueldr> though hitting roadblocks at misc. points
<pie__> spacekookie: just because maybe samueldr would be interested in such as well
<samueldr> I might have to get my hands dirty in writing driver code for the pinephone, which is a scary (not really) new thing
<pie__> samueldr: no i dont have a solution someone just had some seemingly nice infra for x86 platform testing
<gchristensen> I was dreaming about doing that with AMT...
<samueldr> gchristensen: just saying... it looks relatively trivial :)
<gchristensen> :D
<pie__> idk if the guys that do it are on ir
<pie__> irc
<colemickens> samueldr: surely you're not the only point running into such issues with the pinephone? are people who might be writing drivers for it across various projects (postmarketos, etc) collaborating on efforts?
<samueldr> looks like, like for the PBP, most efforts are on the kernel
<samueldr> and leaving u-boot with nothing
<samueldr> so when your boot will fail with all their OS, no way to know why
<samueldr> :/
<samueldr> and, admittedly I wasn't following much in the last few months
<samueldr> not entirely sure what's the status of stuff
<samueldr> and it's hard to get the right info
<pie__> hmhm
<pie__> low level code is black magic to me
* colemickens dreams of a PinePhone dual booting Nix-built-ReplicantOS and NixOS.
<pie__> heh
<pie__> anyone here good with cgroups
<pie__> i still havent limited the ram of my firefoxes.....
<pie__> its the main reason im not improving my nixos config
<samueldr> colemickens: I'm not sure what the support will look like day one
<samueldr> and then, day 100
<samueldr> (once the device is in the hands of end-users)
<colemickens> I'm thankful folks are working on it, I'll actually be most excited if there is still interest/develop/community at day 200 (/me glances at Librem5, etc)
rardiol has quit [Ping timeout: 240 seconds]
drakonis1 has quit [Quit: WeeChat 2.6]
das_j has quit [Remote host closed the connection]
das_j has joined #nixos-chat
<Church-> I hope so. Looking at samueldr's work on mobile nixOS with interest haha
endformationage has quit [Ping timeout: 265 seconds]
cdepillabout has joined #nixos-chat
cdepillabout has quit [Quit: Leaving]
<eyJhb> Any good ways to have a simple way, to start a specific "workspace"? E.g. using i3 I normally start my PDF viewer and have two tabs in my terminal (for editing and for compiling), but would be nice to just have a way of saying "start this", and it would
Jackneill has joined #nixos-chat
drakonis_ has quit [Ping timeout: 250 seconds]
drakonis_ has joined #nixos-chat
__monty__ has joined #nixos-chat
<eyJhb> tilpner: I have looked at that, but doesn't seem like it has the ability to open terminals in X directory and X pdf file I guess
<tilpner> eyJhb: AFAIUI you write a script that does ( cd $WORKDIR; evince foo.pdf ) &
<tilpner> But I must admit to only having tried it a long while ago, and deciding to just keep my laptop turned on :c
<eyJhb> I wish I could do that... :/ But not easy when transporting it around, and *sigh* DisplayLink doesn't play nice with hotswapping it in/out...
<eyJhb> Also, wondering... Am I weird for only having two columns and X rows when working with i3? I feel like having anymore is hard to manage in regards to how it places windows etc.
<tilpner> eyJhb: My battery is terrible, but it's usually good enough to last a day on suspend when at 100% previously
<tilpner> Doesn't help that my laptop now shuts down when the battery says "50%"...
<tilpner> eyJhb: So I only restart every time I upgrade nixpkgs, which is about every 2 or 3 weeks
rardiol has joined #nixos-chat
rardiol has quit [Ping timeout: 265 seconds]
<eyJhb> srhb: how good are you at dependency graphs?
<sphalerite> eyJhb: I often don't even go that far, except within tmux
<sphalerite> eyJhb: as in, I usually just have a bunch of single-window workspaces
<srhb> eyJhb: Insofar as I let nix handle it for me, and Nix is great at it, transitively I am too
<spacekookie> pie__: Hey, yea. "maybe at some point" was the impression I got as well :)
<eyJhb> srhb: I wish Nix could do it for my report... But that is not how it works :p
<eyJhb> spacekookie: I have considered it as well! Using tmux that is
<spacekookie> Hm?
<__monty__> Erroneous ping.
<spacekookie> tmux is cool though
<__monty__> Yep, it's great that you can get the same setup you use on a remote server. None of the tiling WMs can do that.
<adisbladis> __monty__: Emacs (exwm) can do that :)
<sphalerite> adisbladis: if only emacs could do wayland :(
<adisbladis> sphalerite: I wish
<eyJhb> Yeah sorry spacekookie I meant sphalerite :(
<__monty__> adisbladis: OSes don't count >: |
<__monty__> ; )
<etu> __monty__: But exwm is great ;)
<adisbladis> Ugh... Lots of breakage today
<talyz> exwm <3
<eyJhb> ^^^ Sweden unite
<eyJhb> :p
<adisbladis> ^_^
<talyz> ...but yeah, wayland would be nice. Doesn't look like emacs will get wayland support any time soon, though :/
<adisbladis> talyz: I'm still playing with the idea of somehow tying together kwin & emacs
<adisbladis> But... Too many projects
<__monty__> I've never gotten into tiling. None of the layouts work for me. Maybe I should bite the bullet and write some myself.
<talyz> eyJhb: all swedes use exwm, you know ;)
<FireFly> ..exwm? :p
<talyz> adisbladis: That would be really cool
<adisbladis> talyz: It's a statistical fact :)
<adisbladis> FireFly: Emacs window manager
<FireFly> o
<talyz> adisbladis: exactly!
<etu> FireFly: We can help you set it up during 36c3 if you like ;)
<adisbladis> Fun tidbit: https://github.com/nix-community/emacs-overlay started it's life as an overlay for me to get exwm/xelb off master :)
<adisbladis> Then it grew from there
<etu> exwm probably have like hundreds of users or so
<FireFly> I like herbstluftwm as it is :p
<FireFly> ...and vim ;p
<etu> FireFly: you can run evil mode and have vi-bindings for your WM :p
<adisbladis> FireFly: Emacs is a better vim than vim ;)
<FireFly> you might be right about that
<FireFly> I tried emacs & evil a li'l, maybe should give it another shot someday, hehe
<eyJhb> talyz: well! Now I have to play dirty, but does Molly use it, huh? :D
<adisbladis> eyJhb: If not then Molly is not a swede
<talyz> eyJhb: well, no..
<adisbladis> EXWM is a requisite for swedish citizenship
<talyz> eyJhb: ..not yet!
<eyJhb> I don't know what our official Linux requirement is.. I might need to consult with srhb and adamt
<sphalerite> talyz: adisbladis: yeah, I've started running emacs in a terminal again because it sucks in xwayland
<adisbladis> :O
<adisbladis> I cant imagine going back to terminal emacs
<adisbladis> sphalerite: Is it that bad in xwayland?
<__monty__> Only used emacs for a little while but have to say the -nw experience was kinda painful.
<sphalerite> well, probably not in general, but with multiple monitors having different scaling factors, it gets horribly blurry
<adisbladis> Ahh, makes sense
<adisbladis> I "solve" that by only having one massive monitor
<eyJhb> I never really want to go back out of my terminal..
<eyJhb> If I were to run emacs it would be terminal as well
<adisbladis> eyJhb: Emacs is a _much_ better experience outside of the terminal
<sphalerite> nah, emacs really is better in graphical
<talyz> sphalerite: yeah, that's my main issue as well - my current "solution" is to not scale :/
<sphalerite> it feels like a terminal but can do nice things like variable-width fonts and mixing font sizes.
<adisbladis> I often have different font sizes for different buffers :)
<sphalerite> it's like the next step from a plain terminal.
<adisbladis> I'm starting to use eshell a lot these days
<talyz> it would be nice if emacs dropped some of the "pretending everything is a terminal"-mindset and use gtk a bit better
<sphalerite> let's rewrite emacs! /s
<talyz> yay~
<adisbladis> sphalerite: In rust!
<sphalerite> it _is_ very nice that emacs _can_ run in a terminal though
<talyz> sphalerite: indeed, I think it should still be possible, but the modes need to be decoupled..
<sphalerite> it's also funny how I need to deactivate and reactivate xterm-mouse-mode whenever I want to use the mouse in a newly opened client
<__monty__> I think it only works *because* they're entwined.
<__monty__> You'd probably end up with a crufty terminal implementation no one uses and therefore no one wants to work on otherwise.
<adisbladis> __monty__: I think you're underestimating just how many people still use the terminal interface
<__monty__> I don't think so. I only checked out emacs for a number of weeks. So-called "emacsers" I talked to had no idea of all the drawbacks attached to running in a terminal.
<__monty__> Splitting the two up more would just cause divergence.
<__monty__> And split up the team somewhat, because some people might only care about one way of using emacs.
<talyz> __monty__: that's entirely possible - it at least seems to be the case with support for old X toolkits that almost no one uses any more
<__monty__> I think it's the reason for sticking with the "technical debt" of limiting yourself (mostly) to what terminals can do.
<qyliss> ^ Somebody working on making Emacs use gtk3 properly, with proper wayland and stuff
<qyliss> It crashes on startup for me on NixOS at the moment, but I'm hoping it's going somewhere
<infinisil> :O
<etu> Yeah
<etu> I have heard about that one
<etu> been thinking of trying it
<leons> That's amazing
psyanticy has joined #nixos-chat
<leons> "Commit messages are in Japanese." :/
<leons> Probably a bit to complicated to learn, but there's translators :)
<adisbladis> :O
<sphalerite> I didn't see any japanese on the first few pages of commit messages
<gchristensen> wow! yay!
<gchristensen> qyliss: do you have multiple displays?
<qyliss> I've been single display when I've tried it
<gchristensen> did you write a nix expression?
<qyliss> I did
<gchristensen> may I have it? :)
<qyliss> let me see if i still have it
genesis has quit [Ping timeout: 240 seconds]
<qyliss> gchristensen: https://0x0.st/z665.txt
<qyliss> The glib-compile-schemas thing won't be necessary on more recent revs because I made a PR to fix it
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos-chat
waleee-cl has joined #nixos-chat
<talyz> qyliss: cool, I'll definitely try it out! :)
<etu> qyliss: Nice! Thanks for that link :)
<gchristensen> amazing to know that Emacs has 140,000 commits and NixOS has 200,000
<gchristensen> qyliss: so I built it and I'm trying to run it, but it defaults to running in the console. did you have to do anything psecial to make it run as a window?
<gchristensen> oops
dtz has quit [Ping timeout: 250 seconds]
drakonis has joined #nixos-chat
<infinisil> Why are you screwing with me, Apple
<infinisil> A couple days ago it deleted a bunch of my mail folders, including contents
<infinisil> And now I got an email saying I don't have any icloud space left so I can't receive mail anymore..
<gchristensen> :|
endformationage has joined #nixos-chat
drakonis has quit [Quit: WeeChat 2.6]
drakonis has joined #nixos-chat
<gchristensen> scm
drakonis1 has joined #nixos-chat
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos-chat
drakonis_ has quit [Ping timeout: 245 seconds]
drakonis1 has quit [Ping timeout: 252 seconds]
drakonis has quit [Ping timeout: 276 seconds]
drakonis has joined #nixos-chat
<qyliss> omg how
<gchristensen> one sec
<{^_^}> masm11/emacs#5 (by grahamc, 6 seconds ago, open): pgtkfns: g_settings_schema_source_lookup can return NULL
<gchristensen> qyliss: I can't figure out how to fix the obvious problems though
<qyliss> Such as?
<__monty__> The weird scaling of the text?
<gchristensen> yeah, and the bottom bar
<gchristensen> if you zoom in really far you can see it
<gchristensen> maybe the solution is trivial. I have very very minimal experience configuring emacs
<__monty__> "Why do you have a telescope on your desk?" "Oh, I use emacs on wayland, no big deal. No *literally* no _big_ deal."
<gchristensen> lol
<qyliss> lol I totally missed the text scaling
<gchristensen> oh cool, they merged my PR quite fast
<Church-> Nice
<das_j> oh has anyone figured out emoji in fontconfig?
Jackneill has quit [Remote host closed the connection]
<jtojnar> das_j it should work in GTK/Pango apps at least
<das_j> jtojnar: I got them to work in Chrome, but not firefox or dunst (which uses pango)
<das_j> can you share some fontconfig config?
<jtojnar> das_j I just use the default one
<das_j> oof
<das_j> well for me I had to disable dejavu because it is priorized with its ugly emoji
<das_j> looks like more googeling and mybe reading fontconfig doc at some point
<jtojnar> das_j plus
<jtojnar> fonts.fonts = [ pkgs.joypixels ]; fonts.fontconfig.defaultFonts.emoji = [ "JoyPixels" ];
<das_j> I'll check that out, maybe it helps me. Thanks!
<jtojnar> das_j what emoji font are you trying to use?
<das_j> Blobmoji. Not upstream, but just a fork of noto-emoji
<jtojnar> the default config only supports few of them
<jtojnar> oh, that might be the problem
<das_j> oh my, fontconfig is such a mess
<jtojnar> also note that not all programs support all formats (hence the two twemoji packages)
<jtojnar> and you should not have fontconfig-penultimate enabled since it hates emoji with fury
<__monty__> gchristensen: I'll store the telescope, break out the microscope ; )
<gchristensen> lol
<das_j> I'm currently evaluating a container/VM solution where I can quickly spin up a testing network. I hereby regret writing my Firewall in nftables, nobody supports it :/
<samueldr> does "nf" stand for "no fun"?
<das_j> Currently looks like I'm going with Vagrant with the libvirt provider as libvirt will use firewalld if available which will use nftables if available
<das_j> samueldr: It would be a lot of fun if anybody else used it
drakonis_ has joined #nixos-chat
<pie__> iirc nftables is supposed to be good though?
<pie__> or am i just making stuff up
<pie__> also apparently firewall rules are in ebpf these days...(?)
* pie__ wonders how he can get the systemd maintainers to look at his issue without being obnoxious
<gchristensen> drop a link to it in #nixos
<gchristensen> ,pr pie__
<{^_^}> pie__: 0. Sorry. 1. If you complain about PR being in limbo, always provide a number («I have an unmerged PR (#23924)…»), bot will print PR title and it might bring you an impulse-merge. 2. Please look up who commits similar work and ping these people (or request-review), keeping up with all the PRs is not really possible.
<samueldr> wondering if pie__ meant _systemd_ maintainer or systemd maintainers _in nixos_ :)
<samueldr> (I assumed the first)
<pie__> the former :P
psyanticy has quit [Quit: Connection closed for inactivity]
Church- has quit [Quit: WeeChat 1.6]
aleph- has joined #nixos-chat
drakonis_ has quit [Ping timeout: 265 seconds]
Jackneill has joined #nixos-chat
<ashkitten> re: earlier android kernel discussion: why are oems not forced to release sources for their custom kernels?
Jackneill has quit [Remote host closed the connection]
<samueldr> "they are"
<samueldr> they generally do
<samueldr> but some don't care
<samueldr> if none did, no alternative android roms would exist, and projects like pmOS and Mobile NixOS wouldn't really either
<ashkitten> ah
<ashkitten> but they don't want to upstream anything?
<samueldr> that's it :)
<samueldr> most layers don't care about putting it in a shape that's upstreamable
<ashkitten> i dont get the point of that
<ashkitten> oh
<samueldr> it's pump and dump
<ashkitten> because they write shit code that works
<samueldr> once it "works", they go on building the next device
<ashkitten> fair enough
<ashkitten> and ofc they wouldn't bother writing generic drivers for multiple devices i assume
<samueldr> basically
<samueldr> here's an example of bad code
<samueldr> just hardcode "uid" as 1000
<samueldr> pro-tip: that breaks USER_NS
<samueldr> that's because "uid" is not simply a number, necessarily... without USER_NS it is... but with USER_NS it becomes a more complex structure :)
<ashkitten> and i assume it'd be far too much work to try and turn the garbage into workable upstream drivers?
<samueldr> I think the solution is in-between
<samueldr> not working from clean room / clean slate
<samueldr> but not necessarily starting from the bad drivers
<samueldr> I'm not sure yet, haven't worked on *that* yet
<samueldr> one thing I want to do first is make an inventory of the devices I work with, maybe there is shared hardware, and different drivers
<samueldr> that would be interesting to see if there are upgrades to those drivers
<samueldr> or even in other source trees
<ashkitten> hm
drakonis_ has joined #nixos-chat
rardiol has joined #nixos-chat
drakonis_ has quit [Quit: Leaving]
drakonis has quit [Ping timeout: 276 seconds]
drakonis has joined #nixos-chat
drakonis_ has joined #nixos-chat
__monty__ has quit [Quit: leaving]
endformationage has quit [Quit: WeeChat 2.6]
<andi-> pie__: mind sharing a link to the upstream issue?