bennofs is now known as ALLES|bennofs
ALLES|bennofs is now known as bennofs
h0m1 has quit [Ping timeout: 246 seconds]
h0m1 has joined #nixos-aarch64
evils has quit [Remote host closed the connection]
evils has joined #nixos-aarch64
wavirc22 has quit [Ping timeout: 240 seconds]
wavirc22 has joined #nixos-aarch64
evils has quit [Quit: Lost terminal]
evils has joined #nixos-aarch64
wavirc22 has quit [Quit: ZNC 1.7.2+deb3 - https://znc.in]
evils has quit [Quit: Lost terminal]
evils has joined #nixos-aarch64
evils has quit [Client Quit]
evils has joined #nixos-aarch64
mvnetbiz_ has quit [Read error: Connection reset by peer]
evils has quit [Quit: Lost terminal]
evils has joined #nixos-aarch64
leonardp has joined #nixos-aarch64
<leonardp> hello! `nix build -f channel:nixos-19.09 pkgsCross.aarch64-multiplatform.firefox` fails for me with: `error: attribute 'llvmPackages_7' missing` it also fails on nixpkgs-unstable
<leonardp> can anyone shed some light on this?
<tilpner> > builtins.attrNames pkgsCross.aarch64-multiplatform.targetPackages
<{^_^}> [ "__raw" "recurseForDerivations" "stdenv" ]
<tilpner> > :p with lib; unique (map (p: attrNames p.targetPackages) (attrValues pkgsCross))
<{^_^}> [ [ "__raw" "recurseForDerivations" "stdenv" ] ]
<tilpner> So it's the same for all, but llvmPackages_7 expects it to contain more
<leonardp> i don't think i can follow you there. i am relatively new to nixos
<tilpner> I can't really help you, but it doesn't appear to be specific to aarch64
<leonardp> ah, that is already good to know :) thanks for investigating!
<tilpner> It also happens when naively importing nixpkgs with crossSystem set, so without pkgsCross
<tilpner> I could be doing it wrong, but perhaps the top-level llvmPackages_7 definition is not cross-aware?
* tilpner just saw the same conservation on #nixos
<tilpner> Oh, well that's awkward
<leonardp> hehe, #nixos sent me here, but it appears to be related to something with patchelf
* clever waves
<clever> tilpner: i'm not entirely sure what its doing, but the llvmPackages_7 in all-packages.nix references targetPackages.llvmPackages_7, which implies it is cross-aware
<tilpner> I don't really know what I'm talking about, and should probably stop
<clever> yeah, i also dont fully grasp how all of the cross stuff works together
<clever> even after adding a new cross target
<leonardp> oh wow :)
<leonardp> suddenly chrome seems like a viable option for my pinebook
wavirc22 has joined #nixos-aarch64
<leonardp> clever: i saw your binfmt wrapper overlay mentioned in the wiki, is this still needed/functional or is it sufficient to set `boot.binfmt.emulatedSystems`?
<clever> leonardp: ive not gotten around to seeing what boot.binfmt.emulatedSystems does exactly
<leonardp> well, do you know if your overlay working/usable atm?
<clever> leonardp: its still active on my laptop, but i dont remember when exactly i last did an update
<leonardp> so, i could use it to configure my desktop machine as a remote builder for my arm hardware?
<clever> leonardp: yeah
<leonardp> that sounds really nice, i think i'll try that approach :)
<leonardp> thanks!
<thefloweringash> Is the idea here to emulate x86_64 to run chrome?
<thefloweringash> I hope I’ve misread scroll back
<leonardp> nah, i got a pinebook pro (arm laptop) this week, and would like firefox/chrome on it
<leonardp> and *normal* cross-compilation unfortunately appears to be broken right now for both
<thefloweringash> If it helps, chromium builds natively
<clever> leonardp: aarch64 or arm32?
<thefloweringash> the pro is rk3399, aarch64
<clever> leonardp: hydra should be pre-compiling things for aarch64
<thefloweringash> chromium always times out after 2 days :( https://hydra.nixos.org/job/nixpkgs/trunk/chromium.aarch64-linux
<leonardp> thefloweringash: i am building chromium right now natively, but if i can i would like to avoid that
<leonardp> clever: it automatically tries to build these so i was assuming they were not in the binary cache. am i doing something wrong?
<clever> leonardp: what thefloweringash said
<leonardp> clever: do nixops deployments work in cunjunction with your overlay?
<clever> leonardp: should work just fine, just use nixpkgs.overlays in the nixos setup
<leonardp> clever: great! thanks for the heads up!
<thefloweringash> I thought I saw changes to the arm build servers that would make `big-parallel` be sufficient to build chromium, but hydra is still producing timeouts.
<leonardp> i my optionion the problem lies in the compile times of ff/chromium not in the hydra config...
<clever> leonardp: the hydra config can point it towards a machine that dedicates more cores to a single build
<leonardp> that makes sense, but i find the compile times ridiculous nonetheless
<srk> leonardp: +1. as a backup you can use netsurf .. nix-build -E '(import ./. {}).netsurf.browser.override{uilib = "gtk";}'
<srk> I didn't even try building firefox on armv7 as I know there's no much hope to build it anyway, using netsurf and x11spice/spicy from spice-gtk instead :D
<thefloweringash> I don't think all hope is lost for firefox on armv7. I think archlinuxarm does it with some hacks https://github.com/archlinuxarm/PKGBUILDs/blob/c463da744e52f6559e42f0ed551c9c42efaa05d6/extra/firefox/PKGBUILD#L115-L133
<srk> oh cool
<thefloweringash> my rk3328 used to build chromium in ~3 days
<v0|d> thefloweringash: any idea why gcc fails with internal error on 3328?
<v0|d> I've got cc1 processes stuck with running time 37mins.
<v0|d> this is another symptom.
<v0|d> maybe it is broken, not sure, borrowed it to play w/.
<thefloweringash> v0|d: can't say I've seen that sorry. I have been running mainline kernel for quite a while now though
<v0|d> thefloweringash: do I need anything special to run mainline?
<v0|d> thefloweringash: I'm using your blobs from the repo. rkbins.
<v0|d> thefloweringash: also the uboot from nixpkgs.
<v0|d> maybe I need to update the blobs.
<thefloweringash> I'm still carrying around a few extra patches I borrowed from archlinuxarm: https://github.com/archlinuxarm/PKGBUILDs/blob/c463da744e52f6559e42f0ed551c9c42efaa05d6/extra/firefox/PKGBUILD#L115-L133
<thefloweringash> uh, wrong link sorry
<leonardp> clever: as far as i can judge `boot.binfmt.emulatedSystems` does the same as your overlay. if you have time maybe you can confirm it?
<clever> leonardp: that is on my todo list, maybe later today or tomorrow
<misuzu> leonardp: i did test boot.binfmt.emulatedSystems for armv7l and it was 3x slower than building natively
<misuzu> (on armv7l)
<clever> misuzu: the main benefit, is that you may have more cores, and more x86 hardware in general
<clever> though, for a single derivation that takes 3 days normally (chromium), that would be 9 days, for an equal core-count..
<leonardp> misuzu: but i get more than x2 the cores and x10 times the ram and dont have to do any swapping
<clever> yep
<v0|d> thefloweringash: nothing seems special, I'm on sd so don't need usb3, thnx anyway.
leonardp has quit [Remote host closed the connection]
leonardp has joined #nixos-aarch64
t184256 has left #nixos-aarch64 ["Error from remote client"]
t184256 has joined #nixos-aarch64
bennofs has quit [Quit: No Ping reply in 180 seconds.]
bennofs has joined #nixos-aarch64
leonardp has quit [Remote host closed the connection]
ZoomZoomZoom has joined #nixos-aarch64
ZoomZoomZoom has quit [Ping timeout: 256 seconds]
ZoomZoomZoom has joined #nixos-aarch64
<sphalerite> clever: and better IO!
ZoomZoomZoom has quit [Ping timeout: 256 seconds]
v0|d has quit [Remote host closed the connection]
v0|d has joined #nixos-aarch64