<colemickens>
can someone on aarch64 run "nix-build -A esphome --show-trace" in nixpkgs and help me understand the output
<clever>
[clever@amd-nixos:~/apps/nixpkgs-le]$ nix-build -A esphome --show-trace --argstr system aarch64-linux
<clever>
while evaluating the attribute 'passAsFile' of the derivation 'platformio-usr-target' at /home/clever/apps/nixpkgs-le/pkgs/build-support/trivial-builders.nix:7:14:
<clever>
while evaluating the attribute 'buildCommand' of the derivation 'glibc-multi-2.30' at /home/clever/apps/nixpkgs-le/pkgs/build-support/trivial-builders.nix:7:14:
<clever>
i686 Linux package set can only be used with the x86 family.
<clever>
colemickens: i think something is forcing the multilib (32bit + 64bit x86) glibc as an input, which then breaks when your not on x86
<clever>
you can also use --argstr like that, to force an aarch64 eval, even on x86
<clever>
26 multiPkgs = pio-pkgs;
<clever>
colemickens: it might be a side effect of nearly all FHS envs using multiPkgs, why figure out which arch the thing is, when you can just throw every arch at it
<mvnetbiz_>
I saw something confusing there now I forget what it was
<mvnetbiz_>
"From there on, older devices load init from the RAM disk and newer devices load it from the /system partition." Is this saying that new androids don't use an initrd at all?
<samueldr>
"yes"
<samueldr>
it's even harder to answer
<samueldr>
it will use initrd for recovery
<samueldr>
like TWRP
<samueldr>
this is why installing TWRP on new devices is complex
<samueldr>
you need to patch your boot.img to replace the initrd
<samueldr>
though this doesn't affect us
<samueldr>
since it's not actually the bootloadr
<samueldr>
bootloader*
<samueldr>
it's the kernel that skips the initramfs
<samueldr>
we patch it out
<samueldr>
(as do postmarketOS)
<mvnetbiz_>
ok. I thought it maybe used 2 initrds or something