<das_j>
Weird. So do I just have to go through that building process once and hope the next version is cached?
<samueldr>
if it failed for misc. reasons and not for build issues
<das_j>
samueldr: Do you mean the 1481 failed jobs?
<samueldr>
I meant that if it wasn't because 4.18.14 had any real reason to fail consistently, which is doubtful
<samueldr>
one of the jobs I looked at timed out
<samueldr>
another one looked like some resource was exhausted
<samueldr>
but I'm not the best at finding out why it failed through the hydra logs... yet
<das_j>
But hey, apart from the pi building the kernel everything looks great so far
<das_j>
I currently have a stock NixOS, but after the kernel is done, I'll be on my usual config repo
<das_j>
Only thing that still annoys me is the handling of `nixos-rebuild --target-system me@whereever --option system aarch64-linux`. The entire process is built using qemu. Do you think there is a possibility for this to change?
<das_j>
When running nixos-rebuild in qemu, the downside is that derivation like config files are also built in the (slower) qemu and also packages that could be cross-built
<samueldr>
I've only used native on-device rebuild
<das_j>
I thought that would speed up the process by building stuff on my normal machine, but that turns out to be less optimal than expected