{`-`} has joined #robotnix
<
henri>
danielrf[m]: I'm glad that my PRs are helpful.
<
samueldr>
so yeah, at a casual glance, it looks like the lineageos system service for updates is quite different
<
samueldr>
(you probably should, so it's obvious this channel is logged)
<
danielrf[m]>
Will do
<
samueldr>
I wonder if forcing the "system update engine" from android_system_update_engine back to the AOSP one in a LineageOS build would be enough
<
henri>
Well, so far the evidence for it not working is mostly anecdotal.
<
henri>
I literally tried installing the OTA once and then gave up.
<
danielrf[m]>
Even repeatedly trying doesn't make any progress?
<
henri>
I've tried four or five times but it always fails on a different derivation. Probably I haven't tried often enough.
<
danielrf[m]>
yep, I'm fairly sure if you just keep trying it eventually works :)
<
danielrf[m]>
great user experience, I know :D
<
henri>
Or someone could just fix Nix :)
<
danielrf[m]>
I think I only started experiencing the issue after updating the flake.lock, which updated the nixpkgs--so everything had to be rebuilt
<
danielrf[m]>
Earlier, I had a bunch of derivations previously built using `nix-build` that it was using
<
henri>
Since it was so reproducible I built Nix with debug symbols yesterday and obtained a nice stack trace.
<
colemickens>
If it's just an issue with `nix build` you could use flakes-compat to build the derivation with `nix-build` first.
<
colemickens>
(not 100% up on the context, but just throwing that out there)
<
henri>
colemickens: robotnix implements its own flake-compat (sort of)
<
colemickens>
aha nice, I see, thanks for the pointer
henri has quit [Quit: Goodbye!]