samueldr changed the topic of #nixos-infra to: NixOS infrastructure | logs: https://logs.nix.samueldr.com/nixos-infra/
lukegb has quit [Quit: ~~lukegb out~~]
lukegb has joined #nixos-infra
cole-h has quit [Ping timeout: 252 seconds]
<domenkozar[m]> niksnut: sorry for a late follow up, but then there's no objections shutting down wendy?
MichaelRaskin has quit [Ping timeout: 268 seconds]
<niksnut> domenkozar[m]: I'm missing some context, what are you talking about?
<domenkozar[m]> Wendy doesn't have SSE 4.2 that's breaking quite a few data science packages
<domenkozar[m]> And we'll have to retire it anway
<domenkozar[m]> We have enough machines for Linux that it shouldn't make a difference
FRidh has joined #nixos-infra
MichaelRaskin has joined #nixos-infra
FRidh has quit [Quit: Konversation terminated!]
FRidh has joined #nixos-infra
cole-h has joined #nixos-infra
<samueldr> shouldn't we instead make CPU features machine features in the builder configurations, and not send jobs requiring said CPU features on machines not having them?
<samueldr> and conversely, somehow send the installer tests sometimes on a machine with the fewest "newer" features to validate things should still work?
<lukegb> it might make sense to define what the baseline set of CPU features we expect is...
<lukegb> (although, yes, some packages will have special requirements)
<samueldr> yeah, special requirements package having higher requirements is fine (e.g. mathsy stuff)
<samueldr> we'd probably need something similar on ARM to force tests (sometimes?) to run on the less-featured cores we can
<samueldr> it's likely there is no issue, but it'd be better to be able to check
andi- has quit [Ping timeout: 268 seconds]
andi- has joined #nixos-infra
cole-h has quit [Ping timeout: 260 seconds]
andi- has quit [Ping timeout: 258 seconds]
andi- has joined #nixos-infra
FRidh has quit [Quit: Konversation terminated!]
supersandro2000 has quit [Disconnected by services]
supersandro2000 has joined #nixos-infra