| 15 Nov 2022 |
 pie_ | going to try to get a dynamically linked nix on there now that I have static nix at least capable of builds | 21:10:19 |
| pie_ | oh you know what, maybe I didnt make the static-nix with musl | 21:16:50 |
| pie_ | though IDK why that would be a problem when nix-shell is used | 21:17:02 |
| pie_ | ok the problem is that the builder of pkgsStatic.hello for example isnt using the bash from pkgsStatic, its using it from the normal stdenv i thi | 23:40:24 |
| pie_ | * ok the problem is that the builder of pkgsStatic.hello for example isnt using the bash from pkgsStatic, its using it from the normal stdenv i thik | 23:40:27 |
| pie_ | * ok the problem is that the builder of pkgsStatic.hello for example isnt using the bash from pkgsStatic, its using it from the normal stdenv i think | 23:40:29 |
| pie_ | ok I figured out the / part of the problem is that nix-shell invokes stdenv.mkShell implicitly with -p so that's where that bash is coming from | 23:47:43 |
| pie_ | so that at least makes sense (Im going to temporarily hack around it by making the nixpkgs import return pkgsMusl) | 23:48:08 |
| 16 Nov 2022 |
|  FRidh joined the room. | 07:21:59 |
| FRidh | In reply to @jcie74:matrix.org
Im still not sure whether you can just -p python and a bunch of python packages or whether you basically have to withPackages it withPackages is the method to use. -p often works but not always because it relies on an implementation detail, the build-time hooks | 07:25:54 |
| pie_ | ahaa | 08:30:56 |
|  @omlet:matrix.org joined the room. | 20:33:57 |
| 19 Nov 2022 |
| pie_ | proot uh...may? not be great for performance | 01:18:16 |
| pie_ | IDK why but bison's test suite has been running for hours | 01:18:38 |
| @omlet:matrix.org left the room. | 12:13:21 |
| 23 Nov 2022 |
| pie_ | not sure if just slow af or --show-trace is broken under proot | 22:27:31 |
| pie_ | * IDK why but bison's test suite has been running for hours (might be related to fs latency?) | 22:27:47 |
| 24 Nov 2022 |
|  tpw_rules joined the room. | 02:21:39 |
| pie_ | o/ | 02:23:00 |
| tpw_rules | hi | 02:23:21 |
| pie_ | In reply to @jcie74:matrix.org
not sure if just slow af or --show-trace is broken under proot narrator: it was jsut slow af, and probably lustre | 02:23:43 |
| pie_ | tpw_rules: I am currently trying to get some machine learning stuff set up on a 2.6 kernel...mostly for kicks and geek cred, but it's been a little bumpy. | 02:24:57 |
| tpw_rules | with all due respect how are those two things related | 02:25:25 |
| tpw_rules | except for the obvious masochism factor | 02:25:31 |
| pie_ | hah.. I have figured out I mostly don't like ML . I dont like the "am I an idiot or do I just need to throw more compute at it" factor | 02:26:11 |
| pie_ | I like to be certain that im an idiot | 02:26:20 |
| tpw_rules | so you are trying with less? | 02:26:23 |
| tpw_rules | but big mood | 02:26:27 |
| pie_ | the story is I have some access to oldish uni hpc cluster so I figured Id try to un unity's mlagents toolkit on an hpc node for my "intro to reinforcement learning in practice" course | 02:27:12 |
| pie_ | * the story is I have some access to oldish uni hpc cluster so I figured Id try to run unity's mlagents toolkit on an hpc node for my "intro to reinforcement learning in practice" course | 02:27:18 |
