| 19 Oct 2025 |
 matthewcroughan | Ah okay pytest actually only fails to produce devdoc a percentage of the time | 10:18:33 |
| matthewcroughan | The real error is that it's producing a bad output path | 10:18:43 |
 Alyssa Ross | ah that's a common one | 10:18:46 |
| matthewcroughan | error: output '/nix/store/2ldd5ibxb8i538v6cch3j8zkyvkq5mq9-python3.13-pytest-8.4.2-aarch64-unknown-linux-musl' is not allowed to refer to the following paths:
┃ /nix/store/m0b67b3lmjcxa8aplpl75qpb26gr5vsf-python3-3.13.8
| 10:18:58 |
| matthewcroughan | And that is because of this | 10:19:08 |
| matthewcroughan | https://github.com/NixOS/nixpkgs/pull/451506 | 10:19:09 |
| Alyssa Ross | "ac_cv_func_malloc_0_nonnull=yes"
"ac_cv_func_realloc_0_nonnull=yes"
| 10:19:28 |
| Alyssa Ross | that's the solution any time you see an rpl_malloc error | 10:19:34 |
| Alyssa Ross | If you want to be very thorough you can gate between !lib.systems.equals stdenv.buildPlatform stdenv.hostPlatform but I don't expect we support any libcs where the answer wouldn't be yes | 10:20:15 |
| Alyssa Ross | * If you want to be very thorough you can gate behind !lib.systems.equals stdenv.buildPlatform stdenv.hostPlatform but I don't expect we support any libcs where the answer wouldn't be yes | 10:20:23 |
| matthewcroughan | wasm? | 10:20:38 |
| matthewcroughan | https://github.com/virtualsquare/vde-2 | 10:21:45 |
| matthewcroughan | huh | 10:21:46 |
| matthewcroughan |
- With VDE it is possible to create a virtual network of UML machines running
on several real computer
| 10:21:52 |
| matthewcroughan | I wonder if we could make a VM test driver that is powered by UML instead of qemu | 10:22:05 |
| matthewcroughan | it doesn't look like the thing I need to patch is actually in the vde2 source tree though | 10:22:49 |
