| 23 Dec 2025 |
 WeetHet | Is python just broken on unstable or am I missing something | 11:39:18 |
| WeetHet | ❯ nix-shell -I nixpkgs=channel:nixpkgs-unstable -p "python3.withPackages (ps: [ ps.numpy ])"
this derivation will be built:
/nix/store/fsnqmdg1k6cvsvkb24h9bjcx1i3w80ci-python3-3.13.9-env.drv
building '/nix/store/fsnqmdg1k6cvsvkb24h9bjcx1i3w80ci-python3-3.13.9-env.drv'...
created 219 symlinks in user environment
[nix-shell:~]$ python
Python 3.13.9 (main, Oct 14 2025, 13:52:31) [Clang 21.1.2 ] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
Traceback (most recent call last):
File "<python-input-0>", line 1, in <module>
import numpy
ModuleNotFoundError: No module named 'numpy'
>>>
| 11:39:21 |
 K900 | I believe this is fixed on master | 11:49:59 |
 leona | this will be fixed with the next staging-next | 12:41:29 |
| leona | so yes, it's broken currently | 12:41:43 |
| leona | it's fixed on 25.11 | 12:41:55 |
 kdn | How do I make the external USB drives accessible to linux-builder? I'm getting permissions denied | 13:49:42 |
 emily | oh, I was waiting for the patches to be updated | 14:09:06 |
| K900 | Oh I was hoping you would do it | 14:14:02 |
| K900 | I do not understand or trust those patches | 14:14:10 |
| emily | well I showed that they were from an old release of XQuartz and that there is a newer branch of XQuartz that contains different patches but at least one of them was already upstreamed to Xorg | 14:16:15 |
| emily | and at least one of the patches we have currently is a hacky-looking revert that XQuartz hasn't shipped for years | 14:16:37 |
| emily | I don't really have the time but it should be pretty mechanical to check how many commits from the newest XQuartz tag haven't actually made it to an Xorg release since | 14:17:00 |
| emily | I suspect the answer is zero and we can just not fetch any patches | 14:17:09 |
| emily | I might be able to take a look at the commits myself in the new year though | 14:17:42 |
| kdn | I found that I had to find the bash instance referenced by linux-builder-start script and grant it full disk access | 14:32:09 |
| kdn | ok... now how to make that linux-builder accessible from other machines? I want to use it as the primary builder for the aarch64 linux | 14:40:59 |
 Ihar Hrachyshka | kdn: I think you should check this warning. The builder uses the same key for all machines. If you expose it to network, anyone will be able to mine monero run builds on it. | 15:12:07 |
|  undefined01 joined the room. | 15:23:46 |
 debtquity | were you able to figure out a workaround here? i guess I just patch it to use older version? | 23:11:03 |
 niklaskorz | it passes locally for me and I just deal with having to build it from source atm, didn't have time to chase why it's failing in hydra | 23:11:50 |
| niklaskorz | weirdly it's passing for x86_64-darwin now on hydra but aarch64-darwin still isn't | 23:13:34 |
| debtquity | i am updating my flake and it doesn't even build locally. maybe i need to clear nix store on local and remote builder | 23:14:30 |
| debtquity | what's link to the hydra build? | 23:14:53 |
| debtquity | hmm - seems it's not tailscale for me. but gopls dependency?
error: Cannot build '/nix/store/9ziiha04cxm45lsbhw2g65azcsr287a2-gopls-0.21.0.drv'.
Reason: builder failed with exit code 2.
Output paths:
/nix/store/4s7fb36fqfgdhmkjn3p1d4wglcs0avm0-gopls-0.21.0
Last 9 log lines:
> Using versionCheckHook
> Running phase: unpackPhase
> unpacking source archive /nix/store/ba7p1n83qzxd8jskja9w0yxq639z74fr-source
> source root is source
> Running phase: patchPhase
> Running phase: updateAutotoolsGnuConfigScriptsPhase
> Running phase: configurePhase
> Running phase: buildPhase
> Building subPackage ./.
For full logs, run:
nix log /nix/store/9ziiha04cxm45lsbhw2g65azcsr287a2-gopls-0.21.0.drv
| 23:25:19 |
| 24 Dec 2025 |
| debtquity | on the linux-builder, the /nix/.ro-store is at 100%. How do I clean this? Also I use an ephemeral builder, but seems this does not get cleared out | 00:24:13 |
| debtquity | #df -H
...
certs 2.0T 528G 1.5T 27% /etc/ssl/certs
/dev/vdb 1.1G 1.1G 0 100% /nix/.ro-store
overlay 42G 29M 40G 1% /nix/store
shared 2.0T 528G 1.5T 27% /tmp/shared
| 00:24:29 |
 Randy Eckenrode | Why has swift-tools-support-core started crashing the compiler (but only when the build is run non-interactively)? ☹️ | 04:37:34 |
| Randy Eckenrode | Deleting the Clang module cache fixed it. That’s annoying. | 04:47:22 |
| K900 | @emily so I looked at xquartz sources more | 07:34:00 |
