| 12 Nov 2025 |
 aloisw | Will the attrset update thing too, given that even the Snix folks missed it? | 17:32:23 |
 raitobezarius (DECT: 7248) | it shall be | 17:32:35 |
| raitobezarius (DECT: 7248) | :D | 17:32:37 |
| raitobezarius (DECT: 7248) | in other news | 17:32:47 |
| raitobezarius (DECT: 7248) | the Value changes creates this noticeable evaluation changes | 17:33:02 |
| raitobezarius (DECT: 7248) | android32.rustc.x86_64-linux changes error message significantly (was unfree, now is unsupported)
x86_64-freebsd.cargo.aarch64-darwin goes from unsupported to broken
x86_64-netbsd.buildPackages.binutils.aarch64-darwin evals (was infrec)
x86_64-netbsd.cargo.aarch64-darwin is broken (was infrec)
x86_64-netbsd.gmp.aarch64-darwin evals (was infrec)
x86_64-netbsd.libc.aarch64-darwin same
x86_64-netbsd.mesa.aarch64-darwin unsupported (was infrec)
x86_64-netbsd.nix.aarch64-darwin evals (was infrec)
x86_64-netbsd.nixVersions.git.aarch64-darwin same
x86_64-netbsd.rustc.aarch64-darwin same
| 17:33:06 |
| raitobezarius (DECT: 7248) | we deem this an improvement | 17:33:15 |
| raitobezarius (DECT: 7248) | but we also discovered outPath changes | 17:33:26 |
| raitobezarius (DECT: 7248) | (which is the "uh-huh" moment) | 17:33:40 |
| raitobezarius (DECT: 7248) | * the Value changes creates these noticeable evaluation changes | 17:33:54 |
| raitobezarius (DECT: 7248) | my personal explanation is that the sharing broke cycles in the evaluation expression graph, so we can do more | 17:34:08 |
| raitobezarius (DECT: 7248) | the test is 2.93 → post-CL change | 17:34:16 |
| raitobezarius (DECT: 7248) | (this is done from the release-cross.nix file) | 17:34:40 |
| raitobezarius (DECT: 7248) | ok | 17:35:13 |
| raitobezarius (DECT: 7248) | guessing time | 17:35:15 |
| raitobezarius (DECT: 7248) | 
Download clipboard.png | 17:35:16 |
| raitobezarius (DECT: 7248) | (the outPath alert was a false alarm but guess why just by reading this diff) | 17:35:27 |
| raitobezarius (DECT: 7248) | (and then hate nixpkgs) | 17:35:35 |
| aloisw | How on earth does this work in pure mode to begin with? | 17:36:55 |
| raitobezarius (DECT: 7248) | i have no idea :D | 17:37:22 |
| raitobezarius (DECT: 7248) | but this is fucking cursed code that has been accepted in nixpkgs | 17:37:34 |
| raitobezarius (DECT: 7248) | i think nixpkgs CI doesn't run in pure mode actually? | 17:37:44 |
| aloisw | But surely someone runs this with fakes on their system? | 17:40:01 |
| aloisw | But surely someone runs this with flakes on their system? | 17:40:10 |
| raitobezarius (DECT: 7248) | maybe all drbd users are running without flakes | 17:40:23 |
| aloisw | …or uses fish, or zsh, and notices not hitting the cache, one would think. | 17:41:17 |
| raitobezarius (DECT: 7248) | noticing not hitting the cache is definitely an art that is hard to master | 17:41:59 |
| aloisw | (or just break in fish, for that matter, given that they apparently want bash very specifically instead of any POSIX shell, which fish diverges from somewhat) | 17:45:33 |
 just1602 | There's pepole who actually use drbd outside of ganeti users, which are like Tor and Wikimedia? 😅 | 18:31:51 |
| raitobezarius (DECT: 7248) | don't joke about this | 18:32:25 |
