| 10 Apr 2024 |
 Lily Foster | In reply to @cole-h:matrix.org
It's a race then! 😆 yeah i started mine before you stole an ofborg evaluator | 19:27:09 |
| Lily Foster | but the ofborg evaluators are much nicer so might finish first | 19:27:22 |
 cole-h | eval is single-threaded anyways so I think the only benefit will be the 256GiB of memory lol | 19:28:12 |
| Lily Foster | In reply to @cole-h:matrix.org
eval is single-threaded anyways so I think the only benefit will be the 256GiB of memory lol exactly, i'm about to be swapping here in a minute :) | 19:28:25 |
| Lily Foster | (also am on laptop on battery. maybe i should fix the on-battery part) | 19:28:42 |
| cole-h | Last time I tried on my machine and was swapping (without zram at the time, though) it didn't even finish in 12 hours so... Good luck? 😆 | 19:29:01 |
| Lily Foster | oooof | 19:29:07 |
| cole-h | * Last time I tried on my machine and was swapping (without zram at the time, though) it didn't even finish in under 12 hours so... Good luck? 😆 | 19:29:11 |
| Lily Foster | it's usually not too bad here, but does take like 15 mins | 19:29:17 |
| Lily Foster | yeah plugging in power seems to have sped it up significantly, judging by the speed of memory usage growth | 19:29:38 |
| cole-h | . o O ( GitHub adding a "sort by merge date" feature when ) | 19:30:23 |
| cole-h | Definitely didn't forget to actually..... do the revert, so I at least validated that master @ 85d3f97277c0a0b5b542951e8990b4bd2bbe2a26 still fails 😅 | 19:46:45 |
| cole-h | Eval with revert is still running... I'll open a revert PR if that does the trick (if you beat me to it, feel free) | 19:47:18 |
| cole-h | * Eval with revert is still running... I'll open a revert PR if reverting does the trick (if you beat me to it, feel free) | 19:47:31 |
| cole-h | https://github.com/NixOS/nixpkgs/pull/303187 | 19:49:49 |
| cole-h | Yep, revert fixed it. | 19:49:52 |
| cole-h | Self-merged to hopefully limit the amount of new PRs that will run into this | 19:52:45 |
 K900 | Unstable-small started 5 minutes ago | 19:55:22 |
| * K900 kicks | 19:55:27 |
| cole-h | In reply to @me:indeednotjames.com
eval fails for (what looks like) all new PRs with https://gist.github.com/GrahamcOfBorg/455a82c8ebc61cdd44e3a3f5c31502ea and I can't make out what caused this 👀 Thanks for noticing and bringing to our attention here! 💜 | 20:04:17 |
| cole-h | Aaand most of the failing PRs have been notified (I took it upon myself to update the r-ryantm PRs, but the rest I just left a comment on) | 20:09:02 |
 emily | Thanks a lot for looking into this and finding the cause :) | 20:11:15 |
| Lily Foster | well i'm still trying to find the cause | 20:11:37 |
| Lily Foster | as in, why that PR would ever cause that error... | 20:11:49 |
| cole-h | Yeah, found the problem, more like x) | 20:11:55 |
| emily | suuureee. Thanks for finding the PR that caused this then :p | 20:12:07 |
| cole-h | I'm glad my gut feeling was right 😆 If it wasn't that commit, I would have been in for an arduous bisect (every eval takes like 10m even on the beastly ofborg machines......) haha | 20:12:10 |
| Lily Foster | for what it's worth, the attrs that are giving this error are tests.trivial-builders.references.aarch64-darwin and tests.trivial-builders.references.x86_64-darwin | 21:17:30 |
| Lily Foster | i don't get why the initial PR didn't fail ofborg tbh? | 21:18:26 |
| Lily Foster | actually, let me check what step it failed | 21:18:32 |
