!UNVBThoJtlIiVwiDjU:nixos.org

Staging

397 Members
Staging merges | Running staging cycles: https://github.com/NixOS/nixpkgs/pulls?q=is%3Apr+is%3Aopen+head%3Astaging-next+head%3Astaging-next-25.11 | Review Reports: https://malob.github.io/nix-review-tools-reports/129 Servers

Load older messages


SenderMessageTime
8 Jul 2026
@vcunat:matrix.orgvcunat* (wasted many thousand builds because eval wasn't succeeding for many hours)08:07:09
@grimmauld:m.grimmauld.deGrimmauld (any/all) you know what package? Because now it can't be numpy/scipy 08:07:43
@hexa:lossy.networkhexaprobably torchaudio08:08:01
@hexa:lossy.networkhexabut that's on master, soooo08:08:11
@grimmauld:m.grimmauld.deGrimmauld (any/all)ooooh master08:08:19
@grimmauld:m.grimmauld.deGrimmauld (any/all)yeah master doesn't have the hook yet08:08:24
@grimmauld:m.grimmauld.deGrimmauld (any/all)torchaudio gets the hook propagated from scipy08:08:41
@vcunat:matrix.orgvcunat I think I saw *torch* do such things in the past. 08:11:16
@vcunat:matrix.orgvcunat * I think I saw *torch* do such things in the past. (consume much more CPU than "allowed") 08:11:42
@hexa:lossy.networkhexahttps://github.com/NixOS/nixpkgs/pull/53950108:13:09
@hexa:lossy.networkhexalet's see what the bot thinks of me mangling this series08:13:50
@hexa:lossy.networkhexadeleted four new packages, which apparently are not on 26.0508:14:09
@hexa:lossy.networkhexawe sure move fast08:14:12
@raitobezarius:matrix.orgraitobezarius Scrumplex thanks to Sergei Zimmerman (xokdvium), curl did repair the issues we saw with Lix in https://github.com/curl/curl/pull/22274 fwiw 09:31:36
@vcunat:matrix.orgvcunat I suppose that's too little tested to catch the default curl in the current staging-next-26.05 09:48:17
@scrumplex:duckhub.ioScrumplexI'll follow the PR discussion. If they decide to release this in a patch release we can pick it up then09:53:57
@raitobezarius:matrix.orgraitobezariusyeah, it's too fresh10:47:39
@vcunat:matrix.orgvcunat I assume that staging-next will be fine to merge within several hours. 11:05:30
@vcunat:matrix.orgvcunatDone 💥14:48:12
@hexa:lossy.networkhexafinally14:48:53
@vcunat:matrix.orgvcunat5k builds still queued, but that's not much.14:49:30
@hexa:lossy.networkhexa2.7 hours worth14:49:38
@hexa:lossy.networkhexahttps://hydra.nixos.org/build/33670667914:54:00
@hexa:lossy.networkhexa1m45s, so not worth the fuzz14:54:11
@r-burns:matrix.orgr-burnsCould someone outline (or point me to existing docs containing) rough rules of thumb for how long e.g. a full stdenv rebuild, or a full nixos test rebuild + rerun, currently takes on our infra? I think I had a decent intuition for it on our previous sponsored build infra but unfortunately that's no more 😢14:54:18
@hexa:lossy.networkhexaRedacted or Malformed Event14:54:21
@hexa:lossy.networkhexaa full rebuild of the stdenv or from up from and including the stdenv?14:55:11
@hexa:lossy.networkhexaRedacted or Malformed Event14:55:22
@hexa:lossy.networkhexabecause building stdenv takes low single digit hours, we do it in nixpkgs/staging on hydra regularly14:55:45
@hexa:lossy.networkhexa beyond that there is the rate of completed steps 14:57:39

Show newer messages


Back to Room ListRoom Version: 6