coreutils: switch to version 9 for darwin#143800
coreutils: switch to version 9 for darwin#143800happysalada merged 1 commit intoNixOS:staging-nextfrom
Conversation
ofborg
bot
added
the
10.rebuild-darwin-stdenv
ofborg
bot
added
10.rebuild-darwin: 501+
|
Ok just finished testing on darwin-x86_64. Leaving this for one more day open, just in case people want to give feedback. |
|
I thought we wanted to do this in the next cycle as the Darwin workers on hydra are broken/unstable again. (https://matrix.to/#/!UNVBThoJtlIiVwiDjU:nixos.org/$LB3d9bKrXy-aove7ifYFaJPhqT4vEBro15VNlGeEEY8?via=matrix.org&via=helsinki-systems.de&via=nixos.org) |
|
|
||
| # coreutils 9 messes up some binary format on darwin | ||
| # https://github.com/NixOS/nixpkgs/pull/141684 | ||
| # test by trying to build alacritty for example |
There was a problem hiding this comment.
Anyone happen to have an example which doesn't have quite as massive dependency/rebuild footprint for testing?
There was a problem hiding this comment.
r-burns was saying jx was the fastest one for him.
My machine is quite old, it still took me 6 hours I think.
If you leave it overnight, it should be fine.
I made sure by also testing alacritty.
There was a problem hiding this comment.
No, my jx testing was all on linux. I didn't encounter any issues on darwin (that goes for alacritty as well).
|
@mohe2015 I'm happy to do it in the next cycle. it would be a shame to miss the update I felt. |
|
Others please also comment. I'm not a good hydra reader but it seems like only one x86 Darwin builder is working and two are hanging so I assume the reasoning is still valid. I don't know though how long it would take or so. |
|
Was the original issue why this was reverted fixed? If so how? I didn't find any information about that in this PR. |
|
As far as I'm aware, @happysalada was the only one to reproduce the darwin errors locally, so if those are fixed then this is good to go IMO. But if we want to wait until the next staging cycle, that's fine by me. |
|
@SuperSandro2000 the original issue was that the resulting rust binaries had the wrong binary format and couldn't be executed. With the coreutils fix, this doesn't happen anymore (this is what was tested in this PR). |
|
You mean with 924ccbf#diff-88e1a963530d2a05ed02f38aa91b0a292d425396118e1d1a7b05e9132d4f7f5e? Or which fix exactly? |
|
@SuperSandro2000 yes, that fix. |
|
I'm going to say let's merge this one, the build queue is almost empty, and it removes weird logic that is unecessary. |
5 participants
Motivation for this change
switch darwin to version 9 of coreutils.
Currently testing this.
Things done
sandbox = trueset innix.conf? (See Nix manual)nix-shell -p nixpkgs-review --run "nixpkgs-review wip"./result/bin/)