nixVersions.nix_2_20: init at 2.20.5; nixVersions.unstable: 2.19 -> 2.20#285264
nixVersions.nix_2_20: init at 2.20.5; nixVersions.unstable: 2.19 -> 2.20#285264lovesegfault merged 3 commits intoNixOS:masterfrom
Conversation
|
boehm-gc in nixpkgs requires this patch: https://github.com/NixOS/nix/blob/master/dep-patches/boehmgc-traceable_allocator-public.diff |
|
I've added the patch before the other commits. It's not specific to Nix, and is always valid. |
|
This pull request has been mentioned on NixOS Discourse. There might be relevant details there: |
|
Btw the |
Update nix-doc to be API compatible with Nix 2.20 when that is released. cc #285264
|
What work does this need to be ready? |
|
@nyabinary Nixpkgs CI runs the |
|
we are waiting for NixOS/nix#10023 to be included in a tag |
|
2.20.2 is now released. |
6c7a593 to
45e73d0
Compare
|
I'm both glad and sad to see that the |
|
Personally, I'm not convinced that we should provide a Nix that's not compatible with our lib. |
|
Maybe nix should run those tests, so that in the future those regressions are not only noticed after a release and the bump in nixpkgs. |
|
@SuperSandro2000 Agreed, there's even already a draft PR by Robert for that: NixOS/nix#9900 |
There was a problem hiding this comment.
I don't like repeating myself: this has to wait for a patch release from Nix upstream. People wanting the ultimate bleeding edge can apply this PR as a patch themselves. There's no need to rush the unstable package.
For the stable package, this is out of the question given the current regression atmosphere.
|
@RaitoBezarius For which patch specifically? edcb3430ef39a225aada06ef898c907d8277fbe8 and the stuff before that @roberth has added (and also appears to pass CI)? |
It seems recent, as long as this PR is bumped to this patch release and tests passes, we can proceed. |
42c5c2b to
4c66b45
Compare
I was referring to #285264 (comment)
Not part of this PR anyway and I can delegate that decision to other people.
That somehow went passed me. Updated now. |
Sure, that was statement was made proactively just in case people were about to reconsider this decision :) |
|
I couldn't force push to squash the existing commits while fixing merge conflicts, so I created a new PR: #294120 |
8742e49 to
bd698b3
Compare
bd698b3 to
9080c36
Compare
|
So, who wants to press the merge button? |
|
I guess the pr title should be updated to say 2.20.5. |
|
Should this be backported to 23.11? |
|
Usually, we don't aim to backport Because unstable's Nix in stable should be stable and too many people depend on unstable Nix thinking it's stable in general when it's not. I have no strong feeling, but I'm not keeping the pieces if it breaks someone's production. |
|
What if the it's just a new attribute 2_20, 2_21 but leaving nixVersions.unstable untouched? |
Completely fine, per backport policies! |

Description of changes
Fails to compile NixOS/nix#9888
Things done
nix.conf? (See Nix manual)sandbox = relaxedsandbox = truenix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage./result/bin/)Add a 👍 reaction to pull requests you find important.