-
-
Notifications
You must be signed in to change notification settings - Fork 14.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Revert "NixOS Integration Tests: Enable again for darwin" #303187
Revert "NixOS Integration Tests: Enable again for darwin" #303187
Conversation
I don't understand what the problem really is here. Do you have a link to a log file with more context? Reverting this is bad because it breaks NixOS intefgration tests for macOS users which work just fine. |
The problem is that your PR causes ofborg to be unable to evaluate Nixpkgs anymore :P Unfortunately, there is nothing more than what I just edited into the body above. Here's a gist from an actual PR against Nixpkgs that ran into this: https://gist.github.com/GrahamcOfBorg/a5b0c39f39058d8995cf54f554f5aabf I'm not opposed to your PR at all! I'd just like it to not break ofborg's ability to evaluate PRs against Nixpkgs 😄 (If you have more than 32GiB of RAM, you can reproduce the issue with the same command I put in the PR body, where |
The revert was needed because CI would have failed on initial PR and has broken eval on all PRs since it was merged (the original PR was merged before ofborg finished CI checks...) You need to additionally remove this conditional in your original PR and replace it with nixpkgs/pkgs/build-support/trivial-builders/test/default.nix Lines 23 to 27 in 4e2ce36
Do please open a new PR that re-adds the systems and fixes that conditional! |
Thank you everyone for trying to clarify. I do/did understand that the patch was blocking evaluation and would need to be reverted quickly to unblock others. I do however not understand how the concatenation of the lists linux ++ darwin breaks evaluation. Thank you @lilyinstarlight for suggesting to use |
The conditional is the lines I posted ( I just woke up and am on mobile, but I'll send a diff when I'm at a computer in a few hours to more clearly illustrate |
@tfc This diff should do it, but if you open a new PR, we'll watch ofborg to make sure it passes this time: diff --git a/nixos/lib/testing/meta.nix b/nixos/lib/testing/meta.nix
index 529fe714fcf6..bdf313e5b119 100644
--- a/nixos/lib/testing/meta.nix
+++ b/nixos/lib/testing/meta.nix
@@ -36,7 +36,7 @@ in
};
platforms = lib.mkOption {
type = types.listOf types.raw;
- default = lib.platforms.linux;
+ default = lib.platforms.linux ++ lib.platforms.darwin;
description = ''
Sets the [`meta.platforms`](https://nixos.org/manual/nixpkgs/stable/#var-meta-platforms) attribute on the [{option}`test`](#test-opt-test) derivation.
'';
diff --git a/pkgs/build-support/trivial-builders/test/default.nix b/pkgs/build-support/trivial-builders/test/default.nix
index f41372d922bb..e1ed0be72bf3 100644
--- a/pkgs/build-support/trivial-builders/test/default.nix
+++ b/pkgs/build-support/trivial-builders/test/default.nix
@@ -20,11 +20,7 @@ recurseIntoAttrs {
concat = callPackage ./concat-test.nix {};
linkFarm = callPackage ./link-farm.nix {};
overriding = callPackage ../test-overriding.nix {};
- # VM test not supported beyond linux yet
- references =
- if stdenv.hostPlatform.isLinux
- then references
- else {};
+ inherit references;
writeCBin = callPackage ./writeCBin.nix {};
writeClosure-union = callPackage ./writeClosure-union.nix {
inherit (references) samples; |
@lilyinstarlight Now i get the picture. Thank you very much, i will create a PR this afternoon. |
Reverts #303150
This broke ofborg's eval
cc @tfc