!lheuhImcToQZYTQTuI:nixos.org

Nix on macOS

1137 Members
“There are still many issues with the Darwin platform but most of it is quite usable.” — http://yves.gnu-darwin.org181 Servers

You have reached the beginning of time (for this room).


SenderMessageTime
20 Oct 2025
@toonn:matrix.orgtoonn Maybe the result of the check is cached? And the kernel doesn't find the relevant certificate to do the check itself? 09:03:23
@ivy:faggot.shivyim assuming that09:43:22
@ivy:faggot.shivybut it’s odd09:43:28
@ivy:faggot.shivythat you have to «  check » to make it work09:43:42
@ihar.hrachyshka:matrix.orgIhar HrachyshkaSomeone posted copapps for HM: https://github.com/nix-community/home-manager/pull/8031 11:03:18
@ihar.hrachyshka:matrix.orgIhar Hrachyshka* Someone posted copyapps for HM: https://github.com/nix-community/home-manager/pull/8031 11:03:29
@kolargol00:matrix.org@kolargol00:matrix.org left the room.11:40:01
@saiko:knifepoint.netKatalin 🔪 changed their profile picture.16:55:12
@tiferrei:tiferrei.comtiferrei changed their display name from Tiago Ferreira to tiferrei.17:34:14
@niklaskorz:matrix.orgniklaskorzKosmicKrisp has been merged into mesa 3 hours ago 👀21:41:55
@niklaskorz:matrix.orgniklaskorzhttps://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/3752221:42:47
@samasaur:matrix.orgsamasaurhttps://www.macrumors.com/2025/10/20/ios-26-1-liquid-glass-toggle/22:34:08
@reckenrode:matrix.orgRandy EckenrodeIs this like the toggle where you can make it look like someone peed on your screen or not, and Apple defaults to the wrong one?22:35:14
@emilazy:matrix.orgemilyhey I like True Tone22:36:17
@niklaskorz:matrix.orgniklaskorz
In reply to @emilazy:matrix.org
hey I like True Tone
same
22:36:25
@reckenrode:matrix.orgRandy Eckenrode😅22:36:38
@emilazy:matrix.orgemilytechnology is there to save us from having to go back to the bad old days of having to pee on the screen manually.22:37:48
@niklaskorz:matrix.orgniklaskorzI used to have a surface book that came with a heavily green tint ootb22:45:58
@niklaskorz:matrix.orgniklaskorz Manually calibrating it only slightly improved the situation 22:46:18
@reckenrode:matrix.orgRandy EckenrodeIs there any documentation on building it? Do we just enable the driver, and it Just Works™ if you have the required SDK?23:23:45
21 Oct 2025
@reckenrode:matrix.orgRandy EckenrodeI’ve got the source release updates building on aarch64-darwin. I’m glad I did some of this work back in July ….02:26:25
@reckenrode:matrix.orgRandy EckenrodeOnce I confirm x86_64-darwin builds, I’ll split up the changes and open a PR.02:26:43
@rgb:dreyeck.freedombox.rocksRalf joined the room.06:15:26
@reckenrode:matrix.orgRandy Eckenrode

ICU tests fail on x86_64-darwin. ☹️

Triggered by Thread: 0

Exception Type:    EXC_ARITHMETIC (SIGFPE)
Exception Codes:   0x0000000000000001, 0x0000000000000000

Termination Reason:  Namespace SIGNAL, Code 8, Floating point exception: 8
Terminating Process: exc handler [89045]


Thread 0 Crashed:
0   libicui18n.76.1.dylib         	       0x10e162427 icu::Calendar::roll(UCalendarDateFields, int, UErrorCode&) + 1735
1   intltest                      	       0x10465c2f5 CalendarTest::Test22633RollTwiceGetTimeOverflow()::$_0::__invoke(icu::Calendar*, UCalendarDateFields) + 69
2   intltest                      	       0x10465bd35 CalendarTest::RunTestOnCalendars(void (*)(icu::Calendar*, UCalendarDateFields)) + 565
3   intltest                      	       0x1046e271c IntlTest::runTestLoop(char*, char*, char*) + 492
4   intltest                      	       0x1046e2320 IntlTest::runTest(char*, char*, char*) + 128
5   intltest                      	       0x1046e2152 IntlTest::callTest(IntlTest&, char*) + 82
6   intltest                      	       0x1046eee00 IntlTestFormat::runIndexedTest(int, signed char, char const*&, char*) + 4016
7   intltest                      	       0x1046e271c IntlTest::runTestLoop(char*, char*, char*) + 492
8   intltest                      	       0x1046e2320 IntlTest::runTest(char*, char*, char*) + 128
9   intltest                      	       0x1046e2152 IntlTest::callTest(IntlTest&, char*) + 82
10  intltest                      	       0x1046f16ed MajorTestLevel::runIndexedTest(int, signed char, char const*&, char*) + 461
11  intltest                      	       0x1046e271c IntlTest::runTestLoop(char*, char*, char*) + 492
12  intltest                      	       0x1046e2320 IntlTest::runTest(char*, char*, char*) + 128
13  intltest                      	       0x1046e5360 main + 3568
14  dyld                          	       0x2050c5781 start + 3457
12:12:09
@dbaynard:matrix.orgdbaynard

Hello folks, I want to PR darwin support for libtree (analysis of linked libraries) and I have questions about tests.

There's a passthru.tests attrset at https://github.com/NixOS/nixpkgs/blob/5e2a59a5b1a82f89f2c7e598302a9cacebb72a67/pkgs/by-name/li/libtree/package.nix#L28-L46 which tests ls from coreutils. On macos, by default, it checks the macOS build of ls, which is not an ELF. I can get it to work properly using pkgsCross, as in the following.

    (libtree.override {
      inherit (pkgsCross.aarch64-multiplatform) coreutils;
    }).overrideAttrs
      (old: {
        meta = lib.recursiveUpdate (old.meta or { }) { platforms = lib.platforms.darwin; };
      })

I presume that in this case libtree should always test the pkgsCross coreutils… but certainly the tests fail on macos without it. What's the idiomatic way of handling this case?

16:26:19
@dbaynard:matrix.orgdbaynard *

Hello folks, I want to PR darwin support for libtree (analysis of linked libraries) and I have questions about tests.

There's a passthru.tests attrset at https://github.com/NixOS/nixpkgs/blob/5e2a59a5b1a82f89f2c7e598302a9cacebb72a67/pkgs/by-name/li/libtree/package.nix#L28-L46 which tests ls from coreutils. On macos, by default, it checks the macOS build of ls, which is not an ELF. I can get it to work properly using pkgsCross, as in the following.

    (libtree.override {
      inherit (pkgsCross.aarch64-multiplatform) coreutils;
    }).overrideAttrs
      (old: {
        meta = lib.recursiveUpdate (old.meta or { }) { platforms = lib.platforms.darwin; };
      })

I presume that in this case libtree should always test the pkgsCross coreutils… but certainly the tests fail on macos without it. What's the idiomatic way of handling this case?

[Edit: I don't mean idiomatic way of implementing the changes, as opposed to the override, I mean the idiomatic way of handling the cross-platform dependency in the tests.]

16:31:14
@lt1379:matrix.orgLun joined the room.18:35:15

Show newer messages


Back to Room ListRoom Version: 6