!lheuhImcToQZYTQTuI:nixos.org

Nix on macOS

1198 Members
“There are still many issues with the Darwin platform but most of it is quite usable.” — http://yves.gnu-darwin.org200 Servers

Load older messages


SenderMessageTime
15 Mar 2026
@debtquity:matrix.orgdebtquity *

I reproduced it on my machine (aarch64-darwin)

       > Did not find version 2021.01 in the output of the command /nix/store/vamzzqsav4nnyakgcq15v4z5gpqnvmw4-openscad-2021.01/bin/openscad --version
       >
       > Did not find version 2021.01 in the output of the command /nix/store/vamzzqsav4nnyakgcq15v4z5gpqnvmw4-openscad-2021.01/bin/openscad --help
       >
       For full logs, run:
         nix log /nix/store/qdyhkilahqf4ks8h0799f3jwyjagjdnb-openscad-2021.01.drv

for whatever reason it seems openscad --version is writing to stderr instead of stdout. I don't know the internals of versionCheckHook, but I suspect it wants the output in stdout

❯ /nix/store/vamzzqsav4nnyakgcq15v4z5gpqnvmw4-openscad-2021.01/bin/openscad --version > stdout 2> stderr
❯ cat stdout
❯ cat stderr
OpenSCAD version 2021.01

I observed the same with ...openscad --help, although this flag does not output the version

07:44:23
@debtquity:matrix.orgdebtquity

i get same behavior on my machine, weird.

       > Fetching https://github.com/kylef/PathKit.git
       > Fetching https://github.com/kylef/Spectre.git
       > Fetching https://github.com/onevcat/Rainbow.git
       > Fetching https://github.com/mxcl/Version
       > error: Failed to clone repository https://github.com/onevcat/Rainbow.git:
       >     Cloning into bare repository '/nix/var/nix/builds/nix-13475-1657500268/source/.build/repositories/Rainbow-64aa5363'...
       >     fatal: unable to access 'https://github.com/onevcat/Rainbow.git/': SSL certificate OpenSSL verify result: unable to get local issuer certificate (20)
       For full logs, run:
         nix log /nix/store/bqpd1d51vvg0g5nbj6ddj52plx5m1zfn-xcodegen-2.45.3.drv
07:51:31
@samasaur:matrix.orgsamasaur

versionCheckHook is supposed to check both streams (note the 2>&1)

versionOutput="$(@envCommand@ \
    --chdir=/ \
    --argv0="$(basename "${command[0]}")" \
    "${envArgs[@]}" \
    "${command[@]}" 2>&1 \
    | sed -e 's|@storeDir@/[^/ ]*/|{{storeDir}}/|g' \
    || true)"
08:05:15
@niklaskorz:matrix.orgniklaskorzjust wanted to bump vulkan-caps-viewer and now I have to discover it's generally broken on macOS (also with the upstream binary) since the move to Qt6, nice15:51:37
@niklaskorz:matrix.orgniklaskorzlooks like QWindow now has an intermediate QContainerLayer instead of directly containing the CALayer, and since VulkanCapsViewer unfortunately isn't using QVulkanInstance but tries to find the layer by itself... :D16:12:51
@niklaskorz:matrix.orgniklaskorzwell, at least everyone gets a fix now (including upstream)16:13:03
@niklaskorz:matrix.orgniklaskorz* looks like QWindow now has an intermediate QContainerLayer in Qt6 instead of directly containing the CALayer, and since VulkanCapsViewer unfortunately isn't using QVulkanInstance but tries to find the layer by itself... :D16:13:10
@niklaskorz:matrix.orgniklaskorzmhm, actually that's an even more recent change in Qt 6.10 while upstream is on 6.9... guess I'll propose the patch to them and they can merge it whenever they proceed to 6.10+ themselves16:24:25
@isabel:isabelroses.comisabel changed their profile picture.20:55:21
16 Mar 2026
@theo-paris:matrix.org@theo-paris:matrix.org left the room.00:37:48
@kaynetik:matrix.orgkaynetik What would you do if OfBorg check for darwin is stuck for 5+ days on a new PR? 05:29:34
@k900:0upti.meK900 Nothing, that's kind of expected 05:31:06
@k900:0upti.meK900 There is not nearly enough capacity there to keep up with the queue 05:31:15
@kaynetik:matrix.orgkaynetikOh I'm new as a maintainer 05:31:19
@kaynetik:matrix.orgkaynetikSo still learning the ropes 05:31:38
@kaynetik:matrix.orgkaynetikand usually how long does a darwin-rebuild PR take to get in the queue?05:32:45
@kaynetik:matrix.orgkaynetik* and usually how long does a darwin-rebuild PR take to get executed, from the moment it gets onto the queue?05:33:10
@k900:0upti.meK900 Could be a very long time 05:33:47
@k900:0upti.meK900 It's not blocking the merge though 05:33:52
@kaynetik:matrix.orgkaynetikhttps://github.com/NixOS/nixpkgs/pull/49876605:39:38
@kaynetik:matrix.orgkaynetikwell in my case it is, unless I'm blind to it05:39:53
@kaynetik:matrix.orgkaynetik K900: thanks for quick responses, I appreciate it 🫡 05:40:30
@emilazy:matrix.orgemilythe bot comment shows the reasons it's not being merged05:42:11
@kaynetik:matrix.orgkaynetikyep, seen it. So I guess I'd have to wait for either auto-merge, or for the existing maintainer to manually merge05:42:48
@emilazy:matrix.orgemilythere is no auto-merge :)05:44:27
@kaynetik:matrix.orgkaynetik any advice then, do I create a separate PR to first make myself a maintainer, and then rework the existing one to only touch the pkgs/by-name ? 05:46:16
@emilazy:matrix.orgemilyoh, I missed that khaneliman actually hit auto-merge05:47:35
@emilazy:matrix.orgemilyit's not blocked because of Darwin, it's blocked because some of the jobs got skipped for some reason05:48:01
@emilazy:matrix.orgemilykicked CI, maybe that'll fix it05:48:18
@kaynetik:matrix.orgkaynetikyou the best! 05:48:50

Show newer messages


Back to Room ListRoom Version: 6