<aterius>
Is there a better way to more generally override nix to use nixUnstable with nix-darwin? it seems like every time I enter a shell, unless I explicitly at nixUnstable as a build input, it reverts to 2.3.7
<abathur>
I'm not certain, but I seem to recall someone saying you could do something like `nix.package = nixpkgs.nixUnstable;`, maybe in ~/.config/nixpkgs/config.nix?
<aterius>
That's what I'm currently doing
<aterius>
Unfortunately any time I enter a nix-shell, or use nixpkgs-review it uses nix 2.3.7 :/
<abathur>
ah
<abathur>
I haven't tried to daily drive it yet; might be worth searching around github for references to nixUnstable if you haven't already; might stumble onto someone who has it figured out
<aterius>
Always my first step :)
<abathur>
:]
xcmw has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<thefloweringash>
in my extremely special branch of nixpkgs, I'm seeing some qt apps fail to link. The failure seems to be something passing a framework directory directly to the linker, which rightly complains that it can't open the directory as a file.
<thefloweringash>
haven't looked in detail yet, but does this ring any bells for anyone?
<thefloweringash>
it's only the frameworks OpenGL and AGL, which do have some kind of special handling in qt
<thefloweringash>
testing by building my nix-darwin system closure on a big sur machine, it builds quite a lot of packages including ghc, rustc, go, etc. only thing I know to be broken is qt (described above)
<LnL>
oh wow, nice work!
<LnL>
are we passing special stuff for OpenGL or is it the qt build system that does that?
peel has quit [Ping timeout: 244 seconds]
peel has joined #nix-darwin
cbarrett has quit [Ping timeout: 240 seconds]
sparogy has quit [Ping timeout: 244 seconds]
cbarrett has joined #nix-darwin
siraben has quit [Ping timeout: 240 seconds]
cptrbn has joined #nix-darwin
thefloweringash has quit [Ping timeout: 244 seconds]
emily has quit [Ping timeout: 240 seconds]
siraben has joined #nix-darwin
thefloweringash has joined #nix-darwin
sparogy has joined #nix-darwin
emily has joined #nix-darwin
cptrbn has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
cptrbn has joined #nix-darwin
cptrbn has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
eraserhd has quit [Quit: WeeChat 2.9]
<abathur>
<3 thefloweringash
<{^_^}>
thefloweringash's karma got increased to 13
cptrbn has joined #nix-darwin
cptrbn has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
elvishjerricco has quit [Ping timeout: 272 seconds]
elvishjerricco has joined #nix-darwin
<LnL>
abathur: did you run into problems with zshrc/bashrc? I remember also being a bit confused about it but everything worked fine on a clean install
<abathur>
yes
<abathur>
I haven't quite done an install attempt on my latest fix attempt, but there was a PR that changed the profile "target" for zsh to /etc/zshenv instead of /etc/zshrc, but it didn't do anything about the fact that the code checks existence before it writes
<LnL>
ah was that one after the update changes?
<abathur>
one of them, there have been a few
<abathur>
pushed one attempt yesterday but it face-planted, so I force-pushed a fix to the same commit
<abathur>
I think it's the latest one in the PR
<abathur>
I'm trying to sift through some stuff about apfs snapshots
<abathur>
before I try again
<abathur>
I want to take a snapshot and be able to restore to it so that I can iterate on clean attempts faster
<LnL>
oh interesting, totally didn't think about that but I bet that's possible
<abathur>
and I did that yesterday, but the system eats local snapshots over 24h old
<abathur>
yeah, `tmutil localsnapshot /` is sufficient to take one, and then you can restore it (but the UI is wonky/janky at least in big sur) from recovery
<abathur>
but there's an option in apfs.util that is taunting me like it should be able to support a better workflow here
<abathur>
if I can just figure out wtf :]
<LnL>
yeah, all that stuff is pretty hidden and not particularly documented AFAIK
<abathur>
so .../apfs.util -R SNAPSHOT is described as "set the volume to revert to the snapshot named by SNAPSHOT on next mount."
<abathur>
there's also some stuff about reading/editing "purgeability" in apfs.util and I vaguely wonder if there's a way to use that to protect the initial snapshot, but that seems like a secondary concern
<LnL>
is it actually a problem?
<LnL>
the ui might freak out a bit but the installer mostly only touches unixy stuff
<abathur>
if I could get it to just take the snapshot over a simple reboot, I could script something to make it take a fresh snapshot *and* immediately flag it for use on reboot, which should do what I need and only fall over if I keep it booted for more than 24h
<abathur>
which part?
<abathur>
oh huh
<abathur>
I might've found an even nicer way
<abathur>
*still have to solve the syntax quest I think
cptrbn has joined #nix-darwin
<abathur>
I don't know if it'll be more durable than manually-taken localsnapshots, but I see a /dev/disk1s5s1 that I didn't recognize, and when I run `diskutil info /dev/disk1s5s1` it does note at the end that "This disk is an APFS Volume Snapshot"
<abathur>
and then gives its name as com.apple.os.update-...hash?...
cptrbn has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<abathur>
ah, I guess it's probably just of the system volume
xcmw has joined #nix-darwin
<antifuchs>
ok so this time I ran the nix-darwin installer on a fresh machine with an explicit `services.nix-daemon.enable = false;` in the default config set, and it still reset my nix config to be multi-user
<antifuchs>
wtf
<antifuchs>
I think I should make myself a nix-darwin bootstrap script ):
philr has quit [Ping timeout: 265 seconds]
cptrbn has joined #nix-darwin
__monty__ has joined #nix-darwin
<LnL>
is this a fresh install? otherwise your db permissions are probably wrong
<antifuchs>
I was installing nix-darwin on top of a nix install that I'd performed a few weeks ago, so I'm not completely certain it was single or multi-user... the installer just created the N build users that I have come to associate with the nix-daemon