<siraben>
supersandro2000: yeah I requested to join, it's mainly so that I can be pinged as well when darwin issues come up (we don't have nearly enough darwin maintainers as is)
supersandro2000 has quit [Disconnected by services]
<emily>
I guess I might have a busted old nix-daemon that doesn't support Big Sur...
<emily>
heh, `darwin-rebuild --flake` behaves strangely if your `hostname(1)` output has a domain name in it: it wants `darwinConfigurations.yuyuko.emily.moe` (not even `darwinConfigurations."yuyuko.emily.moe"`)
ahmedelgabri has quit [Ping timeout: 244 seconds]
<thefloweringash>
IIUC, there is some metadata for hiding symbols based on the deployment target (it looks like `$ld$hide$os10.12$_dispatch_queue_create_with_target`). but I don't know when it's used. It isn't applied to `_futimens`, for example
<emily>
(one Nix rebootstrap later...) is it just me or is the `inputs` argument to `darwinSystem` broken, at least with flakes? whenever I reference an input I just get an infinite recursion error in the module system
ahmedelgabri has joined #nix-darwin
ahmedelgabri has quit [Ping timeout: 244 seconds]
ahmedelgabri has joined #nix-darwin
ahmedelgabri has quit [Ping timeout: 264 seconds]
<emily>
is `error: getting attributes of path '/usr/lib/libSystem.B.dylib: No such file or directory` when the sandbox is enabled on Big Sur a known Nix issue or should I report it? couldn't find anything from a brief issue tracker search
ahmedelgabri has joined #nix-darwin
ahmedelgabri has quit [Ping timeout: 264 seconds]
ahmedelgabri has joined #nix-darwin
<emily>
well, I reported it :)
ahmedelgabri has quit [Ping timeout: 264 seconds]
<emily>
...does /etc/synthetic.d actually work in Big Sur?
<emily>
I'm not getting any of the directories
ahmedelgabri has joined #nix-darwin
Chiliparrot has joined #nix-darwin
<LnL>
thefloweringash: oh, interesting
<LnL>
emily: sandboxing is a nixpkgs issue, we need to cleanup some of the old impure deps stuff in the stdenv that's not used anymore
<stephank>
I also noticed I couldn’t get synthetic.d to work. Was wondering if it was something silly I did wrong, but maybe it’s buggy.
__monty__ has joined #nix-darwin
andremedeiros has quit [Read error: Connection reset by peer]
andremedeiros has joined #nix-darwin
<abathur>
emily stephank ah, hmm, I noticed it in the docs but don't think I bothered trying it out because I'm annoyed they released one version where they did it the dumb way and *then* bothered doing it right so that we're stuck having to support the dumb way for compatibility
<abathur>
have you tried it both with and without the .conf present?
Chiliparrot has quit [Ping timeout: 264 seconds]
Chiliparrot has joined #nix-darwin
fiddlerwoaroof has quit [Quit: Gone.]
fiddlerwoaroof has joined #nix-darwin
Siyo has quit [Quit: Bye]
supersandro2000 has quit [Ping timeout: 245 seconds]
Siyo has joined #nix-darwin
supersandro2000 has joined #nix-darwin
Ericson2314 has quit [Ping timeout: 240 seconds]
nicolas[m] has quit [Ping timeout: 244 seconds]
thefloweringash has quit [Ping timeout: 258 seconds]
siraben has quit [Ping timeout: 244 seconds]
kloenk has quit [Ping timeout: 240 seconds]
tristan[m] has quit [Ping timeout: 258 seconds]
emily has quit [Ping timeout: 268 seconds]
skrzyp has quit [Ping timeout: 265 seconds]
kalbasit[m] has quit [Ping timeout: 244 seconds]
domenkozar[m] has quit [Ping timeout: 258 seconds]
mjlbach has quit [Ping timeout: 258 seconds]
skrzyp has joined #nix-darwin
Chiliparrot has quit [Ping timeout: 244 seconds]
Ericson2314 has joined #nix-darwin
nicolas[m] has joined #nix-darwin
Chiliparrot has joined #nix-darwin
thefloweringash has joined #nix-darwin
tristan[m] has joined #nix-darwin
siraben has joined #nix-darwin
kalbasit[m] has joined #nix-darwin
mjlbach has joined #nix-darwin
kloenk has joined #nix-darwin
emily has joined #nix-darwin
domenkozar[m] has joined #nix-darwin
philr_ has quit [Ping timeout: 265 seconds]
ahmedelgabri has quit [Ping timeout: 240 seconds]
ahmedelgabri has joined #nix-darwin
ahmedelgabri has quit [Ping timeout: 264 seconds]
ahmedelgabri has joined #nix-darwin
ahmedelgabri has quit [Ping timeout: 272 seconds]
Chiliparrot has quit [Quit: My iMac has gone to sleep. ZZZzzz…]
<stephank>
Don't remember if I did, but also too scared to try at this point. :)
<abathur>
hehe, no worries
<__monty__>
LnL: Having some trouble understanding the configure error. https://git.io/JmxTi
<__monty__>
The clang is question is the one from bootstrapTools.
<__monty__>
Does this mean "System" is a missing header in the include paths?
<emily>
abathur: I had /etc/synthetic.d/nix{,-darwin}.conf and it didn't work; combining them into /etc/synthetic.conf did
<emily>
/etc/synthetic.conf didn't exist while I had synthetic.d
<abathur>
ok
<LnL>
__monty__: it's a linker error so not the header but the library (or it's stub file as a substitute)
remexre has joined #nix-darwin
<__monty__>
LnL: Right, but I thought the local libSystem.B.dylib is exposed to stage0? Isn't that what commonImpureHostDeps is about?
<remexre>
if I'm getting `Dependency "appleframeworks" not found, tried framework', from a build, anyone know what package I need? (new to darwin, heh)
<LnL>
which basically just extracts stuff from the stubs and bootstrapTools and puts them in the right place to get a dummy replacement until we have enough stuff to actually build it
<abathur>
stephank emily yeah; it's the presence of synthetic.conf. On a spare 11.3 beta w/ existing /etc/synthetic.conf for Nix: I created /etc/synthetic.d/blah.conf and rebooted and had the empty directory; I moved /etc/synthetic.conf -> /etc/synthetic.d/nix.conf and rebooted and had *neither* directory; touched /etc/synthetic.conf and rebooted and now have both dirs back
<emily>
huh
<emily>
ok, then I guess I'll have a blank synthetic.conf :)
<abathur>
I wish they'd bother taking the time to get this crap vaguely-right on the first take :/
ahmedelgabri has joined #nix-darwin
<__monty__>
LnL: Right, but it doesn't look like the bootstrapTools clang/ld have access to the stage0 Libsystem? Since this is the only path in the search list, /nix/store/6jysbyhc43sjvfiyh1bpvi1n3zbz212r-bootstrap-tools/lib/clang/7.1.0/include
<LnL>
oh! hold on
<LnL>
you're trying to build something other than the low level stdenv components using a stage0 stdenv
<LnL>
hmm never mind, that's stage1 already
<LnL>
if I start a shell for that and run NIX_DEBUG=1 ld you can see what it links against
<LnL>
for master I get -L/nix/store/hm0z4razdg703xj1dq07a5pbpaxqqacq-bootstrap-stage0-Libsystem/lib which contains libSystem.B.tbd and a bunch of symlinks like libSystem.tbd -> libSystem.B.tbd
<LnL>
along with $out/lib/system/*.tbd which are it's components that contain the actual symbols
<__monty__>
LnL: Hmm, indeed, it's the c-ares from stage2's __bootPackages, so it should be built with stage1's clang, no? Why is configure detecting stage0 clang?
<LnL>
it's building _using_ the stage0 compiler because we don't have any libraries yet to build a compiler that doesn't link against bootstrap tools
<LnL>
if you look at the stage you can see it's where we only have enough of a build environment to bootstrap some tooling like cmake, etc.
<LnL>
building anything else would be pointless since we'd need another build using a proper libc, etc. in the final stdenv
<__monty__>
Yes, but shouldn't stage1 make that compiler link to stage0's Libsystem?
<LnL>
well yes, the stages will use the bootstrap Libsystem for a few stages
<__monty__>
Hmm, this c-ares is actually a dependency of Libsystem if I'm reading the output correctly. Adding to the gist.
<LnL>
right I forgot not Libsystem specifically but other things
<LnL>
which is expected since stage0 can't compile stuff yet
<LnL>
yeah, very much looks like concurrent build output to me
<__monty__>
LnL: Should passing -j1 prevent that? Passing --cores 1 in addition doesn't change anything.
<LnL>
maybe something weird with daemon truster user settings again?
<__monty__>
I'm on a single-user install.
<LnL>
well there goes that idea :D
<__monty__>
LnL: Your reason for thinking it's a parallel build is the "building c-ares" after cpio has already failed, right?
<LnL>
yeah that line is weird
<LnL>
btw how did you end up with the c-ares from stage1? unless you changed stuff in the meantime that drv doesn't match the one from the first log
<remexre>
is there a blocker for rebuilding bash and zsh as universal binaries?
<remexre>
right now it breaks host cpu arch detection that they're not, since ./configure scripts will shell out to uname -m, and run the amd64 version of it since the shell process is amd64
philr_ has joined #nix-darwin
<__monty__>
🤦 I'd move the libSystem.tbd and libc.tbd symlinks out of the way when debugging the previous step. libSystem.B.tbd was still there so it didn't look suspicious to me 😭
<__monty__>
Ok, next thing I'm stuck on is python3-minimal. That's enough for today -_-