hke has quit [Ping timeout: 245 seconds]
hke has joined #nix-darwin
philr has quit [Ping timeout: 250 seconds]
hamishmack has joined #nix-darwin
johnw has quit [Quit: ZNC - http://znc.in]
johnw has joined #nix-darwin
liamdaws_ has joined #nix-darwin
liamdaws_ is now known as liamdawson_
<liamdawson_> so if pinentry-qt5 won't build on darwin, can I write an expression for a package override `foo` that uses pinentry-qt5 on Linux and pinentry-mac on Darwin?
liamdawson_ has quit [Quit: Textual IRC Client: www.textualapp.com]
liamdaws_ has joined #nix-darwin
jtojnar has joined #nix-darwin
<jtojnar> What linker do we use on Darwin?
carlosdagos has quit [Quit: Connection closed for inactivity]
philr has joined #nix-darwin
<LnL> dyld
<jtojnar> LnL: is that the system linker?
<LnL> yeah, only /usr/bin/dyld is allowed by the kernel
<jtojnar> LnL: not even for linking during compilation?
<LnL> ah no, that's a different one I think
<jtojnar> I got this out of borg https://paste.gnome.org/pcvthtuo9
<LnL> for builds it's a wrapped "${darwin.binutils-unwrapped}/bin/ld"
<jtojnar> and I see that comes from cctools
<LnL> yeah probably
nikivi has quit [Quit: ZNC is awesome]
nikivi has joined #nix-darwin
philr has quit [Ping timeout: 246 seconds]
Nikita has joined #nix-darwin
nikivi has quit [Read error: Connection reset by peer]
Nikita is now known as Guest38017
Guest38017 is now known as nikivi
tnks has joined #nix-darwin
<tnks> I always get lost every time I try to click around and find things on Hydra. Unison stopped building on Darwin, and it seems to relate to libglade no longer building.
<tnks> Can someone help me figure out the best way to hunt this down on Hydra? I can't find glade in jobset nixpkgs:nixpkgs-18.09-darwin.
<cransom> tnks: looks like it was fixed recently, https://hydra.nixos.org/build/87367133#tabs-summary
<tnks> cransom: do you mind teaching me how to fish? I never feel like I'm using Hydra's web UI correctly.
<cransom> i stuck unison in the search box on top and scrolled till i found darwin and 18.09
<tnks> cransom: cool... that's easy enough. I guess I tried to start with the darwin 18.09 build and drill down.
<tnks> (I got lost).
<cransom> i did too and then realized i was trying too hard
<clever> ive got a user on darwin, that has reinstalled nix, and then `nix-build '<nixpkgs>' -A hello -vvvv` fails with this at the end of the log: https://gist.github.com/cleverca22/78cb1d7d991f30efb526250290050289
<LnL> clever: fresh install?
<clever> LnL: he removed nix and then re-installed it
<clever> and has also restarted nix-daemon
<LnL> can you send the output of nix doctor
<LnL> I'm afraid it's https://github.com/NixOS/nix/issues/2523 tho :/
<{^_^}> nix#2523 (by periklis, 9 weeks ago, open): darwin: nix-daemon crashes due to OBJC_DISABLE_INITIALIZE_FORK_SAFETY
<clever> LnL: what is `nix doctor`? it doesnt work on my nixos machine?
<LnL> it's in 2.2
<clever> evaluating file '/nix/store/1jz25hcma179wbpi56blgajw47n5kgqd-nix-2.2.1/share/nix/corepkgs/derivation.nix'
<clever> ah, perfect, and he does have 2.2.1 on the mac
<LnL> the 2.1 installer doesn't suffer from the fork issue, but I'm afraid 2.2 might
<clever> will take a while to get a reply though
<LnL> however, I just came across this and it sounds very suspicious https://github.com/NixOS/nixpkgs/issues/41312#issuecomment-395896215
<clever> LnL: oh, you should be able to `nix-store -r` the above path i gave, and then inspect its curl dep
<clever> $ nix-store -qR /nix/store/1jz25hcma179wbpi56blgajw47n5kgqd-nix-2.2.1 | grep curl
<clever> /nix/store/w72babkd4n20qdjr3pn67r47why7j0h7-curl-7.62.0
<LnL> well yeah, but I've been trying to debug this for a while
<LnL> it's a bit of a mysterious issue
<clever> does that version of nix at least reproduce the issue? or is there state involved?
<LnL> kind of annoying that there's no consistent way to reproduce it
<LnL> ok yeah, that looks like the problem
<clever> ?
<LnL> that version has the same problem
<clever> ah
trcc has joined #nix-darwin
<LnL> gchristensen: ^ FYI, might want to avoid reinstalling darwin builders at the moment
<gchristensen> ...ack
<gchristensen> thanks
trcc has quit [Remote host closed the connection]
trcc has joined #nix-darwin
trcc has quit [Remote host closed the connection]
trcc has joined #nix-darwin
trcc has quit [Remote host closed the connection]
trcc has joined #nix-darwin
trcc has quit [Remote host closed the connection]
trcc has joined #nix-darwin
trcc has quit [Remote host closed the connection]
trcc has joined #nix-darwin
trcc has quit [Remote host closed the connection]
philr has joined #nix-darwin
philr has quit [Ping timeout: 272 seconds]
<LnL> NOOO! the daemon still dies :(
<LnL> but only when substituting from my local cache
* LnL (╯°□°）╯︵ ┻━┻
<clever> LnL: that reminds me, i setup a custom cache cache, using warp/wai/servant, and then discovered a new bug, when xz became the bottleneck
<clever> warp hard-closes http requests, if the download takes over 60 seconds
<clever> and nix streams the dl thru xz and into nar extract, and will throttle the dl if it cant unpack fast enough
<LnL> hmm, interesting
<LnL> kind of makes sense tho, otherwise it would have to be stored twice
<clever> yeah
<clever> LnL: but that also ties into the ipfs problem
<clever> if you want to share a storepath on ipfs, you currently need the entire .nar.xz on disk
philr has joined #nix-darwin
<clever> LnL: this project i worked on: https://github.com/taktoa/narfuse allows mounting a dir full of nar's at /nix/store/
<LnL> neat, so the storage is nar blobs in ipfs but the mount is the unpacked store paths?
philr has quit [Ping timeout: 250 seconds]
hamishmack has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<clever> LnL: yeah
<clever> LnL: i was able to boot a systemd-nspawn container on that fuse fs