| 26 May 2021 |
 Finn Behrens | In reply to @sternenseemann:systemli.org
aarch64-darwin builds are sandboxed?! I think lowdown create a sandbox in an own way somehow | 15:13:39 |
 sterni (he/him) | andi-: yeah clang 7 is too old for no-semantic-interposition | 15:14:22 |
| sterni (he/him) | In reply to @kloenk:petabyte.dev
I think lowdown create a sandbox in an own way somehow …that doesn't work on aarch64-darwin, but on x86_64-darwin? | 15:14:54 |
 andi- | @sterni: Do we have a way to detect that without hardcoding yet another llvm version lookup everywhere? | 15:14:54 |
| Finn Behrens | In reply to @sternenseemann:systemli.org
…that doesn't work on aarch64-darwin, but on x86_64-darwin? seems like it, I did manage to build nix master for x86 but not for arm | 15:15:21 |
| sterni (he/him) | In reply to @andi:kack.it
@sterni: Do we have a way to detect that without hardcoding yet another llvm version lookup everywhere? better than stdenv.cc.isClang && lib.versionAtLeast stdenv.cc.version "<whatever the min version is>"? | 15:15:38 |
| andi- | Yes, I dislike Clang but it clutters all the derivations with logic like that.. | 15:16:02 |
| andi- | Companies praise Clang for being great and all.. I am just like "ugh, why?" | 15:16:20 |
| sterni (he/him) | andi-: you could also just not care about darwin for a few weeks | 15:16:26 |
| sterni (he/him) | and hope the clang upgrade is done by then | 15:16:34 |
| andi- | sterni (he/him): I have to work on a Darwin box :P | 15:16:42 |
| Finn Behrens | I lost track of the bootstrap tarball. Where is it currently. Do I still have to use github:thefloweringash/nixpkgs/aarch64-darwin-bootstrap-tools as my nixpkgs, or is it in master/staging? | 15:18:50 |
 thefloweringash | it's in master! | 15:19:09 |
| Finn Behrens | Hope ghc has support for arm soon. Don't like that I can't compile exa without a great amount of work | 15:20:02 |
| andi- | thefloweringash: do you mind popping into #infrastructure:nixos.org? I asked Rob if the builders there are also panicing. Would be good to figure that out but my Mac foo is too weak :D | 15:21:36 |
| Finn Behrens | In reply to @kloenk:petabyte.dev
nice, just have to get it into nix, as the use an own version of lowdown :-) Sadly still seeing this:
GEN doc/manual/nix-build.1
lowdown: sandbox_init: Operation not permitted
lowdown: sandbox_init: Operation not permitted
| 15:23:15 |
| sterni (he/him) | In reply to @kloenk:petabyte.dev
Hope ghc has support for arm soon. Don't like that I can't compile exa without a great amount of work it does? you just need to make sure it has llvm available | 15:26:00 |
| sterni (he/him) | someone should just add aarch64-darwin to platforms for ghc and see how it goes I guess | 15:26:18 |
| sterni (he/him) | aarch64-linux is working perfectly fine | 15:26:25 |
| Finn Behrens | oh? last time I checked the website of ghc the said the are still building some fancy new shit | 15:26:52 |
| Finn Behrens | Never really learned haskell myself though | 15:27:02 |
| Finn Behrens | thefloweringash: how was this tool you have written named? The one to sign nix pathes in a reproducible way. I looked in my DiagnosticReports folder and so quite a lot crash files | 15:28:05 |
| sterni (he/him) | Finn Behrens: https://hydra.nixos.org/eval/1673050?filter=aarch64&compare=1672940&full=#tabs-still-succeed :) | 15:28:13 |
| Finn Behrens | * thefloweringash: how was this tool you have written named? The one to sign nix pathes in a reproducible way. I looked in my DiagnosticReports folder and saw quite a lot crash files | 15:28:22 |
| thefloweringash | I think it'll show up as "sigtool" | 15:28:42 |
| Finn Behrens | In reply to @sternenseemann:systemli.org
Finn Behrens: https://hydra.nixos.org/eval/1673050?filter=aarch64&compare=1672940&full=#tabs-still-succeed :) Looks good. Will try to fix my nix, and that maybe I can build exa. yay | 15:29:01 |
| Finn Behrens | In reply to @thefloweringash:matrix.org
I think it'll show up as "sigtool" yeah, have like 6 files from it in that folder | 15:29:17 |
| sterni (he/him) | you probably need to mess with a bunch of files for it to allow aarch64-darwin though | 15:29:33 |
| sterni (he/him) | or set allowUnsupported | 15:29:38 |
| Finn Behrens | In reply to @sternenseemann:systemli.org
you probably need to mess with a bunch of files for it to allow aarch64-darwin though as long as it work, messing with files and than creating a pr would be a preferable way for me :-) | 15:30:06 |
