| 16 Jun 2021 |
 Sarah Hodne | After a bunch of changes to my environment it seems git now starts fast again, so I'm not sure what was broken :/ | 15:54:59 |
| Sarah Hodne | * After a bunch of changes to my environment it seems Nix git now starts fast again, so I'm not sure what was broken :/ | 15:54:59 |
 LnL | maybe something with configuration or plugins? | 16:37:34 |
|  rosariopulella joined the room. | 18:52:07 |
 toonn | I *think* I managed to update Security. Not sure why it was on an older version. | 19:53:15 |
| toonn | Gave up on hfs for now though : / | 19:53:24 |
| LnL | yeah sounds ok to use the old version, either that or Kernel.framework if it's not during bootstrapping | 19:56:08 |
| toonn | LnL: Ah, that's probably what boot.nix in Security is about? How do I test its default.nix? Build darwin.Security after stdenv builds? | 20:01:59 |
| LnL | I think so yes | 20:02:26 |
| LnL | it should be the default while boot can only be referenced through stdenv.__bootPackages...darwin.Security | 20:03:33 |
| 17 Jun 2021 |
 John Ericson | thefloweringash: did you encounter errors with the same -arch ... being passed twice? | 00:55:26 |
| John Ericson | I recall that you did | 00:55:28 |
| John Ericson | i am looking at 20.03 cross with still clang 7 | 00:58:23 |
| John Ericson | looks like the issue is old clang is issuing -arch, there is also manual -arch | 00:58:41 |
| John Ericson | and new clang is having none of it | 00:58:48 |
| John Ericson | ah nevermind, looks like I just need to skip ``-sdk_version` | 01:40:21 |
| John Ericson | * ah nevermind, looks like I just need to skip -sdk_version | 01:40:25 |
 thefloweringash | I don’t remember the general problem of -arch being passed twice. I think if both values are the same there’s no warning or error, but a couple of packages were hard coding values which conflicted with the wrapper value. | 04:31:44 |
| thefloweringash | As for -sdk_version, that’s carefully handled by the logic added in “Darwin platform versions” PR in the ld wrapper. It’s valid to pass an explicit -sdk_version to the wrapper, and it will do the right thing. | 04:34:08 |
| thefloweringash | * As for -sdk_version, that’s carefully handled by the logic added in “Darwin platform versions” PR in the ld wrapper. It’s valid to pass an explicit -sdk_version to the wrapper, and it will do the right thing. (https://github.com/NixOS/nixpkgs/pull/111988) | 04:36:50 |
| toonn | LnL: Is there a fix when a framework is supposed to have a header but doesn't? | 19:28:44 |
| toonn | I don't think we aggregate headers from SDKs the way we can from source releases versions? | 19:29:08 |
| toonn | Hmm, something must be going on. I checked the .pkg and that header *is* in there. So why isn't it in darwin.apple_sdk 🤔 | 19:44:19 |
| LnL | hmm for frameworks that's unexpected, as far as I remember the headers for those come straight out of the apple package | 20:37:42 |
| toonn | Can't say I was expecting it : ) | 20:59:39 |
| toonn | I'll look into it further day after tomorrow. | 20:59:52 |
| 19 Jun 2021 |
|  nazarii left the room. | 20:16:44 |
| 20 Jun 2021 |
 siraben | Anyone using gccemacs on macOS with a binary cahce? | 15:54:27 |
 mjlbach | Me | 15:55:29 |
| siraben | mjlbach: do the instructions in https://gist.github.com/mjlbach/179cf58e1b6f5afcb9a99d4aaf54f549 still work? | 15:56:11 |
