| 25 May 2021 |
 LnL | oh, CF is just gone from the list for newer releases | 16:47:10 |
| LnL | https://opensource.apple.com/release/macos-10126.html | 16:47:25 |
| LnL | That CF-1349.80 (coming soon!) number could be found as a release number back then IIRC | 16:48:16 |
| LnL | our current build returns 1454.90.0 while eg. Catalina has 1677.201.0 but we also just write a random version based on guessing | 17:00:39 |
 toonn | Ah, I meant version of swift-corelibs though, like swift-5.4-RELEASE. | 17:07:55 |
| toonn | As in, I'm looking into updating swift-corelibs because of an issue and I'm not sure whether I need to try to get the OSX-10.13-ish version of swift-corelibs. | 17:09:28 |
| toonn | This issue in particular, https://git.io/JGejW | 17:10:00 |
| LnL | yeah, see https://github.com/NixOS/nixpkgs/blob/684963881b19d2c93adf4423d293357a89b9f253/pkgs/os-specific/darwin/swift-corelibs/corefoundation.nix#L64-L66 | 17:11:14 |
| LnL | so our best bet I guess is to find the pr/commits like https://github.com/apple/swift-corelibs-foundation/pull/2782 and pick the first stable release that includes those changes | 17:13:38 |
| toonn | LnL: The current version being your guess for High Sierra kinda obviates my need to update it. | 17:23:05 |
| toonn | I was hoping the issue I'm seeing was due to a 10.12 CF with 10.13 XNU. But if the CF is already the one for 10.13 that seems unlikely. | 17:23:45 |
| toonn | I'm kinda stuck on the problem. This is pretty much the only relevant search result, https://trac.macports.org/ticket/54993 | 17:24:22 |
| toonn | And that was a problem specifically with the order of the attributes and GCC. | 17:24:37 |
| toonn | Our CF already has the attribute order fix. And darwin stdenv isn't built with GCC anyway. | 17:24:58 |
| LnL | was about to say, do you mean clang | 17:26:25 |
| toonn | No, just saying a GCC-related bug that has already been fixed is the only relevant thing I've come across : ) | 17:30:14 |
| toonn | A veritable dearth of information if I've ever seen one. | 17:30:27 |
| LnL | can you post the actual traceback somewhere | 17:35:59 |
| toonn | LnL: Added nix-store --read-log to the gist. | 17:43:51 |
| LnL | did any of that change compared to the current Libsystem? | 17:49:33 |
| LnL | hmm wait, Libsystem-1238.60.2 is the current version | 17:50:22 |
| toonn | Yeah, haven't updated Libsystem yet. Only XNU. | 18:03:41 |
| LnL | hmm right, the actual source of the header is xnu itself | 18:09:54 |
| LnL | there is a diff between the current version and 10.13 however | 18:11:17 |
| LnL | the /nix/store/jhcpg6l57syfnyqbk3bc1zg1vjwksshx-Libsystem-1238.60.2 is headers from bootstrap tools right? | 18:12:36 |
| LnL | * the /nix/store/jhcpg6l57syfnyqbk3bc1zg1vjwksshx-Libsystem-1238.60.2 is using headers from bootstrap tools right? | 18:12:56 |
| toonn | I think so, yes. | 18:15:45 |
| toonn | You think updating the Libsystem in bootstrap-tools is the fix? | 18:17:33 |
| LnL | not sure, but trying a build with a hacked up Libsystem which has the new include/sys/syslog.h is definitively something to try | 18:19:54 |
| LnL | if you drill down to the correct stdenv.__bootPackages... attribute you can .override it | 18:21:42 |
