| 26 Jun 2026 |
 emily | so hopefully we won't need the fixup-gcc-15-darwin-aarch64-support.patch | 07:31:58 |
| emily | btw if you care about GNAT on Darwin you might want to try building it on that branch :) | 07:32:25 |
| emily | I am pretty sure the patches should no longer matter at all but good to double-check | 07:32:56 |
| emily | the problem is that it was rebased from a years-old branch but not updated for all the changes in GCC since then | 07:34:01 |
| emily | so it has been behind the GCC packaging we use since it landed | 07:34:13 |
| emily | and AIUI it also has no working Linux stdenv bootstrap story | 07:34:50 |
| emily | so if we were to switch to it, it would need someone to go through years of GCC changes or compare it to the current GCC packaging with a fine-toothed comb and also write a stdenv bootstrap | 07:35:17 |
| emily | and I don't think there's been progress on that since it was merged a year ago, so I am worried it is just going to bitrot | 07:35:56 |
| emily | hm, I guess I should add the Darwin patches for GCC 16 to the main gcc while I'm at it :) | 07:36:43 |
| emily | Myria: also, wait, are you saying stdenv builds but stdenvNoCC fails?! | 07:40:33 |
 Myria | Hmmm, stdenvNoCC succeeded before I rebased, now it fails | 07:44:17 |
| Myria | Will need to investigate more later. My thoughts are it's a problem with the provided bootstrap-tools? But idk | 07:45:54 |
| Myria | Will take a look, thanks for all that! | 07:46:13 |
 Grimmauld (any/all) | disallowed requisites on gcc can do that | 07:46:26 |
| Myria | I updated them in my rebase so it properly builds, avoiding my janky fix | 07:47:15 |
| emily | oh, I meant on Darwin | 07:47:17 |
| emily | but I guess that's a separate issue | 07:47:20 |
| Myria | Yeah, stdenv is also failing now rebased on aarch64-darwin | 07:48:02 |
| emily | it looks like 15.3.0 is probably just leaking more into outputs | 07:49:00 |
| emily | do you have the error? | 07:49:28 |
| emily | perhaps we should just revert the bump (on staging too) until there is time to look at this properly | 07:50:46 |
| Myria | https://pastebin.com/LPiBEQXq | 07:50:52 |
| Myria | Probably ideal | 07:51:14 |
| Myria | Seems like it's the same error as before | 07:51:33 |
| Grimmauld (any/all) | yeah the non-rebased version seems to succeed, the rebased version fails | 08:11:45 |
| Grimmauld (any/all) | weird | 08:11:53 |
| emily | are we 100% positive it's the GCC bump that did it? | 08:47:15 |
| emily | it may be the combination of the bump with another change in the worst case | 08:47:27 |
| emily | oh you said "rebased on aarch64-darwin" :) | 08:47:39 |
| Myria | Oh shoot, I meant aarch64-linux sorry | 08:48:16 |
