| 25 Jun 2026 |
 Grimmauld (any/all) | can confirm, and not an issue on x86_64-linux | 18:12:19 |
 Vladimír Čunát | staging-next merged! | 19:03:57 |
 Gaétan Lepage | That was the small one, best on 26.05's staging, right? | 20:12:19 |
| Gaétan Lepage | * That was the small one, based on 26.05's staging, right? | 20:12:32 |
 hexa | yes | 20:17:21 |
| 26 Jun 2026 |
 emily | working on this now | 03:43:22 |
| emily | plan still for 26.11 imminently? | 03:43:31 |
| Vladimír Čunát | I guess so. I'm not aware of any relevant news. | 04:21:09 |
| Grimmauld (any/all) | In reply to @vcunat:matrix.org
I guess so. I'm not aware of any relevant news. May want to look into why stdenvNoCC fails on aarch64-linux on staging branch before starting | 05:48:24 |
| Grimmauld (any/all) | Or well, we know why (GCC 15.3.0) but I don't think we have a fix | 05:49:03 |
| Vladimír Čunát | You know which particular change inside 15.2 -> 15.3? | 05:52:19 |
| Grimmauld (any/all) | In reply to @whispers:catgirl.cloud
hm. i don't know where to ask about this/search for information (especially since it might be my fault, computers have been acting up), it may or may not be known already, but is stdenvNoCC broken on staging for aarch64-linux? am on commit 03357d87710d8b303d13d8889ff7a7a428e338ea. error i get is
error: output '/nix/store/057dp7l3976rlnj4rfskssn0s2jfh5d1-stdenv-linux-no-cc' is not allowed to refer to the following paths:
/nix/store/1bdzriipn18184zpwh2lfdabk2gqcgxn-glibc-2.42-61
/nix/store/f04zlspmsqzyv15kcxbn6mx4wb71a0jm-libidn2-2.3.8
/nix/store/hihrq5y6ls54k8p3r7d1xal0np2k8a54-libtool-2.5.4-lib
/nix/store/xag03kgklavhhq66qnac59pqxpxvjqga-libunistring-1.4.2
/nix/store/xzs7w65n77a94hmpwkfsnwqkj2k5akn4-xgcc-15.2.0-libgcc
full text: https://paste.debian.net/hidden/d3d8863c. i don't know enough about stdenv to know where to begin looking at this Only repeating what whispers found out. I didn't do the bisect, I just confirmed the failure | 05:54:31 |
| Vladimír Čunát | One option is to just revert the gcc bump for now, until this is fixed 🤷 | 05:58:30 |
| Vladimír Čunát | That's what I plan to do when starting staging-next. | 05:59:13 |
| emily | https://github.com/NixOS/nixpkgs/pull/535508 should cover everything that has enough rebuilds to want to land before staging starts; I'll build stdenv on that now (cc Randy Eckenrode too). https://github.com/NixOS/nixpkgs/pull/535513 fixes a regression caused by the drop. cleaning up my big branch of fixing packages that will have their update scripts break from the drop, but that shouldn't cause many rebuilds | 07:05:59 |
 Myria | I have a fix for the aarch64-darwin stuff with GCC 15.3.0, not sure about the linux stuff? | 07:17:06 |
| Grimmauld (any/all) | if you have a PR i can poke a build on aarch64-linux and see if stdenvNoCC happens to be fixed by it "accidentally" (worth a shot, even if unlikely) | 07:20:51 |
| Myria | It would just be my gcc update PR: https://github.com/NixOS/nixpkgs/pull/531338 | 07:22:02 |
| Grimmauld (any/all) | Didn't we already do this: https://github.com/NixOS/nixpkgs/pull/532074 ? | 07:22:26 |
| Myria | My main thoughts would just be the minimal-bootstrap update will change hopefully the paths required | 07:22:33 |
| Myria | Will need to rebase it now though | 07:22:52 |
| Grimmauld (any/all) | poked a build | 07:24:30 |
| emily | why don't we fix it by bumping to GCC 16 | 07:24:49 |
| Grimmauld (any/all) | I am confused why a later PR doing the same thing (but less thorough) was merged with barely any review within hours of it being opened... | 07:25:19 |
| Myria | Yeah, me too, but that's how it goes I guess | 07:25:50 |
| emily | we probably need to figure out what's happening with gccNGPackages at some point, since it's been years behind the GCC derivation for some things since it landed and afaik has no active maintenance / hasn't picked up many of the changes to GCC packaging since… | 07:25:59 |
| emily | could you rebase this on staging given the other one was merged? or I guess we could revert that one and re-apply this | 07:26:47 |
| emily | actually since there's no conflicts I'll just prod it in GitHub | 07:27:16 |
| emily | though I suppose the commit messages are now a bit confusing if both get landed | 07:27:27 |
| emily | (e.g. if bisection hits one) | 07:27:30 |
