| 6 Dec 2025 |
John Ericson | It reeks of "configuration not composition" | 16:34:47 |
jade_ | raitobezarius: fyi there's internal server errors on zulip.lix.systems when trying to log in | 19:52:38 |
raitobezarius | fixed, that's due to a keycloak breaking change | 19:53:10 |
jade_ | ty | 19:53:46 |
raitobezarius | https://cl.afnix.fr/c/infra/+/175 | 19:53:54 |
Winter | raitobezarius are you aware of similar issues when logging into forgejo? this has been happening for months though but just curious if you’re aware | 20:03:49 |
jade_ | is this 100% reproducible or inconsistent? | 20:04:15 |
raitobezarius | not-100% reproducible | 20:04:21 |
raitobezarius | those issues afaik are related to the IPv6 only stack | 20:04:27 |
raitobezarius | they were fixed once we ironed out all the DNS resolution mess | 20:04:36 |
raitobezarius | (just retried again right now to check and it works) | 20:04:44 |
raitobezarius | but we did have such a mess | 20:04:47 |
| 7 Dec 2025 |
thubrecht | Welp https://github.com/NixOS/nixpkgs/commit/b3cf9ce0f9d01b9c1dc4b7e04fb8b28f79ee68c4 | 09:21:41 |
raitobezarius | sounds like it's worth sending a revert | 14:13:32 |
Rutile (Commentator2.0) feel free to ping | what context am i missing? | 19:11:45 |
Sergei Zimmerman (xokdvium) | In reply to @commentator2.0:elia.garden what context am i missing? The env variable is CppNix specific and is kind of a kludge | 19:12:21 |
Sergei Zimmerman (xokdvium) | Also builtins.warn hasn’t landed yet I think because of the whole env variable? | 19:12:55 |
Sergei Zimmerman (xokdvium) | Anyway minver is 2.18 and the abort on warm behavior can’t be relied on to be supported if it’s polyfilled | 19:13:45 |
Sergei Zimmerman (xokdvium) | * Anyway minver is 2.18 and the abort on warn behavior can’t be relied on to be supported if it’s polyfilled | 19:13:58 |
Sergei Zimmerman (xokdvium) | Just grepping for “warn” would have been a better approach. Or do a scopedImport to make builtins.warn into a throw. | 19:16:20 |
thubrecht | The issue at play I think is that lix now has builtins.warn but it righfully has no knowledge of the env variable, hence the lib polyfill does not execute and no throws happen | 19:54:04 |
thubrecht | What is weird though is that there are similar checks that are done before, but they still pass... | 19:55:25 |
Sergei Zimmerman (xokdvium) | In reply to @thubrecht:matrix.org What is weird though is that there are similar checks that are done before, but they still pass... Exit code is not checked? | 20:15:37 |
thubrecht | It is checked | 20:53:56 |
thubrecht | But the other tests that use NIX_ABORT_ON_WARN don't actually fail because the warning in considered an error | 20:54:24 |
thubrecht | But because of error: cannot convert a thunk to JSON | 20:54:31 |
Sergei Zimmerman (xokdvium) | Ugh, wth | 20:54:57 |
Sergei Zimmerman (xokdvium) | That error doesn't make sense, nothing in user code should be able to observe thunks | 20:55:50 |
thubrecht | It's evaluation an option | 20:57:29 |
thubrecht | * It's evaluating an option | 20:57:34 |