| 16 Nov 2025 |
Grimmauld (any/all) | do i just add that? What is our stance on silencing warnings like that? | 12:28:20 |
Grimmauld (any/all) | i can't find any upstream issues about it either | 12:29:05 |
symphorien | the opinion of the nixpkgs core team is linked in this post https://discourse.nixos.org/t/nixpkgs-core-team-update-2025-11-04/71742 | 12:58:53 |
| 18 Nov 2025 |
| grimmauld (any/all) joined the room. | 08:24:55 |
| jopejoe1 (4094@39c3) changed their display name from jopejoe1 to jopejoe1 (4094@39c3). | 20:18:56 |
| 19 Nov 2025 |
Grimmauld (any/all) | nix why-depends /nix/store/64324wnvzc6308591l5ckwmwyjbx4ay6-krb5-x86_64-unknown-linux-musl-1.22.1-lib /nix/store/2fwx03hyi4azqfwa0ayx6y6affy06rxr-x86_64-unknown-linux-musl-clang-wrapper-21.1.2 --precise --all
/nix/store/64324wnvzc6308591l5ckwmwyjbx4ay6-krb5-x86_64-unknown-linux-musl-1.22.1-lib
└───lib/krb5/plugins/preauth/pkinit.so: …inux-musl-21.1.2/lib:/nix/store/7pcr012rp5r4fiyz6mnpjkssx08wac5v-openssl-x86_64-unknown-linux-mu…
lib/krb5/plugins/preauth/spake.so: …inux-musl-21.1.2/lib:/nix/store/7pcr012rp5r4fiyz6mnpjkssx08wac5v-openssl-x86_64-unknown-linux-mu…
lib/krb5/plugins/tls/k5tls.so: …inux-musl-21.1.2/lib:/nix/store/7pcr012rp5r4fiyz6mnpjkssx08wac5v-openssl-x86_64-unknown-linux-mu…
→ /nix/store/7pcr012rp5r4fiyz6mnpjkssx08wac5v-openssl-x86_64-unknown-linux-musl-3.6.0
└───lib/libcrypto.so.3: …~#.o~#.....compiler: /nix/store/2fwx03hyi4azqfwa0ayx6y6affy06rxr-x86_64-unknown-linux-musl-clang…
→ /nix/store/2fwx03hyi4azqfwa0ayx6y6affy06rxr-x86_64-unknown-linux-musl-clang-wrapper-21.1.2
Uhh, i am seeing build clang leaking into host musl, this is not intended, right? | 16:05:06 |
Grimmauld (any/all) | (on llvm musl, of course) | 16:05:26 |
K900 | Via openssl? | 16:07:13 |
K900 | Probably needs to be patched out | 16:07:17 |
Grimmauld (any/all) | yeah, downstream things break, like libpq having disallowed requisites on stdenv.cc | 16:10:21 |
Grimmauld (any/all) | thats how i noticed | 16:10:33 |
Grimmauld (any/all) | wait i can't read, this is krb5 being wonky, not musl itself | 16:12:06 |
Grimmauld (any/all) | but yes, via openssl | 16:13:34 |
Grimmauld (any/all) | compiler: /nix/store/2fwx03hyi4azqfwa0ayx6y6affy06rxr-x86_64-unknown-linux-musl-clang-wrapper-21.1.2/bin/x86_64-unknown-linux-musl-clang -fPIC -pthread -m64 -Wa,--noexecstack -Qunused-arguments -Wall -O3 -DOPENSSL_USE_NODELETE -DL_ENDIAN -DOPENSSL_PIC -DOPENSSL_BUILDING_OPENSSL the entire compiler CLI is in there... Wild. | 16:22:00 |
Grimmauld (any/all) | Found the culprit! | 16:27:55 |
Grimmauld (any/all) |  Download image.png | 16:27:55 |
K900 | Oof | 16:28:16 |
Grimmauld (any/all) | openssl throws the compiler info into the build | 16:28:18 |
K900 | Should probably just patch it | 16:28:21 |
K900 | To strip it out always | 16:28:23 |
Grimmauld (any/all) | yep | 16:28:27 |
K900 |  Download image.png | 16:29:30 |
K900 | https://github.com/openssl/openssl/blob/8da3d5dd9f7058335794190a11bf1b764536c716/util/mkbuildinf.pl#L37 | 16:29:36 |
K900 | sob | 16:29:36 |
Grimmauld (any/all) | we could probably just remove-references-to -t ${stdenv.cc} buildinf.h in postConfigure | 16:32:35 |
Grimmauld (any/all) | i feel like that'd be the easiest solution | 16:32:59 |
K900 | Honestly maybe just delete the entire line | 16:33:18 |
Grimmauld (any/all) | nah that'll cause issues | 16:33:26 |
K900 | Or maybe there's some way to just not write it in the first place | 16:33:29 |
Grimmauld (any/all) | https://github.com/openssl/openssl/blob/8da3d5dd9f7058335794190a11bf1b764536c716/crypto/build.info#L118 we could remove $(CC) from ehre | 16:34:05 |