| 7 Mar 2026 |
 dramforever | so it's at least fragile somehow? | 11:54:18 |
 Yureka (she/her) | https://github.com/openMathLib/openBLAS/commit/05fc88180c42eae7588fac4bd59c3b86ffefbc23 | 12:10:18 |
| Yureka (she/her) | reverting this fixes the build on both Ampere and M1 Pro | 12:10:33 |
| Yureka (she/her) | https://github.com/NixOS/nixpkgs/pull/497538 | 12:15:41 |
 K900 | Can't you just makeflags BUILD_BFLOAT16=0? | 12:21:56 |
| dramforever | oh come on
postPatch = ''
# cc1: error: invalid feature modifier 'sve2' in '-march=armv8.5-a+sve+sve2+bf16'
substituteInPlace Makefile.arm64 --replace-fail "+sve2+bf16" ""
'';
| 12:23:04 |
| dramforever | come on | 12:23:07 |
| dramforever | turns out we played ourselves | 12:23:53 |
| dramforever | commented on the pr | 12:28:34 |
| Yureka (she/her) | wait what? | 12:36:24 |
| Yureka (she/her) | wouldn't replace-fail fail if it didn't replace anything? | 12:36:34 |
| Yureka (she/her) | or does it only replace the first occurance | 12:38:36 |
| Yureka (she/her) | No | 12:41:02 |
| Yureka (she/her) | Oh I understand now | 12:42:26 |
| Yureka (she/her) | We shouldn't be replacing/removing this flag anymore | 12:42:37 |
| Yureka (she/her) | And when keeping it, the new version builds | 12:42:46 |
| dramforever | now the question is how did this ever build on aarch64-linux | 12:48:07 |
| Yureka (she/her) | maybe it didn't and someone misclicked? | 13:09:59 |
 ghpzin | it doesn't fix opencv thing right ? | 13:16:09 |
| dramforever | what's the opencv thing | 13:16:58 |
| Yureka (she/her) | https://github.com/NixOS/nixpkgs/issues/495748
https://github.com/NixOS/nixpkgs/issues/495726
are now also happening on staging-next (I previously reported them on staging) | 13:22:37 |
| dramforever | get these googlers IWYU | 13:30:48 |
| Yureka (she/her) | These ones are new: protobufc, onnxruntime
https://github.com/NixOS/nixpkgs/issues/497555
https://github.com/NixOS/nixpkgs/issues/497556 | 13:33:15 |
| ghpzin | There is some separate openblas build pulled by opencv that is broken on x86_64-linux:
...
/nix/store/rfp8lhk4dl9syfn64rwb3h3c73426p08-binutils-2.44/bin/ld: /build/ccz2kZvK.o: in function `main':
linktest.c:(.text.startup+0x3142): undefined reference to `openblas_set_threads_callback_function'
collect2: error: ld returned 1 exit status
make[1]: *** [Makefile:211: ../libopenblas-r0.3.31.so] Error 1
make[1]: Leaving directory '/build/source/exports'
make: *** [Makefile:149: shared] Error 2
but I guess it is a different issue from aarch64-linux one. | 13:35:11 |
| dramforever | different issue | 13:35:44 |
| ghpzin | There is some separate openblas build pulled by opencv that is broken on x86_64-linux:
...
/nix/store/rfp8lhk4dl9syfn64rwb3h3c73426p08-binutils-2.44/bin/ld: /build/ccz2kZvK.o: in function `main':
linktest.c:(.text.startup+0x3142): undefined reference to `openblas_set_threads_callback_function'
collect2: error: ld returned 1 exit status
make[1]: *** [Makefile:211: ../libopenblas-r0.3.31.so] Error 1
make[1]: Leaving directory '/build/source/exports'
make: *** [Makefile:149: shared] Error 2
but I guess it is a different issue from aarch64-linux one.
---
Upstream issue: https://github.com/OpenMathLib/OpenBLAS/issues/5607
Probably need this: https://github.com/OpenMathLib/OpenBLAS/commit/14594773a0c06b1fb40208dcdbd3aad4c39d845d | 13:51:01 |
 matthewcroughan | Any objections to merging https://github.com/NixOS/nixpkgs/pull/497358 | 16:09:15 |
| matthewcroughan | Seems like it has been done before, and merged then, and seems to be a recurring theme | 16:09:25 |
| matthewcroughan | Maybe we should never ever use the "latest" in this derivation, and always pin it | 16:10:28 |
| matthewcroughan | the fact that it was ever allowed to be pointing to latest, seems to be the issue, no? | 16:10:40 |
