Gleb Popov
Gleb Popov
Hum, maybe. We're now also seeing GHC failures that look similar: https://pkg-status.freebsd.org/package19/data/142amd64-default-build-as-user/311324bc7461/logs/ghc-9.8.4.log
What needs a confirmation, exactly? It is a regression from 3.14. The bug itself may be somewhere inside GHC, but the fact is that the problem wasn't seen until cabal-install...
I agree, the hooks feels like a more proper place for this feature, but from what I gather, even the "jail start" hook is ran after the jail is actually...
How do we move this forward?
Bump. Can we get this in?
Thank you. I only rebased this PR to catch up with upstream changes and solve conflicts. This is now a very low priority for me, so maybe you should remove...
pkg prints an upgrading plan for you and asks you're ok with it. This was supposedly caused by a partial upgrade, but without logs nothing can be done with this...
I tried to compile the 0.7.0 release with Flang 21. Here are some errors that I got: ``` error: Semantic errors in /wrkdirs/usr/ports/devel/fortran-stdlib/work/.build/src/stdlib_specialfunctions_gamma.f90 /wrkdirs/usr/ports/devel/fortran-stdlib/work/.build/src/stdlib_specialfunctions_gamma.f90:16:5: error: REAL(KIND=-1) is not a supported...
Right, here's the explanation from upstream: https://github.com/llvm/llvm-project/issues/158444#issuecomment-3289580940 I will retry with quadmath enabled and report back.
If you have a reproducer, maybe you can compile pkg WITH_DEBUG=yes from ports to get a proper backtrace?