It looks like GCC 16.1.0 is now released. Yay!
Tasks to do:
- Build on bucy (Linux)
- Build and test on macOS
The first should be "simple" in the sense that gcc + Linux is usually well tested at least for build. My guess is any issues will be at the Open MPI or Baselibs level.
The second? 🤷🏼 I see an auto-PR into homebrew-core (Homebrew/homebrew-core#280204) but it apparently failed to build on Tahoe (see https://github.com/Homebrew/homebrew-core/actions/runs/25163315545/job/73763626598?pr=280204). Well, it seemed to have failed to build everywhere even Linux. My guess is @fxcoudert or @iains might need to do some brew and macOS magic as they have in the past (see #290 for 15.1 fun).
Bucy
Open MPI
Open MPI has issues (see #376) but it builds with some hacks.
Baselibs
All of Baselibs that I care about built! Some issues with SDPToolkit, but eh, I don't care.
GFE
One pFUnit test fails:
Goddard-Fortran-Ecosystem/pFUnit#548
Claude helped figure out a workaround:
Goddard-Fortran-Ecosystem/pFUnit#549
and I've made a bugzilla report:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125117
MAPL2
100% tests passed, 0 tests failed out of 56
No CMake changes needed!
MAPL3
100% tests passed, 0 tests failed out of 61
Yay.
GEOSgcm
It looks like GCC 16.1.0 is now released. Yay!
Tasks to do:
The first should be "simple" in the sense that gcc + Linux is usually well tested at least for build. My guess is any issues will be at the Open MPI or Baselibs level.
The second? 🤷🏼 I see an auto-PR into homebrew-core (Homebrew/homebrew-core#280204) but it apparently failed to build on Tahoe (see https://github.com/Homebrew/homebrew-core/actions/runs/25163315545/job/73763626598?pr=280204). Well, it seemed to have failed to build everywhere even Linux. My guess is @fxcoudert or @iains might need to do some brew and macOS magic as they have in the past (see #290 for 15.1 fun).
Bucy
Open MPI
Open MPI has issues (see #376) but it builds with some hacks.
Baselibs
All of Baselibs that I care about built! Some issues with SDPToolkit, but eh, I don't care.
GFE
One pFUnit test fails:
Goddard-Fortran-Ecosystem/pFUnit#548
Claude helped figure out a workaround:
Goddard-Fortran-Ecosystem/pFUnit#549
and I've made a bugzilla report:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125117
MAPL2
No CMake changes needed!
MAPL3
Yay.
GEOSgcm