Releases: Goddard-Fortran-Ecosystem/GFE
Releases · Goddard-Fortran-Ecosystem/GFE
Various updates to submodules and CI compilers
Changed
- Update gFTL to v1.14.0
- Added support for non-poly allocatable containers
- Update gFTL-shared to v1.9.0
- Update fArgParse to v1.8.0
- Update yaFyaml to v1.4.0
- Workaround for gfortran 13.3 and minor updates to CI
- Update pFUinit to v4.10.0
- Migrated to use v2 interfaces from gFTL
- Update CI to match that of pFUnit
- Remove
macos-11
from GitHub Actions, add macos-12
and gfortran-14
to Ubuntu 24.04
- Update CI to NVHPC 24.5
Update pFlogger submodule to workaround issues in gfortran 13.2
Changed
- Update pFlogger to v1.14.0
- Added
-quiet
flag for NAG Fortran
- Workaround additional polymorphic assignment bug in gfortran 13.2 (in build_locks)
[1.14.0] - 2024-03-20
Changed
- Update gFTL to v1.13.0
- Update pFlogger to v1.13.2
- Removed
macos-11
from CI as it is deprecated. Add macos-13
Port to ifx (and minor bugfix for unit tests)
Changed
- Update gFTL to v1.13.0
- Update pFlogger to v1.13.2
- Removed
macos-11
from CI as it is deprecated. Add macos-13
Port to Fujitsu and various workarounds for GCC 13
Changed
- Update gFTL to v1.12.0
- Fujitsu compiler support (and other minor bugfixes)
- Update gFTL-shared to v1.8.0
- Update fArgParse to v1.7.0
- Fujitsu compiler support
- Update to Apache 2.0 license
- Update pFUnit to v4.9.0
- Fujitsu compiler support
- Some bug fixes
- Update yaFyaml to v1.3.0
- Fujitsu compiler support
- GCC 13 fixes
- Update pFlogger to v1.13.1
- Fujitsu compiler support
- GCC 13 fixes
- Update CI to use Intel LLVM
Updated submodules (mostly Intel LLVM port)
Package |
Version |
gFTL |
v1.11.0 |
gFTL-shared |
v1.7.0 |
fArgParse |
v1.6.0 |
pFUnit |
v4.8.0 |
yaFyaml |
v1.2.0 |
pFlogger |
v1.11.0 |
bugfix updates to gFTL-shared and pFUnit
Current versions
Package |
Version |
gFTL |
v1.10.0 |
gFTL-shared |
v1.6.1 |
fArgParse |
v1.5.0 |
pFUnit |
v4.7.3 |
yaFyaml |
v1.1.0 |
pFlogger |
v1.10.0 |
Changed
- Update gFTL-shared to v1.6.1
- Fixed issue where the names of some iterators for containers were not being correctly named. E.g., StringSetIterator was only named SetIterator.
- Update pFUnit to v4.7.3
- Fixed cmake issue where target "pfunit-mpi-defines" is defined more than once
- Missing variable declaration in parameterized test case boiler plate code.
Updated various submodules
Current versions
Package |
Version |
gFTL |
v1.10.0 |
gFTL-shared |
v1.6.0 |
fArgParse |
v1.5.0 |
pFUnit |
v4.7.0 |
yaFyaml |
v1.1.0 |
pFlogger |
v1.10.0 |
Changed
- Update gFTL to v1.10.0
- Added
IntelLLVM.cmake
file as a copy of Intel.cmake
to support the LLVM Intel compiler frontends
(Note - this has not been tested due to lack of system access.)
- Added Fortran-friendly iterator factories (and tests)
- Update gFTL-shared to v1.6.0
- Added
IntelLLVM.cmake
file as a copy of Intel.cmake
to support the LLVM Intel compiler frontends
- Updated gFTL submodule to v1.10.0
- Update fArgParse to v1.5.0
- Added
IntelLLVM.cmake
file as a copy of Intel.cmake
to support the LLVM Intel compiler frontends
- Updated submodules for gFTL-shared (v1.6.0)
- implemented workaround for GFortran which was not correctly handling
deferred-length allocatable string arrays. Used StringVector instead.
- Update pFUnit to v4.7.0
- Update fArgParse submodule to v1.5.0
- Added IntelLLVM.cmake to support ifx
- Added interface for
@asertEquals
for arrays of strings. Previously only string scalars could be compared.
- Added check in pFUnit preprocessor that raises an exception if the module name and filename do not agree unless
@suite
is used to override default assumptions.
- Added option to set labels to ctests
- Added changelog enforcer GitHub Action
--verbose option
is now passed through by ctest runner
- Converted GitHub CI to use cmake abstract build commands
- Update yaFyaml to v1.1.0
- Added
IntelLLVM.cmake
file as a copy of Intel.cmake
to support the LLVM Intel compiler frontends
- Updated required gFTL version to v1.10.0
- Updated required gFTL-shared version to v1.6.0
- Update pFlogger to v1.10.0
- Added
IntelLLVM.cmake
file as a copy of Intel.cmake
to support the LLVM Intel compiler frontends
- Updated required version of gFTL to v1.10.0
- Updated required version of gFTL-shared to v1.6.0
- Updated required version of yaFyaml to v1.1.0
Added
Important compiler workarounds and code fixes for NVIDIA
Current versions
Package |
Version |
gFTL |
v1.8.3 |
gFTL-shared |
v1.5.1 |
fArgParse |
v1.4.2 |
pFUnit |
v4.6.3 |
yaFyaml |
v1.0.7 |
pFlogger |
v1.9.3 |
Fixed
- Update gFTL to v1.8.3
- Fixes for GNU Make builds
- Missing
KIND=
on size()
procedure. Gave incorrect response for large containers. Only affects V1 containers.
- Update gFTL-shared to v1.5.1
- Fixes for GNU Make builds
- Update fArgParse to v1.4.2
- Fixes for GNU Make builds
- Update pFUnit to v4.6.3
- Fix for compilers that do not support 128 bit reals
- Fixed build_submodule for old git versions
- Fixed for use with FetchContent
- Fix CMake logic in
add_pfunit_ctest.cmake
for MPIEXEC_EXECUTABLE
. Problem not exposed by common MPI flavors which use mpirun
- Fix GitHub CI workflow by pinning to CMake 3.24.3
- Fixes for GNU Make builds
- Update fArgParse submodule to v1.4.2
- Fix
pFUnitParser.py
in cases where there is no module name
- Update yaFyaml to v1.0.7
- Workaround for NVIDIA compiler
- Fixes for GNU Make builds
- Workarounds for Intel 2021.7
- Out of date examples have been updated
- Update pFlogger to v1.9.3
- Fixes for GNU Make builds
- Package was incorrectly assuming that all extant compilers support 128 bit reals. Now a check is performed to optionally include support for 128 bit reals.
- Fix
Parser()
declaration for Intel 2021.7
Improved default install path (for most users)
Changed
- Change behavior if user does not pass in
CMAKE_INSTALL_PREFIX
to put install directory inside build (as we are know we can write there; default for CMake install is /usr/local
)
Added
Update pFUnit submodule (bugfix)
Changed
- Update pFUnit to v4.6.1 (bug fix)