It was using the global parallel value in cmake integration tests, while
on macos runners, this should be lower by one, as used in other macos
jobs. Performance impact is minimal.
Follow-up to
fb70812437ad28b74dbdc1031e46c1d86bc9db3c #16126
Closes #18701
shell: ${{ contains(matrix.image, 'windows') && 'msys2 {0}' || 'bash' }}
env:
CC: ${{ !contains(matrix.image, 'windows') && 'clang' || '' }}
+ MAKEFLAGS: ${{ contains(matrix.image, 'macos') && '-j 4' || '-j 5' }}
MATRIX_IMAGE: '${{ matrix.image }}'
TESTOPTS: ${{ contains(matrix.image, 'macos') && '-D_CURL_PREFILL=ON' || '' }} ${{ contains(matrix.image, 'windows') && '-DCMAKE_UNITY_BUILD_BATCH_SIZE=30' || '' }}
OLD_CMAKE_VERSION: 3.11.4