Rebuild for compiler_version 9

[h-vetinari]

Needed to resolve failing dependency resolution e.g. in conda-forge/scipy-feedstock#153

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[h-vetinari]

I hadn't seen #19 (which however didn't bump the linux compiler version); now opened conda-forge/blas-feedstock#59 as well

[h-vetinari]

@isuruf, this is green. :)

[h-vetinari]

@isuruf, any objections to merging this?

[isuruf]

This wouldn't change anything as we don't use a fortran compiler here.

[h-vetinari]

This is about the c_compiler_version, and -- AFAICT -- the reason why things don't resolve anymore on linux for e.g. scipy. Not least because the resolution errors started on the compilers migration, and were clearly visible in the sense that, in the same CI run, v7 resolved and v9 didn't. These resolution errors in conda-forge/scipy-feedstock#130 & conda-forge/scipy-feedstock#153 haven't gone away since.

[isuruf]

The only thing that would change with this PR is that the run dependency of libgcc-ng>=7.5.0 will become libgcc-ng>=9.3.0. I don't see how that will help anything.

[h-vetinari]

Have a look at some of the resolution errors for adding blis to the scipy run-reqs... I'm having a hard time coming up with a definite culprit, but libgcc-ng is part of the game there:

Package libgcc-ng conflicts for:
numpy[version='>=1.16.6,<2.0a0'] -> python[version='>=3.7,<3.8.0a0'] -> libgcc-ng[version='>=4.9|>=7.2.0']
mpmath -> python -> libgcc-ng[version='>=4.9|>=7.3.0|>=7.5.0|>=9.3.0|>=7.2.0']
pytest -> python[version='>=3.9,<3.10.0a0'] -> libgcc-ng[version='>=4.9|>=7.3.0|>=7.5.0|>=9.3.0|>=7.2.0']
numpy[version='>=1.16.6,<2.0a0'] -> libgcc-ng[version='>=7.3.0|>=7.5.0|>=9.3.0']
libgcc-ng[version='>=9.3.0']
python[version='>=3.7,<3.8.0a0'] -> libgcc-ng[version='>=4.9|>=7.3.0|>=7.5.0|>=9.3.0|>=7.2.0']
blas=[build=*blis] -> liblapack==3.8.0[build=*netlib] -> libgcc-ng[version='>=9.3.0']
libcblas[version='>=3.8.0,<4.0a0'] -> libgcc-ng[version='>=7.3.0|>=7.5.0|>=9.3.0']
liblapack[version='>=3.8.0,<4.0a0'] -> libgcc-ng[version='>=7.3.0|>=7.5.0|>=9.3.0']
python_abi=3.7[build=*_cp37m] -> python=3.7 -> libgcc-ng[version='>=4.9|>=7.3.0|>=7.5.0|>=9.3.0|>=7.2.0']
blas=[build=*blis] -> libgcc-ng[version='>=7.3.0|>=7.5.0']
libblas[version='>=3.8.0,<4.0a0'] -> libgcc-ng[version='>=7.3.0|>=7.5.0|>=9.3.0']
pytest-xdist -> python[version='>=3.5'] -> libgcc-ng[version='>=4.9|>=7.3.0|>=7.5.0|>=9.3.0|>=7.2.0']

I had pinged @conda-forge/blis a couple of times in conda-forge/scipy-feedstock#130 (not realising it was just you 😅), and so I just thought I'd now try fixing it myself. Looking at this from the other side, what possible downside could it have to rerender this feedstock?

[h-vetinari]

Closing as superseded by #21