Why is "fortran" censored in the log display?

Code to Cloud GitHub Actions
#1

This is an odd one, but I noticed it today in trying to get GitHub Actions working on a repo. Namely, when I watch the log output scroll by I see:

  make verify ESMF_COMM=openmpi ESMF_COMPILER=g***
  shell: sh -e {0}
  env:
    OMPI_ALLOW_RUN_AS_ROOT: 1
    OMPI_ALLOW_RUN_AS_ROOT_CONFIRM: 1
    OMPI_MCA_btl_vader_single_copy_mechanism: none
    FC: g***
    CC: gcc
    CXX: g++

for this bit of the yaml file:

    env:
      OMPI_ALLOW_RUN_AS_ROOT: 1
      OMPI_ALLOW_RUN_AS_ROOT_CONFIRM: 1
      OMPI_MCA_btl_vader_single_copy_mechanism: none
      FC: gfortran
      CC: gcc
      CXX: g++

and in one of the steps I have:

      - name: Versions etc.
        run: |
          gcc --version
          gfortran --version
          mpirun --version
          mpicc --show
          mpic++ --show
          mpifort --show
          echo $BASEDIR
          git submodule status

and the output is:

gcc (GCC) 10.2.0
Copyright (C) 2020 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

GNU Fortran (GCC) 10.2.0
Copyright (C) 2020 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

mpirun (Open MPI) 4.0.5

Report bugs to http://www.open-mpi.org/community/help/
gcc -I/opt/openmpi-4.0.5/include -pthread -Wl,-rpath -Wl,/opt/openmpi-4.0.5/lib -Wl,--enable-new-dtags -L/opt/openmpi-4.0.5/lib -lmpi
g++ -I/opt/openmpi-4.0.5/include -pthread -Wl,-rpath -Wl,/opt/openmpi-4.0.5/lib -Wl,--enable-new-dtags -L/opt/openmpi-4.0.5/lib -lmpi
g*** -I/opt/openmpi-4.0.5/include -pthread -I/opt/openmpi-4.0.5/lib -Wl,-rpath -Wl,/opt/openmpi-4.0.5/lib -Wl,--enable-new-dtags -L/opt/openmpi-4.0.5/lib -lmpi_usempif08 -lmpi_usempi_ignore_tkr -lmpi_mpifh -lmpi

And then later on in the logging:

/bin/bash ../../libtool --tag=CXX  --tag=CC   --mode=link mpicc  -fPIC   -fopenmp -std=c99 -D_BSD_SOURCE -D_POSIX_SOURCE  -L/__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib -lnetcdf -ljpeg -lmfhdf -ldf -ljpeg -lhdf5_hl -lhdf5 -lm -L/__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib -lsz -ljpeg -L/__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib -lcurl -lssl -lcrypto -lz -lm -ldl -lm -L/__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib -L/__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux//lib   -L/__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib -o ncpdq ncpdq.o libnco.la -lexpat -lgsl -lm -lnetcdf -lnetcdf -lnetcdf -L/__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib /__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib/libhdf5_hl.a /__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib/libhdf5.a /__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib/libz.a /__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib/libsz.a /__w/ESMA-Baselibs/x86_64-pc-linux-gnu/g***/Linux/lib/libcurl.a -lmfhdf -...

Each one of those g*** is a gfortran so for some reason “fortran” is changed to “***”…but why? gcc, good; g++, good, gfortran…bad? I mean, is there some “Fortran is not a Bad Word” setting I need to add to my workflow?

Note: things seem to work, but I’m just really baffled by this.

#2

That is a bit weird. :thinking: Do you have a secret that contains fortran in it? It might be the secret masker that’s doing it.

#3

@thomasshaped Oooh. I did not know that exisited!

Guess what I added to this workflow when I was updating it? Docker Hub Authorization for my image pull because of the whole “limits on anonymous pulls” thing they announced. And want to guess my Docker Hub username? (Not too much of a reveal as it’s my default internet username almost everywhere.)

Though now as I read https://github.com/actions/virtual-environments/issues/1445, maybe I don’t need to worry? Sounds like GitHub might have reached an agreement that it isn’t needed for public images pulls into GitHub Actions?

Sigh. Thanks. Weird mystery solved. (Though I guess this could make a very inefficient attack on secrets.)

1 Like
#4

Oh no way, wild guess but I’m glad it was right! :sweat_smile:

I believe this is the case, yep! Public pulls of images should not be affected by these rate limits when used on GitHub-hosted runners.