detect submodule issues in autogen.sh

[jeffhammond]

This was discussed over email on August 29, 2018 but I am creating an issue for tracking.

GitHub makes it really easy for users to clone without --recursive and thus not get the submodules.

./autogen.sh should detect the lack of submodules and provide a helpful error message rather failing as follows:

=== Applying patches ===
Applying patch to src/hwloc...
Applying: hwloc: separated visibility cflags for better control.
error: sha1 information is lacking or useless (config/hwloc.m4).
error: could not build fake ancestor
Patch failed at 0001 hwloc: separated visibility cflags for better control.
The copy of the patch that failed is found in: .git/rebase-apply/patch
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".
done

=== Running third-party initialization in src/hwloc ===
./autogen.sh: line 147: ./autogen.sh: No such file or directory

[nitinehpc]

Ha Ha, I am also seeing the same issue.

$ ./autogen.sh
Checking for autoconf version...done
Checking for automake version...done
Checking for libtool version...done

Checking header file mpi.h at /apps/mpi/intel64/bin/../ ...done
Found /apps/mpi/intel64/include/mpi.h
Generating MPI wrappers... mpihstyle=_mpich
done

Checking header file mpio.h at / apps/mpi/intel64/bin/../ ...done
Found / apps/mpi/intel64/include/mpio.h
Generating MPI IO wrappers... mpihstyle=_mpich
done

Syncronizing confdb -> test/confdb... done

=== Applying patches ===
Applying patch to src/hwloc...
fatal: Not a git repository (or any parent up to mount point /home)
Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).
done

=== Running third-party initialization in src/hwloc ===
./autogen.sh: line 147: ./autogen.sh: No such file or directory

[jeffhammond]

The solution is to git clone --recursive...

[devreal]

It appears that http://git.mpich.org/ is not available anymore (timing out). Looks like openpa has moved to https://github.com/pmodels/openpa...

[jeffhammond]

Yeah, I think the migration of OpenPA is recent. I'm sure one of the Argonne folks will fix this soon enough.

[jeffhammond]

The OpenPA repo location issue was fixed in 242c7ca and merged just now (#34).

[jeffhammond]

Attempting a fix in #38...