ERI test creates more sym links than it needs to

[billsacks]

The ERI test creates a bunch of unnecessary sym links in the main case (i.e., the case without a 'ref' extension), pointing to history files in the ref2 archive space. As far as I can tell, this doesn't cause problems with the test, but it makes it hard to understand what's going on when you're trying to debug a failed test: It means that the history files created by the main case are actually put in the ref2 archive space! I could also imagine this causing problems in some edge cases, e.g., involving rerunning a test, but I'm not sure about that.

Specifically, I ran ERI_Lm36.T31_g37_gl4.IGCLM45.yellowstone_intel.clm-glcMEC_long, but this should show up in many ERI tests.

It looks like these come from this code in eri.py:

        # the following lines creates the initial component history files for the restart test
        for item in glob.glob("%s/*/hist/*nc" % dout_sr2):
            newfile = "%s" % item.replace(".ref2", "")
            newfile = os.path.basename(newfile)
            dst = os.path.join(rundir, newfile)
            if os.path.exists(dst):
                os.remove(dst)
            os.symlink(item, dst)

This block follows immediately after:

        _helper(dout_sr2, refdate_3, refsec_3, rundir)

and I would think that that _helper is enough to create the necessary sym links to run the main case.

@jgfouca @mvertens @jedwards4b - it looks like you're the ones who wrote the python version of this test. Can any of you comment on this?