Do we care about malloc failure in tests?

[Roland McGrath]

We have exit(77) to yield UNSUPPORTED.  But IMHO using that for cases like
malloc failure is worse than FAIL.  We could add another test outcome for
such badness, but I don't see any urgency in doing any such thing.  If
tests fail due to such an issue, you'll look at what went wrong, be
confused, and try running the tests again, when they'll work.