I'm not sure why this was timing out when running on my machine, since it took only 45 second for the tests to fail and the timeout seem to be set to 240 seconds.
Simply skipping these cases wasn't enough because the inferior was being built 12 times (once for each call to setUp). So I moved the setUp work into the tests themselves. Skipping stops wasting time building and rebuilding the inferiors. It also seems to avoid the last lingering intermittent GNU Make crash.
Before going this route, I spent a lot of time trying to get the inferior to build on Windows. I had to fix some POSIX assumptions in common.h and enable exceptions (which have been disabled on Windows because clang-cl exception support on Windows was lagging). Enabling exceptions caused four other tests to fail, so we have that to look forward to. And I wasn't able to figure out the linking problem where it complained that the -L option was ignored and then complained that it couldn't find liblldb.lib (because it ignored the -L option).
I'm open to other ideas, but this is the best way I can see to get unblocked.