[qmtest] Patches for scheduler updates

[Mark Mitchell]

> Tomorrow I'll make an attempt at digging into these failures in the
> hopes of figuring out what's going on, but thought people might like a
> heads up and a chance to try to reproduce this... or point out the
> pilot error causing all this ;-).

The problem with bad_target2 is that bad_target.qmt has this target
group pattern:

  $^

And the GNUmakefile.in creates the thread_target without giving it a
group, which makes it default to the empty string.

So, actually, QMTest is honoring the target specification.

The way to fix that is to explicitly use a bogus name rather than "$^"
as a group pattern.  

The reason this did not show up without parallelism is that QMTest puts
the default "local" target in the "local" group.  We could debate
whether or not QMTest should use the empty string in that case, or
whether every target's group should default to its name, if no explicit
group is specified.

I've got fixes for the bad_target tests locally and will check them in
soonish.

-- 
Mark Mitchell
CodeSourcery, LLC
mark at codesourcery.com