> Can you file a PR and then make the test KFAIL for i386,
> with reference to the bug report.  My recollection of the
> policy is: new PASS is good, new KFAIL is really good,
> new FAIL is bad.

By necessity, there is a great deal of flexability here.  While it is 
reasonable to expect the new tests to be fully resolved for one specific 
architecture, that expectation really doesn't extend be extended to 
other architectures - otherwize we'd never get new tests :-)  Anyway, 
I'll try to make the mechanism of adding the KFAILs easier, at present 
it is a bit messy (as in almost impossible :-) :-(

I've attached a more current work-in-progress.

Note that for PPC, you'll want to apply this unapproved patch:
http://sources.redhat.com/ml/gdb-patches/2003-10/msg00626.html
as otherwize the test trips over PPC GDB bugs.

Andrew