[RFA/RFC] Tweak for a gdb.mi test.

[Andrew Cagney]

> Andrew Cagney wrote:
> 
>> 
> 
>> >>  From memory, a suggestion was to let people select the back-trace
>> >> policy independant of the current architecture.
> 
>> >
>> >
>> > I thought we also had a policy of not inserting tests
>> > that we knew would fail on some targets?  Something about
>> > this being a regression test...
> 
>> 
>> My understanding of the ``the regression test'' is that it stops people
>> adding ``feature'' tests that demonstrate bugs in things that have never
>> worked (or are not even implemented).  It avoids, among other things,
>> the problem of not knowing of a FAIL is a bug in GDB or in a testcase
>> that never worked.
>> 
>> I look at this test as something similar to call-ar-st.  It has been
>> demonstrate to work on one platform (I think the original was Arm/eCos
>> but more recently GNU/Linux i386) and should work on other platforms.
> 
> 
> OK, I yield.  Patch withdrawn.

The sooner we get KFAIL the better!

Andrew