Should there be a gdb_test_backtrace?

[Andrew Cagney]

Randolph, Michael,

I've been thinking about that HP/UX problem (where there's an extra 
frame) along with a problem of my own - needing to change:
	<signal handler called>
to
	<signal trampoline> at 0x12345

I'm wondering of the backtrace tests should be handled using a new 
expect procedure like:

gdb_test_backtrace \
	"keeper at line ..."
	signal-trampoline
	"bowler at ..."
	"main ..."

it would then:
	- add 'N#' prefixes
	- expand signal-trampoline per local os/format

thoughts?  Should it take an optional backtrace limit?

Andrew

(no I've not implemented it, and I don't know how hard the 
implementation would be :-)