[JVMTI] RFA: JVMTI Stack Tracing

[Hans Boehm]

On Fri, 20 Oct 2006, David Daney wrote:
[Use of SIGPWR]
> It would.  I almost flagged this myself, but I didn't have the time to
> verify which signals the GC uses.
>
> Would it be possible to use a Real-Time signal?  That might not be as
> portable as using a 'Standard' signal.
>
The GC uses SIGPWR on Linux primarily for historical reasons.  If I
were writing it now, I would probably use a real-time signal, as the GC
does on some platforms.

It may make sense to also change the GC and use real-time signals
consistently everywhere.  But that would force all of
us who are used to ignoring SIGPWR and SIGXCPU in gdb to reprogram our
fingers or gdb scripts.  A secondary issue is that there seems to be no
standard mechanism for allocating real-time signals, and hence I think
you have to worry about conflicts there as well.

Hans