Handling of SIGINT

[Alex Vandiver]

On Thu, 2009-02-19 at 12:34 -0500, David Golden wrote:
> On Thu, Feb 19, 2009 at 12:18 PM, Andy Armstrong <andy at hexten.net> wrote:
> > Why isn't the SIGINT going to the test script?
> 
> I think on Win32, at least, SIGINT goes to everything on the console
> simultaneously.  [snip]

Hrm.  The odd thing is, this isn't Win32, this is SPAR^H^H^H^HLinux.  At
first I assumed something odd relating to double-forking, but even that
can't actually explain what I'm seeing -- certainly, simple test cases
correctly pass on the INT to their kids correctly.  The odd thing is,
the $SIG{INT}=sub{kill INT=>$self->{pid}} actually seems to solve the
problem, which is even more inexplicable, now that I think about it.
*mumble*  I'll see if I can strip this down to a smaller test case later
tonight.
 - Alex

-- 
Networking -- only one letter away from not working