[IRCServices Coding] DUMPCORE only unhandles SIGSEGV?

[Finny Merrill]

On Sat, 12 Jul 2003 03:19:34 JST, Andrew Church <achurch at achurch.org> 
wrote:

> SIGFPE only happens on divide by zero, and in the few places where
> Services divides by a variable the variable is already tested for being
> nonzero.  SIGQUIT is user-generated.  SIGABRT isn't generated at all.
> SIGILL only occurs if your compiler is broken or you run a binary 
> compiled
> for the wrong CPU.
>

Well, we recently had a problem where our services (not ircservices but 
based on it)
went into a spinlock, and there was no way we could even find out what it 
was because
as soon as it happened the process monitor killed it for eating CPU, and 
there was no
way we could get it to core. This has never happened to me with 
ircservices, but what
if it did one day? The same problems would occur.