From: Emil Ong <emilong@...>
Date: 2001-07-06T04:11:17+09:00
Subject: [ruby-talk:17337] Re: SIGUSR1/2 ?

Hi,

On Fri, 6 Jul 2001, ts wrote:

> >>>>> "E" == Emil Ong <onge@mcs.anl.gov> writes:
> 
> E> I'm embedding ruby into an application that already uses SIGUSR1 and
> E> SIGUSR2.  Is there any reason why ruby catches these signals?  Can I
> E> safely modify signal.c so that it doesn't catch them?  If they aren't
> E> really necessary, would it be reasonable to have a function that stopped
> E> ruby from catching them in the C API (e.g. rb_ignore_sigusr())?
> 
>  Have you tried to re-install your sighandler after the call to 
>  ruby_init() ? 

I didn't try that, but I think it may be difficult.  I'm using a library
which implements a standard API, so I'd like to be able to use it 
opaquely.  Moreover, this library's initialization function has to be
called before anything else (including ruby_init()) because of a unique
runtime system. 

If it's not feasible, I understand, but I thought it might be useful to
other people using SIGUSR[12].

Thanks,
Emil