signals on latest linux kernels
Jauder Ho
jauderho@transmeta.com
Thu Dec 4 00:44:00 GMT 1997
yeah, I was compiling this against a stock libc5.3.12
On Wed, 3 Dec 1997, David S. Miller wrote:
> Date: Thu, 04 Dec 1997 00:53:10 -0700
> From: Jeffrey A Law <law@cygnus.com>
>
> In message < Pine.LNX.3.96.971203233738.29782A-100000@turtle.transmeta.com >you
> write:
> > linus changed the way signals are handled in the latest kernels..
> > compiling egcs will fail at libiberty/strsignal.c
> >
> > at line 246...
> >
> > the NSIG should be changed to a _NSIG and that should cause gcc to
> > compile fine then..
> This is not a change we can make since it would break all the existing
> systems that use that code -- including all versions of linux that
> don't have Linus's recent change.
>
> We will need to find some other way to handle this problem that doesn't
> break other targets.
>
> Suggestions anyone?
>
> The correct fix is in GLIBC, it should define NSIG to _NSIG or
> something along these lines. In fact with 2.0.x GLIBC I notice the
> following in /usr/include/signal.h
>
> #ifdef __USE_MISC
> #define NSIG _NSIG
> #endif
>
> I suppose Richard of Ulrich can comment more precisely.
>
> Later,
> David S. Miller
> davem@dm.cobaltmicro.com
>
IMPORTANT NOTICE TO RECIPIENTS: The Entire Physical Universe, Including
This Product, May One Day Collapse Back Into an Infinitesimally Small
Space. Should Another Universe Subsequently and Spontaneously Re-Emerge,
the Existence of This Message in That Universe Cannot be Guaranteed.
More information about the Gcc-bugs
mailing list