Re: Positive feedback on new gcc :) & fp exception question
On Wed, 21 Jan 1998, Guenter Geiger wrote:
> Sorry I was a little bit confused here. In fact I was using a .67
> with applied pnp patch. I get SIOCADDRT: invalid argument and
> different RPC call error messages. This is probably related to nfs.
> I'll try .72 too. Do you use nfs ?
Yes, I do. In fact, that's how I compile most of my packages since I have
such a small hard drive :) As far as the SIOCADDRT problem, when do you
get this error? If it's on startup, that's because sysvinit hasn't been
revised yet to conform to the new routing requirements of the 2.1 kernels
yet. They are alot stricter but also only need one 'route' command
instead of three.
> Unlukily it doesn't seem to work. Still a floating point exception.
> But you have pointed me to the right direction, since in the first
> place I thought these exceptions are handled by the kernel. I will
> try to investigate further on this.
Hmmm...try turning off optimisation on the compile and see if it works
with the -mieee switch. Worse comes to worse, check the info files on
gcc. It has ALOT of options related to the new -m switches.
> I had a fatal signal 6 when compiling csound -
> I don't know what this is meant to be, but I
> can try to figure out where in the code it occured exactly.
Change the -O2 to -O and it should work. There's still a bug in the
optimisation code that Richard Henderson is trying to work out. I have
the same problem with qmail (egcs and gcc 2.8 suffer from it still).
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
Trouble? e-mail to email@example.com .