This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
RE: Compilation problem in Dynix/ptx
- To: robertlipe at usa dot net, biswapesh dot chattopadhyay at bt dot com
- Subject: RE: Compilation problem in Dynix/ptx
- From: biswapesh dot chattopadhyay at bt dot com
- Date: Mon, 11 Dec 2000 17:38:36 -0000
- Cc: gcc at gcc dot gnu dot org
Thanks Robert - see my previous post - I *did* configure with --with-gnu-as
--with-gnu-ld - the problem was solved when I moved as to gas, according to
a previous message on the ML.
Basically three patches need to be made to compile 2.95.2 succesfully (c and
c++) on Dynix/ptx V4.4.8 i386:
a) This problem with as (needs to have a better fix)
b) The problem of s_sys_siglist not being defined (needs better patch than
NO_SYS_SIGLIST in config.h ?)
c) The patch in gcc/cp/class.c as detailed in
http://gcc.gnu.org/ml/gcc-help/2000-06/msg00088.html .
I'll try and compile a recent snapshot and see what I get.
Thanks for your help.
Biswa.
> -----Original Message-----
> From: Robert Lipe [mailto:robertlipe@usa.net]
> Sent: 11 December 2000 17:32
> To: biswapesh.chattopadhyay@bt.com
> Cc: gcc@gcc.gnu.org
> Subject: Re: Compilation problem in Dynix/ptx
>
>
> biswapesh.chattopadhyay@bt.com wrote:
> > Well, I thought (a) would work but now I get something
> completely different:
>
> Completely different problem.
>
> > done
> > _muldi3
> > /var/opt/spool/out/em/biswa/gnu/i386-sequent-sysv4/bin/as:
> unrecognized
> > option `-no_0f_fix'
>
> It looks like you're runing GNU assembler without having configured
> --with-gas. config/ptx4.h adds the '-no_0f_fix'. Rats. And
> it does it
> in a dorky way that ignores --with-gas.
>
> If you're using an a newer/older PTX assembler or GAS that doesn't
> understand this flag (and I have no idea what it is) and you
> don't think
> you need it, just whack it from ASM_SPEC. Other than the addition of
> that flag, it looks very SVR4-ish.
>
> You should also propose a Real Fix to this problem for the GCC3 branch
> for PTX. This is not a heavily tested target, so the onus
> really is on
> you to fix it. :-)
>
> RJL
>