sco5.h patch

sco5.h patch

[J. Kean Johnston]

Hello all.

Below is a small patch to sco5.h which corects the crt* inclusion. When
I made ELF the default rather than COFF, I missed this stuff. Without
this patch users cannot use gcc -static to create a statically linked ELF
file. Not good.

Can this be included in both egcs 1.0.1 and gcc 2.8.0. It is a very low
risk change and affects, obviously, only SCO Open Server 5. The benefit
of this is it allows the compiler to work as documented.

And just to give Kenner and the Cygnus folks a heads up: there is probably
going to be one more chnge to this file as soon as I can verify and rerun
all the regression tests. Seems like in COFF mode the compiler is suffering
from considerable bloat. Initial analysis indicates a reversal of a previous
patch by Robert Lipe to disable HAVE_ATEXIT is called for.

Kean.

PS. This patch works on both, but is offset by 1 line in egcs. This was made
against the gcc-2.8.0-971225 snapshot.

*** config/i386/sco5.h.jkj	Mon Jan  5 13:22:44 1998
--- config/i386/sco5.h	Mon Jan  5 13:26:08 1998
***************
*** 828,839 ****
         %{!Xc:%{Xk:values-Xk.o%s} \
          %{!Xk:%{Xt:values-Xt.o%s} \
           %{!Xt:values-Xa.o%s}}}}}} \
!   %{mcoff:crtbeginS.o%s} \
!   %{!mcoff:%{!static:crtbegin.o%s}%{static:crtbeginS.o%s}}"
  
  #undef ENDFILE_SPEC
  #define ENDFILE_SPEC \
!  "%{!mcoff:%{!static:crtend.o%s}%{static:crtendS.o%s}} \
    %{mcoff:crtendS.o%s} \
    %{pg:gcrtn.o%s}%{!pg:crtn.o%s}"
  
--- 828,838 ----
         %{!Xc:%{Xk:values-Xk.o%s} \
          %{!Xk:%{Xt:values-Xt.o%s} \
           %{!Xt:values-Xa.o%s}}}}}} \
!   %{mcoff:crtbeginS.o%s} %{!mcoff:crtbegin.o%s}"
  
  #undef ENDFILE_SPEC
  #define ENDFILE_SPEC \
!  "%{!mcoff:crtend.o%s} \
    %{mcoff:crtendS.o%s} \
    %{pg:gcrtn.o%s}%{!pg:crtn.o%s}"
  

Re: sco5.h patch

[Jeffrey A Law]

  In message < 19980105133410.03908@sco.com >you write:
  > Can this be included in both egcs 1.0.1 and gcc 2.8.0. It is a very low
  > risk change and affects, obviously, only SCO Open Server 5. The benefit
  > of this is it allows the compiler to work as documented.
Unfortunately, egcs-1.0.1 was already on the ftp sites when you sent in
the patch, so it was too late to get it into 1.0.1.

However, if we have egcs-1.0.2, we could/should include this patch.

I've installed it into the mainline sources.

jeff