OSR5.02 and whoops solved (?)

OSR5.02 and whoops solved (?)

[Bill Walker]

The problem I am having with OSR 5.0.2 getting va_list defined wrongly seems
to be due to the existence of an _old_ gcc 2.7.2.1 on the system I am
making egcs on.  That gcc has a varargs.h file that is being
picked up by the top level "make bootstrap".  The top level "make bootstrap"
is using that old gcc as the native compiler.

A "work around" might be this:

	Go ahead and make the gcc in the egcs distribution in three stages,
and install it.  Then run make at the top level, and install again.


	I've got a _lot_ of folks depending on the gcc 2.7.2.1 compiler
to incorrectly compile a large software project, so I can't just remove
it right now.  That would be the simplest solution by far.

	(The software project depends on "amulet", which compiles
cleanly under 2.7.2.1, but the better error detection under egcs
stops it!  The folks involved rather have it compileable, but wrong.
Go figure. )







-- 
Bill Walker Ph.D.
Chairman, Dept. of Computer Science
East Central University
Ada, Oklahoma 74820-6899

e-mail:  bw@cs.ecok.edu 
phone:   580 332 8000 ext. 594
FAX:     580 332 4616

Re: OSR5.02 and whoops solved (?)

[Jeffrey A Law]

  In message < m0xVggb-0000O5C@student.ecok.edu >you write:
  > 
  > The problem I am having with OSR 5.0.2 getting va_list defined wrongly
  > seems to be due to the existence of an _old_ gcc 2.7.2.1 on the system I am
  > making egcs on.  That gcc has a varargs.h file that is being
  > picked up by the top level "make bootstrap".  The top level "make bootstrap"
  > is using that old gcc as the native compiler.
Hmmm, if the problem is just during the stage1 build you might be able to
do something like

setenv CC /bin/cc  (or some other non-broken compiler)

configure
make bootstrap
make install


jeff

Re: OSR5.02 and whoops solved (?)

[Robert Lipe]

> The problem I am having with OSR 5.0.2 getting va_list defined wrongly seems
> to be due to the existence of an _old_ gcc 2.7.2.1 on the system I am
> making egcs on.  That gcc has a varargs.h file that is being
> picked up by the top level "make bootstrap".  The top level "make bootstrap"
> is using that old gcc as the native compiler.

Yes, I've been looking at your system {as you requested} this morning.

The gcc you have on this system does not support this system.


$ uname -X | grep Rel
Release = 3.2v5.0.2
$ gcc -v /tmp/x.c
Reading specs from /usr/local/lib/gcc-lib/i486-unknown-sco3.2v4.2/2.7.2.1/specs
gcc version 2.7.2.1

[ Many other clues that this is a 3.2v4 targeted gcc that was 
**almost** hacked to run on OpenServer 5 (a.k.a. 3.2v5).

What you're seeing is one of the many many things between "almost 
ported" and "ported".

In fact, this is covered in the FAQ for c.u.s.p.

	http://www.dgii.com/cgi-bin/scoprgfaq/faq.pl?file=6

with some additional information at:
	
	ftp://ftp.dgii.com/users/robertl/scods/gds_faq.html


If you knew how broken it was, you'd be afraid.  Very afraid.

> 	I've got a _lot_ of folks depending on the gcc 2.7.2.1 compiler
> to incorrectly compile a large software project, so I can't just remove
> it right now.  That would be the simplest solution by far.

Simpler yet:

CC=path to workign compiler configure
CC=path to working compiler make bootstrap

More specifically:

CC="/bin/cc -belf" configure
CC="/bin/cc -belf" make bootstrap

Enjoy.

RJL