Hi,
> Has anyone managed to compile and succesfully run GC 4.01 on a Solaris
> 2.4 machine?
>
> I'm thinking I'll have to rvert back to 3.53. This is the output when
> I start the program.
I've solved the problem, thanks to "Don". Turns out there's a compiler bug
in gcc 2.7.x which prevents it from compiling gs4.01 properly. I wonder if
that doesn't explain some other bugs I've found in other programs... Hmmmm.
My thanks also to
Tim Evans
Todd C. Miller
for their input. For completeness's (sp?) sake, here is what the readme for
gs4.01 says:
gcc (all platforms):
There is an optimizer bug in gcc 2.7.0, 2.7.1, and 2.7.2 that causes
these versions of gcc to generate incorrect code. You can work around this
by adding the switch
-Dconst=
to the compilation switches (CFLAGS). Alternatively, you can rebuild gcc
with the following patch:
*** clean-ss-951203/expr.c Sun Nov 26 08:18:07 1995
--- ss-951203/expr.c Mon Dec 4 11:41:18 1995
*************** expand_expr (exp, target, tmode, modifie
*** 4582,4587 ****
through a pointer to const does not mean that the value there can
never change. Languages where it can never change should
also set TREE_STATIC. */
! RTX_UNCHANGING_P (temp) = TREE_READONLY (exp) | TREE_STATIC (exp);
return temp;
}
--- 4582,4588 ----
through a pointer to const does not mean that the value there can
never change. Languages where it can never change should
also set TREE_STATIC. */
! RTX_UNCHANGING_P (temp) = TREE_READONLY (exp) & TREE_STATIC (exp);
return temp;
}
If possible, compile with gcc 2.5.8 or 2.6.3, which don't have this bug.
-- Laurent Duperval, Systems Analyst Grafnetix Systems Inc. Tel: (514) 861-3389 777, de la Commune Ouest Fax: (514) 866-6206 Suite 101, Montreal, Qc Laurent@Grafnetix.COM Canada, H3C 1Y1 URL: http://www.Grafnetix.COM/~laurent
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:11:06 CDT