This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Bootstrap failure w/3.2 trunk on sparc-solaris2.7 in insn-attrtab.c
- From: "Kaveh R. Ghazi" <ghazi at caip dot rutgers dot edu>
- To: gcc-bugs at gcc dot gnu dot org
- Cc: davem at redhat dot com, vmakarov at redhat dot com
- Date: Thu, 2 May 2002 13:15:08 -0400 (EDT)
- Subject: Bootstrap failure w/3.2 trunk on sparc-solaris2.7 in insn-attrtab.c
There's a new bootstrap failure on the 3.2 trunk for
sparc-sun-solaris2.7 when using cc for stage1. I get this:
> cc -DIN_GCC -g -DHAVE_CONFIG_H -I. -I. -I../../egcc-CVS20020501/gcc
> -I../../egcc-CVS20020501/gcc/. -I../../egcc-CVS20020501/gcc/config
> -I../../egcc-CVS20020501/gcc/../include -c insn-attrtab.c -o
> insn-attrtab.o
> "insn-attrtab.c", line 11546: syntax error before or at: }
> "insn-attrtab.c", line 11546: warning: syntax error: empty declaration
> cc: acomp failed for insn-attrtab.c
> make[2]: *** [insn-attrtab.o] Error 2
The relevant declaration in insn-attrtab.c:11546 is:
> const struct function_unit_desc function_units[] = {
> };
GCC also warns about this not being valid:
> insn-attrtab.c:11546: warning: ISO C forbids empty initializer braces
So any chip/platform for which this struct is empty will choke when
using an ISO C compliant compiler.
I'm pretty sure it started happening around the time of the dfa-branch
merge. My guess is that you either need to avoid emitting the struct
if its empty or insert a dummy value if that's too problematic.
--Kaveh
--
Kaveh R. Ghazi Director of Systems Architecture
ghazi@caip.rutgers.edu Qwest Global Services