This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Someone broke bootstrap with gfortran, again!


On Sat, Jul 23, 2005 at 10:17:25AM -0700, Steve Kargl wrote:
> Note, you need to start with either an empty object tree
> and do a complete bootstrap or remove the libgfortran directory
> and do a bubblestrap.

I always start from zero.

I can see a failure from pinski's reduced test case.  Fixed
by the following.  You see if it fixes whatever the hell you
are seeing that I can't on amd64-linux.


r~



Index: tree-complex.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/tree-complex.c,v
retrieving revision 2.37
diff -u -p -d -r2.37 tree-complex.c
--- tree-complex.c	22 Jul 2005 00:33:46 -0000	2.37
+++ tree-complex.c	23 Jul 2005 17:30:24 -0000
@@ -517,7 +517,7 @@ set_component_ssa_name (tree ssa_name, b
 	  && !DECL_IGNORED_P (SSA_NAME_VAR (ssa_name)))
 	{
 	  comp = get_component_var (SSA_NAME_VAR (ssa_name), imag_p);
-	  SSA_NAME_VAR (value) = comp;
+	  replace_ssa_name_symbol (value, comp);
 	}
 
       VEC_replace (tree, complex_ssa_name_components, ssa_name_index, value);


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]