Re: [tree-ssa] Bootstrap failure on powerpc-apple-darwin

[Andrew Pinski <pinskia at physics dot uc dot edu>]

On Nov 14, 2003, at 5:21 PM, Jan Hubicka wrote:
> The patch...
>
> Index: gimple-low.c
> ===================================================================
> RCS file: /cvs/gcc/gcc/gcc/Attic/gimple-low.c,v
> retrieving revision 1.1.4.11
> diff -c -3 -p -r1.1.4.11 gimple-low.c
> *** gimple-low.c	12 Nov 2003 22:06:25 -0000	1.1.4.11
> --- gimple-low.c	15 Nov 2003 01:06:12 -0000
> *************** lower_bind_expr (tree_stmt_iterator *tsi
> *** 157,163 ****
>     tree old_block = data->block;
>     tree stmt = tsi_stmt (*tsi);
>
> !   if (BIND_EXPR_BLOCK (stmt))
>       {
>         data->block = BIND_EXPR_BLOCK (stmt);
>
> --- 157,163 ----
>     tree old_block = data->block;
>     tree stmt = tsi_stmt (*tsi);
>
> !   if (BIND_EXPR_BLOCK (stmt) && BIND_EXPR_BLOCK (stmt) != old_block)
>       {
>         data->block = BIND_EXPR_BLOCK (stmt);
>
> *************** expand_used_vars (void)
> *** 339,345 ****
>     cfun->unexpanded_var_list = nreverse (cfun->unexpanded_var_list);
>
>     for (cell = cfun->unexpanded_var_list; cell; cell = TREE_CHAIN 
> (cell))
> !     expand_var (TREE_VALUE (cell));
>
>     cfun->unexpanded_var_list = NULL_TREE;
>   }

This patch fixed it for me, will send results after finishing bootstrap

Thanks,
Andrew Pinski