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: Regression for trunk on i686-pc-linux-gnu



On Jul 27, 2004, at 3:41 PM, Richard Kenner wrote:


cvs up -D20040722 tree.def tree-ssa-operands.c gimplify.c tree-gimple.c
ought to give you a working compiler.


OK, it was a typo in tree-gimple.c.  But I can't test it since the
tree is broken on my machine.  Seems obviously the error, though:

*** tree-gimple.c 24 Jul 2004 01:29:11 -0000 2.18
--- tree-gimple.c 27 Jul 2004 19:38:10 -0000
*************** is_gimple_addr_expr_arg (tree t)
*** 324,328 ****
{
return (is_gimple_id (t) || handled_component_p (t)
! || TREE_CODE (t) == REALPART_EXPR || TREE_CODE (t) == IMAGPART_EXPR);
}


--- 324,329 ----
{
return (is_gimple_id (t) || handled_component_p (t)
! || TREE_CODE (t) == REALPART_EXPR || TREE_CODE (t) == IMAGPART_EXPR
! || TREE_CODE (t) == INDIRECT_REF);
}



Uh, according to the grammar:


   addr-expr-arg: ID
    | compref

so the original code is correct.

I thought &* got folded away anyway.


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