Hi, In r193053, input_location was mistakenly used. This patch changed to use curr_insn_location instead. This looks like an obvious patch, can I check it in to trunk?
Bootstrapped and passed gcc/gdb regression test. Thanks, Dehao gcc/ChangeLog: 2012-11-01 Dehao Chen <de...@google.com> * expr.c (expand_expr_real_1): Change to not using input_location.
Index: gcc/expr.c =================================================================== --- gcc/expr.c (revision 193071) +++ gcc/expr.c (working copy) @@ -9221,13 +9221,11 @@ expand_expr_real_1 (tree exp, rtx target, enum mac if (g) { rtx r; - location_t saved_loc = input_location; + location_t saved_loc = curr_insn_location (); - input_location = gimple_location (g); - set_curr_insn_location (input_location); + set_curr_insn_location (gimple_location (g)); r = expand_expr_real (gimple_assign_rhs_to_tree (g), target, tmode, modifier, NULL); - input_location = saved_loc; set_curr_insn_location (saved_loc); if (REG_P (r) && !REG_EXPR (r)) set_reg_attrs_for_decl_rtl (SSA_NAME_VAR (exp), r);