Hi, Richard,
Please ignore my previous email sent this morning. -:)
Actually, I studied more about the heuristic, and I agreed that the simplified
loop as you suggested below is the correct approach to go, my current
two-level nested loop is a little too complicate and confusing.
> On Aug 13, 2025, at 05:24, Richard Biener <richard.guent...@gmail.com> wrote:
>
> So the overall CFG walk should be like
>
> do
> {
> cond_bb = get_immediate_dominator (CDI_DOMINATORS, cur_bb);
> if (single_pred_p (curr_bb))
> {
> auto gsi = gsi_last_bb (cond_bb);
> if (!gsi_end_p (gsi)
> && stmt_ends_bb_p (gsi_stmt (gsi)))
> {
> ... handle control stmt at the end on cond_bb ...
> }
> }
> curr_bb = cond_bb;
> }
> while (depth < flag_diagnostics_show_context
> && curr_bb != ENTRY_BLOCK_PTR_FOR_FN (cfun));
>
> stmt_ends_bb_p is is_ctrl_stmt () || is_ctrl_altering_stmt where the
> former includes GIMPLE_COND (and switch and more) and the
> latter includes for example a call that internally throws. Whether
> we terminate the walk upon unhandled control stmts would be
> up to debate, I'd skip for now, as you do and as you also skip
> GIMPLE_COND without location.
>
Thanks a lot for your suggestion.
Qing
