On Mon Jun 09 17:12:36 2008, [EMAIL PROTECTED] wrote:
> On Tue Jul 17 22:23:38 2007, petdance wrote:
> > In function branch_branch, there's a line like this:
> >
> > ins->r[get_branch_regno(ins)] = next->r[0];
> >
> > That call to get_branch_regno() can return -1. That would make a
> > negative array index, and stomp on who knows what.
> >
>
>
> Grepping ./compilers/imcc/optimizer.c for get_branch_regno, I get:
>
> 365: (reg = get_branch_regno(last)) >= 0) {
> 1082: if (get_branch_regno(ins) >= 0) {
> 1094: const int regno = get_branch_regno(ins);
> 1251: reg_index = get_branch_regno(cond);
> 1327: get_branch_regno(cond) >= 0) {
>
> ... which suggests that the code Andy was referencing no longer appears
> in this file.
>
> Are the issues raised in this thread still pertinent?
>
> Thank you very much.
> kid51
>
The code seems to check for a negative value and error now, so no, it doesn't
seem to be an
issue. I think this is the same issue that was picked up in RT#46227, we just
didn't notice
until now.
Resolving ticket.
--
Will "Coke" Coleda
