PR16585: Fix setting/resetting of current_function_has_computed_jump

[Steven Bosscher]

On Tuesday 26 October 2004 00:13, Richard Henderson wrote:
> On Mon, Oct 25, 2004 at 11:36:23PM +0200, Steven Bosscher wrote:
> > Does this approach look sane, or should I perhaps move this code
> > to cfgcleanup.c?
>
> This doesn't look like the best fix.  I think that at minimum,
> the _clearing_ of current_function_has_computed_jump should be
> moved somewhere else.  It might be best to still set it here,
> but I'm not sure.

Yeah, I thought it was an odd place to have this, too.  In fact
I was thinking perhaps we should just not know about this
current_function_has_computed_jump thing until much later in
the compilation.  The only pass that actually cares about this
is sched-rgn.

Gr.
Steven