On Sun, Apr 21, 2019 at 08:53:11PM +0200, Alexander Bluhm wrote:
> I wonder whether SIZE_T_MAX would be better than -1. But they seem
> to be equivalent.
I prefer SIZE_T_MAX as well.
> > Index: min_heap.h
> > ===================================================================
> > RCS file: /cvs/src/lib/libevent/min_heap.h,v
> > retrieving revision 1.5
> > diff -u -p -r1.5 min_heap.h
> > --- min_heap.h 20 Apr 2019 23:22:28 -0000 1.5
> > +++ min_heap.h 20 Apr 2019 23:28:09 -0000
> > @@ -33,7 +33,7 @@
> >
> > typedef struct min_heap {
> > struct event **p;
> > - unsigned n, a;
> > + size_t n, a;
> > } min_heap_t;
Changing n means that at least timeout_correct in event.c must be
adjusted as well, because it is used as an unsigned:
static void
timeout_correct(struct event_base *base, struct timeval *tv)
{
unsigned int size;
...
size = base->timeheap.n;
for (; size-- > 0; ++pev) {
struct timeval *ev_tv = &(**pev).ev_timeout;
timersub(ev_tv, &off, ev_tv);
}
...
}
Looking only at min_heap.h it is okay, but takes a bit polishing to be
applicable for all files in libevent.
I haven't inspected all occurrences, so no updated patch yet.
Tobias
