On Wed, 20 Nov 2019, Jan Hubicka wrote:
> Hi,
> sadly this patch seems to trigger false positive
> ../../gcc/vec.h:311:10: error: attempt to free a non-heap object ???@Xa???@Y
> [-Werror=free-nonheap-object]
>
> which is about vec destructor freeing heap allocated memory if the
> vector was ever resized from hits stack allocated place. In this case
> it never is.
>
> I am testing the following workaround which turns it into simple array
> (only benefit of auto_vec here is the bounds check that I implemented by
> hand).
>
> I will commit it once x86_64-linux boostrap finishes unless someone
> comes with better alternative. Sorry for this - it did not show up with
> LTO build for me.
I've committed a workaround for a similar issue to vec.h, so you
might want to re-check.
Richard.
> Honza
>
> * fibonacci_heap.h (fibonacci_heap<K,V>::consolidate): Turn auto_vec
> to ordinary array.
>
> Index: fibonacci_heap.h
> ===================================================================
> --- fibonacci_heap.h (revision 278501)
> +++ fibonacci_heap.h (working copy)
> @@ -648,17 +648,18 @@ template<class K, class V>
> void fibonacci_heap<K,V>::consolidate ()
> {
> const int D = 1 + 8 * sizeof (long);
> - auto_vec<fibonacci_node<K,V> *, D> a;
> + fibonacci_node<K,V> *a[D];
> fibonacci_node<K,V> *w, *x, *y;
> int i, d;
>
> - a.quick_grow_cleared (D);
> + memset (a, 0, sizeof (a));
>
> while ((w = m_root) != NULL)
> {
> x = w;
> remove_root (w);
> d = x->m_degree;
> + gcc_checking_assert (d < D);
> while (a[d] != NULL)
> {
> y = a[d];
>
--
Richard Biener <rguent...@suse.de>
SUSE Software Solutions Germany GmbH, Maxfeldstrasse 5, 90409 Nuernberg,
Germany; GF: Felix Imendörffer; HRB 36809 (AG Nuernberg)
