On 21/09/18 18:10 +0200, François Dumont wrote:
Here is the patch complement.
load_factor.cc failure revealed a bug in load factor management. Now
computation of _M_next_resize is consistent throughout the different
places where it is set.
The 2 other tests only have to be adapted.
PR libstdc++/87135
* src/c++11/hashtable_c++0x.cc (_Prime_rehash_policy::_M_next_bkt):
Use __builtin_floor to compute _M_next_resize.
* testsuite/23_containers/unordered_set/hash_policy/71181.cc: Adapt.
* testsuite/23_containers/unordered_set/hash_policy/prime_rehash.cc:
Adapt.
Now fully tested under x86_64.
Ok to commit ?
OK, thanks.
