3 Jul
2017
3 Jul
'17
12:17 p.m.
On 03.07.2017 14:06, François Kawala wrote:
Another related question, why it is necessary to copy the reached array ?
Because ownership issues. The function works with a std::vector<> internally, because numpy's ndarrays cannot be dynamically increased. We could return a wrapped std::vector (called Vector_size_t internally in graph-tool), but these are not documented. So instead, the std::vector's contents are copied as a ndarray. -- Tiago de Paula Peixoto <tiago@skewed.de>