On 11/16/13 3:29 AM, Ivan Kazmenko wrote:
On Friday, 15 November 2013 at 21:46:26 UTC, Vladimir Panteleev wrote:
I've been investigating an instance that I ran into while processing
some data, in which multiSort had apparently stalled whereas two
successive sort calls processed the data quickly. After reducing the
code and 20 million points, I've reduced the problem to this test case:
I've been hit by this case too, half a year ago. See the relevant
discussion on D.learn:
http://forum.dlang.org/thread/[email protected]
The fact that the issue is raised once again perhaps means it is worth a
bugreport. What do you guys think?
Ivan Kazmenko.
I think a median of five pivot selection is in order.
Andrei
