Date: Mon, 18 Apr 2016 09:25:40 -0400 From: Ryan Stone <rysto32@gmail.com> To: Hans Petter Selasky <hps@selasky.org> Cc: Aleksander Alekseev <afiskon@devzen.ru>, FreeBSD Current <freebsd-current@freebsd.org> Subject: Re: qsort() documentation Message-ID: <CAFMmRNx9ETbT9M8L=CBYk6VPWQYu39e5Mh-QCHPj_6K8Hs32kA@mail.gmail.com> In-Reply-To: <5714DC98.7090208@selasky.org> References: <5714C86A.8050204@selasky.org> <20160418151639.634d571d@fujitsu> <5714DC98.7090208@selasky.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Apr 18, 2016 at 9:09 AM, Hans Petter Selasky <hps@selasky.org> wrote > I think the algorithm is switching to mergesort. I'll look up the paper > and add that correctly before commit. > No, it switches to insertion sort, assuming that it's acting on an already sorted array. If that assumption is wrong you still get O(n**2) complexity.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAFMmRNx9ETbT9M8L=CBYk6VPWQYu39e5Mh-QCHPj_6K8Hs32kA>