Shlomi Fish wrote:
>
> Hi!
>
> I downloaded the most up to date version of kpat from anoncvs.kde.org and
> noticed that the method used is Hard DFS. Since kpat halts and resumes the
> solve a lot, I suggest Soft-DFS will be used instead, now that it has a
> stable, hopefully memory-leak free code.
>
> Hard DFS uses procedural recursion to perform a Depth-First Search scan.
> Thus, it requires O(d) time to halt and resume. Soft-DFS on the other
> hand, has its own dedicated stacks, and thus requires O(1) time. So, it
> makes for a better option, and it is possible that it is even faster then
> Hard-DFS, without being interrupted. From my impression it seemed roughly
> the same speed, but I have yet to do a full benchmark.
>
> Stephan, please take note.
>
As you might remember I had quite some crashes with soft DFS, so I never
really
switched away. As you know KDE (and kpat) are short before a release and
so I
didn't want to risk something for the sake of speed.
Greetings, Stephan
--
People in cars cause accidents. Accidents in cars cause people.
Received on Tue Feb 20 2001 - 06:02:28 IST