Re: filling area by color atack safety

Michael S <already5chosen@yahoo.com> writes:

[..various fill algorithms and how they scale..]

>
One thing that I were not expecting at this bigger pictures, is good
performance of simple recursive algorithm.  Of course, not of original
form of it, but of variation with explicit stack.
For many shapes it has quite large memory footprint and despite that it
is not slow.  Probably the stack has very good locality of reference.
>
[algorithm]

You are indeed a very clever fellow.  I'm impressed.

Intrigued by your idea, I wrote something along the same lines,
only shorter and (at least for me) a little easier to grok.
If someone is interested I can post it.

I see you have also done a revised algorithm based on the same
idea, but more elaborate (to save on runtime footprint?).
Still working on formulating a response to that one...