Re: [glob2-devel] Todo List For Beta 4

[Kai Antweiler]

> although we might tweak them for 1 wide passages.

The easiest way that I can think of:

Calculate a gradient, that does not propagate through 1 wide
passages. Also store those fields, were the gradient did stop to
propagate due to this special rule. Then the real gradient can be
calculated quickly, because it only improves values.

You can compare the distances according to those two
gradients. If the difference of the distances is larger than
x, use the normal gradient. Otherwise use that special
gradient.

x could depend on the number of globs, that are currently using the
special gradient - i.e: that are heading for 1
wide passages.
-- 
Kai Antweiler