[PEAK] dispatch -- chaining decorators

[Phillip J. Eby]

At 11:29 PM 12/9/2005 +0100, Simon Belak wrote:

>Will next_method always be the first argument or is it better to check 
>throughout?

It currently has to be first.  There is a possibility in the future that 
might change, if I figure out a good way to incorporate bind variables in 
rules.  Even then, however, it's likely that it would be the second argument.


>You have put my mind to rest.
>Comparing to the last branch not using generic functions extensively, the 
>size has gone down 23% and number of code lines for ~10% while introducing 
>additional functionality. So I guess I'm doing fine.

Only 10% reduction?  I'm kind of surprised it's so little improvement.  :)