> Could you please sent mail in plain text? Thanks.

I use a free mail, so ... but I've changed my options. I hope so ^^

> *grin* Do you really want to know?

Just want! Apache module made me go to gory apache sources 8 months ago, and apache made me be here a month ago. Newbie ^^;

> Well, if it only has one comment, it should be very clear ;).

oot!


> IOW, take the number of free bytes, round it up to the next
> multiple of BOUNDARY_SIZE and then devide it by BOUNDARY_SIZE.
> This gives us an index which we can use to compare free amounts
> of memory in a node by multiples of BOUNDARY_SIZE. The
> advantage is that nodes will be grouped together and the
> priority queue has to be reshuffled less.

I got it.

> The part of the code you are looking at checks if the new node
> has more memory free (after the allocation) than the currently
> active node. If it does, the new node is made the active node.
>
> The node with less free memory is moved back in the chain, so
> that the node with the most memory available is at the front
> of the list.

I got it.

> This scheme is a way to use memory a bit more optimal than the
> original pools implementation, in a cheap way.

I will ask why it is some times later.

With your help, I am drawing a linked list(node) and printf-ing between lines.

Thanks! Sander

Reply via email to