[Click] [PATCH] usermode Router::run_selects() not robust under handler deletion

[Peter Swain]

On Fri, Oct 31, 2003 at 03:08:57PM -0800, Eddie Kohler wrote:
> I'm not sure I really understood the problem -- why not just fix up the 'p'
> pointer, as attached?

Much better, Eddie.
It minimises change, keeps the more compact pointer notation,
but renormalizes p after potential damage.

I was agonizing over a complete fix which coped with handlers doing
arbitrary remove_select() calls but not missing events. All we're going
to see in practice is removing their own entries, and it's only an
optimization anyway.

Thanx

^..^
(oo)