Gene,

it is not the case that callbacks are not nested. They
frequently are, and the windows spec calls (not by name)
for all callbacks to be re-entrant. Each must have a
unique user area.

The only problems it creates are from user assumptions
that values stored into the user area will be constant
from one invocation to another.

I don't think that there is any alternative, but will
listen if you want to propose one and implement it
as a demonstration.

Rick

>From: "Gene LeFave" <gene_at_tekdata.com>

>I'm sure you gave a lot of thought to the way that call backs are
>handled. Although, I can see the logic of a dummy user task and
>associated stacks being created on the return stack, I think that a
>dedicated user/stack area for call backs might be better.

>From what I know of windows, call backs are never nested. So all of
>the windows call backs could share a user table. Or perhaps you
>could designated a user table to be used for a given call back. I'm
>just kind of brainstorming here but I'm just not confortable with the
>dynamic user table concept. It just seems to me to be a real
>source of potential problems.

.
Received on Thu Apr 15 1999 - 10:29:45 PDT