In my C/C++ code I have an array of object "slots". these slots are initialized with some data.
A "slot manager" awaits a request for a slot and delivers a member of the array to a user.
The user pushes its "instance data" into a "slot" and does some work.
At some point the work is done and the "slot" is returned to the "slot
manager"
OK... so I am thinking about adding scripting as a feature of a
slot. My current idea is that I will have a master LUA state, and each
slot will get a lua thread.
My initial thought was that I would create a new LUA thread when the
slots are initialized, but this would require somehow resetting the
thread's state between slot allocations... so that each slot user gets
a "clean" lua state. I don't see anyway of doing this in the lua API.
I guess the alternative is to remove all references to the thread
when the slot is freed (so that the GC can free the memory) and then
create a new thread when a slot is allocated.
My concern is thrashing the memory allocator. In my application it is
very important to pay careful attention to memory allocations and
fragmentation AND to runtime speed.
Can anyone point me in a direction?
Thanks,
Buzz
