Pool.new(capacity, factory?, reset?) → poolRequire scripts.s3.pool. A pool keeps reusable objects ready for short bursts of work. It is an optimization for measured allocation pressure, not a replacement for ordinary tables.
Construction
capacity must be a positive integer. Construction immediately creates that many objects. The default factory creates empty tables; the default reset clears their keys in place.
A custom factory takes no arguments and must return a non-nil value. A custom reset receives the released object. Non-table objects require a custom reset.
local Pool = require 'scripts.s3.pool'
local vectors = Pool.new(4, function()
return {
x = 0,
y = 0,
}
end, function(vector)
vector.x = 0
vector.y = 0
end)
local vector = vectors:acquire()
vector.x = 12
print(vector.x)
vectors:release(vector)Members
| Call | Result / behavior |
|---|---|
pool:acquire() | Checks out an object; allocates a new one if none are available. |
pool:release(object) | Resets it, then retains it if there is room. Excess objects are dropped. |
pool:available() | Number of idle objects. |
pool:capacity() | Maximum idle objects retained, not a limit on concurrent acquisitions. |
pool:allocated() | Current pool-created objects: idle plus checked out. |
pool:stats() | Four return values: available, in_use, allocated, capacity; no result table. |
There is no ownership or double-release check. Release each acquired object exactly once to the pool it came from. Double release can cause two callers to receive the same object; foreign releases also invalidate accounting. Never release nil.
See Pooling and Signals for synchronous payload reuse and error-safe release.