fix: Memory leak

[ioworker0]

Fixes #113

[codecov]

Codecov Report

Merging #114 into master will decrease coverage by 1.90%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #114      +/-   ##
==========================================
- Coverage   95.84%   93.94%   -1.91%     
==========================================
  Files           9        9              
  Lines         554      380     -174     
==========================================
- Hits          531      357     -174     
+ Misses         16       14       -2     
- Partials        7        9       +2     

Impacted Files	Coverage Δ
pool.go	91.34% <100.00%> (-3.93%)	⬇️
pool_func.go	93.04% <100.00%> (-0.86%)	⬇️
ants.go	80.00% <0.00%> (-5.00%)	⬇️
worker_loop_queue.go	92.72% <0.00%> (-3.48%)	⬇️
worker.go	100.00% <0.00%> (ø)
options.go	100.00% <0.00%> (ø)
worker_func.go	100.00% <0.00%> (ø)
worker_array.go	100.00% <0.00%> (ø)
... and 1 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ef60172...00691c6. Read the comment docs.

[panjf2000]

Also please resolve this one:

ants/pool_func.go

Lines 274 to 286 in ad86bfa

	func (p *PoolWithFunc) revertWorker(worker *goWorkerWithFunc) bool {
	if atomic.LoadInt32(&p.state) == CLOSED || p.Running() > p.Cap() {
	return false
	}
	worker.recycleTime = time.Now()
	p.lock.Lock()
	p.workers = append(p.workers, worker)
	// Notify the invoker stuck in 'retrieveWorker()' of there is an available worker in the worker queue.
	p.cond.Signal()
	p.lock.Unlock()
	return true
	}

as well.

[ioworker0]

Also please resolve this one:

ants/pool_func.go

Lines 274 to 286 in ad86bfa

	func (p *PoolWithFunc) revertWorker(worker *goWorkerWithFunc) bool {
	if atomic.LoadInt32(&p.state) == CLOSED || p.Running() > p.Cap() {
	return false
	}
	worker.recycleTime = time.Now()
	p.lock.Lock()
	p.workers = append(p.workers, worker)
	// Notify the invoker stuck in 'retrieveWorker()' of there is an available worker in the worker queue.
	p.cond.Signal()
	p.lock.Unlock()
	return true
	}

as well.

Sure, no problem

[panjf2000]

I believe that this bugfix is intuitive enough, no need to add unit tests, please remove it.

[ioworker0]

I believe that this bugfix is intuitive enough, no need to add unit tests, please remove it.

Gotcha😃

[panjf2000]

Thanks~

[ioworker0]

Thanks~

@panjf2000
Bro, when will you consider adding a new tag？😃

[coticom]

I have a point that I don't quite understand: After the double check, isn't it still possible that the pool might be closed? So, what's the purpose of this double check?

[OliverChao]

@coticom If the pool has been closed, revertWorker will never insert an un-released worker because of double check.

[ioworker0]

I have a point that I don't quite understand: After the double check, isn't it still possible that the pool might be closed? So, what's the purpose of this double check?

IIUC, the first one looks more like is a best-effort early check. But it's
unreliable without the lock, so we also require the second one.

IOW, I don't believe that the first one is necessary in common cases :)