Fix #2023 by handling "empty Callable" case in Flux#collectList

[bsideup]

It seems that we were handling Callable (but not Fuseable.ScalarCallable) incorrectly
and, instead of producing an empty list, were returning an empty Mono instead.

The change also simplifies handling of Fuseable.ScalarCallable

[codecov-io]

Codecov Report

Merging #2024 into 3.2.x will increase coverage by 0.01%.
The diff coverage is 100%.

@@             Coverage Diff              @@
##              3.2.x    #2024      +/-   ##
============================================
+ Coverage     84.61%   84.62%   +0.01%     
- Complexity     3966     3968       +2     
============================================
  Files           364      364              
  Lines         30177    30178       +1     
  Branches       5611     5612       +1     
============================================
+ Hits          25533    25537       +4     
+ Misses         3037     3035       -2     
+ Partials       1607     1606       -1

Impacted Files	Coverage Δ	Complexity Δ
...ore/src/main/java/reactor/core/publisher/Flux.java	98.63% <100%> (ø)	517 <0> (+1)	⬆️
...actor/core/publisher/FluxOnBackpressureLatest.java	80.61% <0%> (-3.07%)	2% <0%> (ø)
...in/java/reactor/core/publisher/FluxWindowWhen.java	80.28% <0%> (-1.45%)	2% <0%> (ø)
...ava/reactor/core/publisher/WorkQueueProcessor.java	69.42% <0%> (-0.83%)	19% <0%> (ø)
...ava/reactor/core/publisher/EventLoopProcessor.java	79.79% <0%> (-0.52%)	51% <0%> (-1%)
...ain/java/reactor/core/publisher/FluxPublishOn.java	86.39% <0%> (-0.21%)	6% <0%> (ø)
...c/main/java/reactor/core/publisher/FluxReplay.java	84.78% <0%> (+0.15%)	25% <0%> (ø)	⬇️
...rc/main/java/reactor/core/publisher/Operators.java	81.4% <0%> (+0.42%)	125% <0%> (ø)	⬇️
...in/java/reactor/core/publisher/TopicProcessor.java	74.85% <0%> (+0.59%)	16% <0%> (ø)	⬇️
...main/java/reactor/core/publisher/FluxGenerate.java	82.2% <0%> (+0.61%)	7% <0%> (ø)	⬇️
... and 3 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 b05e933...2a44b7f. Read the comment docs.

[simonbasle]

I would suggest to use the form onAssembly(new MonoCallable<>(...)) here and below, so that there is no additional step in the backtrace that corresponds to the fromCallable call.

[reactorbot]

@bsideup this PR seems to have been merged on a maintenance branch, please ensure the change is merge-forwarded to intermediate maintenance branches and up to master 🙇