NB. I do not like the name customized_pool_executor and I think we should rename it early - before it starts to get used in examples and by users in the wider community

Instead of customized_pool_executor I suggest

1. custom_pool_executor (better English in my view)
2. pool_executor (since this is what it is, if existing classes use this name, we should adjust them accordingly)
3. please suggest something.

pool_executor is fine.

ok. I will merge this commit (once approved) and then do a new one with the new name of pool_executorin the next day or two.

Let's change the name right away. If it's a matter of a day or two - no problem.

OK. executor renamed. I've got several other PR's on the way that depend on this, so I'm keen to get this merged ASAP. (= before tutorial!)

Let's merge this as soon as the tests have cycled.

hmm. The appveyor test failed for this, but I cannot find anything wrong.

For some reason, hello_world prints:

[00:41:36] python build\bin\hpxrun.py -l2 -t2 build\Debug\bin\hello_world.exe -- --hpx:bind=none
[00:42:33] hello world from OS-thread 1 on locality 0
[00:42:33] hello world from OS-thread 1 on locality 0
[00:42:33] hello world from OS-thread 0 on locality 0
[00:42:34] hello world from OS-thread 0 on locality 1
[00:42:34] hello world from OS-thread 1 on locality 1

(which is clearly wrong) and hangs afterwards. I'm not sure if this is caused by your changes, though.