New function: default_thread_pool_shutdown()

[lgritz]

This cleanly terminates all worker threads of the default thread pool
by resizing the pool to 0 threads... but only if the pool has been
created in the first place. Note that the alternate idiom

default_thread_pool()->resize(0);

seems like it should accomplish the same thing, but in fact will start
by creating a pool and its threads, if it has not done so before, only
to immediately terminate them.

[fpsunflower]

I'm not sure I follow the logic.

If the thread pool already exists, it will stop all the threads. But if the thread pool hasn't been created yet, it isn't going to prevent them from being created in the future.

So someone might call default_thread_pool_shutdown() only to see the default thread pool get created in the future.

Am I missing something?

[lgritz]

It's intended to be called at shutdown. Hence the name, default_thread_pool_shutdown().

The intent is to solve a problem that occurs for some apps on Windows, where apparently the app upon terminating can kill the threads before calling the static destructors (like that of the thread pool), and if one of those threads held a mutex, it can cause a deadlock.

I don't 100% understand the situation of what set of circumstances was a concern, but this was a request from a software vendor and it seemed reasonable. The problem is "solved" for them by calling default_thread_pool->resize(0), but they wanted a way to avoid that idiom's tendency to create a bunch of threads for the case where the pool had never been created in the first place.

[lgritz]

Pause this, I'm still haggling over the technically correct solution with the requestors...

[lgritz]

Abandoning this, it seems not to be needed at this time.

[lgritz]

This look ok to you, @LazyDodo ?

[LazyDodo]

LGTM!