Fix set_threadpool_size(n) for n == 1

[texodus]

• Fixes a deadlock condition when set_threadpool_size(1) was called in Python.
• Add documentation to set_threadpool_size() to note that it must be called before the global thread pool is instantiated.
• Removed sloppy parallel_for() API in C++. Confirmed that this generates nearly identical bytecode in wasm/release, and confirmed no benchmark regressions in single- or mult- threaded runtimes.
• Cherrypick Add a small tweak to gnode to allow for building python code without PSP_PARALLEL_FOR #1851

[timkpaine]

the python test failure looks unrelated, could be a dependency version issue

[timkpaine]

the wasm test is that chromedriver/perspective-viewer timeout issue again!

Set timeout to 60 seconds...
Executing bindgen...                              
                                                  
Running headless tests in Chrome on `[http://127.0.0.1:37831/`](http://127.0.0.1:37831/%60)
Try find `webdriver.json` for configure browser's capabilities:
Not found
Starting new webdriver session...                 
Visiting http://127.0.0.1:46747...                
Loading page elements...                          
Waiting for test to finish...                     
                                                  
Failed to detect test as having been run. It might have timed out.
output div contained:
    Loading scripts...

                                                  
driver status: signal: 9
driver stdout:
    Starting ChromeDriver 102.0.5005.61 (0e59bcc00cc4985ce39ad31c150065f159d95ad3-refs/branch-heads/5005@{#819}) on port 37831
    Only local connections are allowed.
    Please see https://chromedriver.chromium.org/security-considerations for suggestions on keeping ChromeDriver safe.
    ChromeDriver was started successfully.

driver stderr:
    [1655104268.249][SEVERE]: Timed out receiving message from renderer: 299.698
    [1655104268.252][SEVERE]: Timed out receiving message from renderer: 299.698

[texodus]

The python failure is due to int literal 32/64 discrepancy on windows

The rust failure is endemic to wasm_bindgen_test - I don't have a fix but we can remove these tests if they are causing more harm than good.