Python hangs fetching data with urllib in parallel

```
import time
from urllib import request
from queue import Queue
from threading import Thread

slots = 4

def dl(q):
while True:
url = q.get()
if isinstance(url, int):
time.sleep(1)
print(url)
else:
print("Fetching...")
print(f"{url} size is: {len(request.urlopen(request.Request(url)).read())}")
print("SUCCESS")
q.task_done()

q = Queue()
for i in range(slots):
t = Thread(target=dl, args=(q,))
t.daemon = True
t.start()

while True:
for n in range(4):
if n == 0:
q.put("https://example.com/")
elif n == 2:
q.put("https://example.net/")
else:
q.put(n)
q.join()
input("Enter to continue")
```

There are data fetching and 1s timers arranged in odd/even pattern. Success with slots = 1 (no parallel), success with slots = 2 if data fetch can complete before 1s timer.

Resorting to Requests same result.

Start by running the supplied Python reproducer in ish with different slot counts, then compare the urllib and Requests cases described in the report. Done means parallel data fetching no longer hangs when timer and fetch tasks are interleaved.

Written by the indexing model from the issue text.