Threads from _log_start_if_long might not be reaped

[inducer]

@benSepanski showed me a long with zillions of them live at one point.

Here's what that log looked like:

Thread 0x00007f6535c23700 (most recent call first):
  File "/home/bensepan/firedrake/lib/python3.9/site-packages/pytools/__init__.py", line 2424 in _log_start_if_long
  File "/usr/lib/python3.9/threading.py", line 892 in run
  File "/usr/lib/python3.9/threading.py", line 954 in _bootstrap_inner
  File "/usr/lib/python3.9/threading.py", line 912 in _bootstrap

Thread 0x00007f6331ffb700 (most recent call first):
  File "/home/bensepan/firedrake/lib/python3.9/site-packages/pytools/__init__.py", line 2424 in _log_start_if_long
  File "/usr/lib/python3.9/threading.py", line 892 in run
  File "/usr/lib/python3.9/threading.py", line 954 in _bootstrap_inner
  File "/usr/lib/python3.9/threading.py", line 912 in _bootstrap

Thread 0x00007f63327fc700 (most recent call first):
  File "/home/bensepan/firedrake/lib/python3.9/site-packages/pytools/__init__.py", line 2424 in _log_start_if_long
  File "/usr/lib/python3.9/threading.py", line 892 in run
  File "/usr/lib/python3.9/threading.py", line 954 in _bootstrap_inner
  File "/usr/lib/python3.9/threading.py", line 912 in _bootstrap

Thread 0x00007f6332ffd700 (most recent call first):
  File "/home/bensepan/firedrake/lib/python3.9/site-packages/pytools/__init__.py", line 2424 in _log_start_if_long
  File "/usr/lib/python3.9/threading.py", line 892 in run
  File "/usr/lib/python3.9/threading.py", line 954 in _bootstrap_inner
  File "/usr/lib/python3.9/threading.py", line 912 in _bootstrap

Thread 0x00007f63337fe700 (most recent call first):
  File "/home/bensepan/firedrake/lib/python3.9/site-packages/pytools/__init__.py", line 2424 in _log_start_if_long
  File "/usr/lib/python3.9/threading.py", line 892 in run
  File "/usr/lib/python3.9/threading.py", line 954 in _bootstrap_inner
  File "/usr/lib/python3.9/threading.py", line 912 in _bootstrap

Thread 0x00007f6537426700 (most recent call first):
  File "/home/bensepan/firedrake/lib/python3.9/site-packages/pytools/__init__.py", line 2424 in _log_start_if_long
  File "/usr/lib/python3.9/threading.py", line 892 in run
  File "/usr/lib/python3.9/threading.py", line 954 in _bootstrap_inner
  File "/usr/lib/python3.9/threading.py", line 912 in _bootstrap

(on a presumably unrelated crash in PETSc/Cython)

[benSepanski]

I've now figured out the bug, it was just a few python errors (type mismatch, forgetting to put self. in front of an attribute) inside of a firedrake matrix-free PETSc matrix