Patches for 0.59.1

[sklam]

• Cherrypick for 0.59.1
• change log
• skipped PPC test failure for test_issue9490_non_det_ssa_problem

TODO:

• cherrypick db03908 and b39bdf2 for main. Remember to keep note to Bad interaction (hang) involving TBB and subprocess fork and maybe MKL OMP #9501 (comment)
• cherrypick release note update
• cherrypick towncrier removal

[sklam]

BFID: numba_yaml_439

[sklam]

#9480 (comment)

passed. subsequent commits have no code changes.

[sklam]

Waiting for release blocker fix

[sklam]

numba_yaml_440

updated PR

[sklam]

BFID: numba_yaml_441

[sklam]

Note for reviewer:

After cherrypicks and changelog update, the following tests adjustments to test_issue9490_non_det_ssa_problem were made:

• skip LLVM codegen failure on PPC (db03908)

AssertionError: subprocess failed with output:
Invalid PPC CTR loop!
UNREACHABLE executed at /opt/conda/conda-bld/llvmdev_1684265335202/work/llvm/lib/Target/PowerPC/PPCCTRLoops.cpp:179!
Fatal Python error: Aborted

I believe this is purely a LLVM codegen problem.

• On linux-64 py3.9 npy1.25 , the buildfarm consistently fail the test due to a TBB problem (b39bdf2).
  GDB points to:

(gdb) info threads
  Id   Target Id                                    Frame
* 1    Thread 0x7ffff7c39280 (LWP 3954038) "python" __pthread_clockjoin_ex (threadid=140736834631424, thread_return=0x0,
    clockid=<optimized out>, abstime=<optimized out>, block=<optimized out>) at pthread_join_common.c:145
  2    Thread 0x7fffdc49c700 (LWP 3954042) "python" syscall () at ../sysdeps/unix/sysv/linux/x86_64/syscall.S:38
  3    Thread 0x7fffdc09b700 (LWP 3954043) "python" 0x00007ffff7d3f89b in sched_yield () at ../sysdeps/unix/syscall-template.S:78
  4    Thread 0x7fffdbc9a700 (LWP 3954044) "python" 0x00007ffff7d3f89b in sched_yield () at ../sysdeps/unix/syscall-template.S:78
  5    Thread 0x7fffdb899700 (LWP 3954045) "python" 0x00007ffff7d3f89b in sched_yield () at ../sysdeps/unix/syscall-template.S:78
  6    Thread 0x7fffdb498700 (LWP 3954047) "python" syscall () at ../sysdeps/unix/sysv/linux/x86_64/syscall.S:38
  7    Thread 0x7fffdac96700 (LWP 3954048) "python" 0x00007ffff7d3f89b in sched_yield () at ../sysdeps/unix/syscall-template.S:78
  8    Thread 0x7fffdb097700 (LWP 3954046) "python" 0x00007ffff7d3f89b in sched_yield () at ../sysdeps/unix/syscall-template.S:78
  12   Thread 0x7fffd9c92700 (LWP 3954054) "python" 0x00007ffff7d3f89b in sched_yield () at ../sysdeps/unix/syscall-template.S:78
  13   Thread 0x7fffd9490700 (LWP 3954053) "python" 0x00007ffff7d3f89b in sched_yield () at ../sysdeps/unix/syscall-template.S:78
  14   Thread 0x7fffda895700 (LWP 3954050) "python" 0x00007ffff7d3f89b in sched_yield () at ../sysdeps/unix/syscall-template.S:78
  15   Thread 0x7fffd908f700 (LWP 3954055) "python" 0x00007ffff5b4349a in __kmp_acquire_ticket_lock ()
   from /home/siu/miniconda3/envs/testenv_bad/lib/python3.9/site-packages/mkl/../../../libiomp5.so

I believe this is bad interaction between TBB and MKL OMP on a machine that has 72 cpu cores. The resolution is to force the numba code to use only 1-thread.