logging.StreamHandler produces BrokenPipeError when cut short by head utility

Bug report

Bug description:

The following code:

# test.py
import sys
import logging

logger = logging.getLogger(__name__)
logger.setLevel(logging.DEBUG)
logger.addHandler(logging.StreamHandler(sys.stdout))
logger.debug('foo')
logger.debug('bar')

when piped to head -n1, produces:

# python test.py | head -n1
foo
--- Logging error ---
Traceback (most recent call last):
  File "/src/cpython/Lib/logging/__init__.py", line 1155, in emit
    self.flush()
    ~~~~~~~~~~^^
  File "/src/cpython/Lib/logging/__init__.py", line 1137, in flush
    self.stream.flush()
    ~~~~~~~~~~~~~~~~~^^
BrokenPipeError: [Errno 32] Broken pipe
Call stack:
  File "test.py", line 8, in <module>
    logger.debug('bar')
Message: 'bar'
Arguments: ()
Exception ignored while flushing sys.stdout:
BrokenPipeError: [Errno 32] Broken pipe

This is because logging.StreamHandler.emit does not handle BrokenPipeError when flushing the stream.
I will fix this with the approach recommended by Note on SIGPIPE.

CPython versions tested on:

CPython main branch, 3.13

Operating systems tested on:

Linux

Linked PRs

• gh-134569

Le chemin signalé est Lib/logging/init.py, en particulier StreamHandler.emit et flush ; reproduisez le problème avec test.py redirigé vers head -n1 sous Linux. Comparez le correctif proposé dans la PR liée gh-134569 et ses tests ; le travail est terminé lorsque le pipeline court n’émet plus le traceback de logging ni le BrokenPipeError final.