[doc] multiprocessing.Queue fails intermittently with "Broken pipe"

[SamuelGrayson]

BPO	38561
Nosy	@iritkatriel
Files	test.py

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = None
created_at = <Date 2019-10-23.01:40:48.912>
labels = ['type-bug', '3.9', '3.10', '3.11', 'library', 'docs']
title = '[doc] multiprocessing.Queue fails intermittently with "Broken pipe"'
updated_at = <Date 2022-01-19.10:16:13.205>
user = 'https://bugs.python.org/SamuelGrayson'

bugs.python.org fields:

activity = <Date 2022-01-19.10:16:13.205>
actor = 'iritkatriel'
assignee = 'docs@python'
closed = False
closed_date = None
closer = None
components = ['Documentation', 'Library (Lib)']
creation = <Date 2019-10-23.01:40:48.912>
creator = 'Samuel Grayson'
dependencies = []
files = ['48674']
hgrepos = []
issue_num = 38561
keywords = []
message_count = 2.0
messages = ['355204', '410932']
nosy_count = 3.0
nosy_names = ['docs@python', 'Samuel Grayson', 'iritkatriel']
pr_nums = []
priority = 'normal'
resolution = None
stage = None
status = 'open'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue38561'
versions = ['Python 3.9', 'Python 3.10', 'Python 3.11']

[SamuelGrayson]

See this SO post for more details

[iritkatriel]

I don't think you're supposed to close the queue more than once. The documentation can be improved on that. Currently the documentation of close is:

close()
Indicate that no more data will be put on this queue by the current process. The background thread will quit once it has flushed all buffered data to the pipe. This is called automatically when the queue is garbage collected.

https://docs.python.org/3/library/multiprocessing.html#multiprocessing.Queue.close

That could be read to imply to each process needs to indicate when it is done.

[slateny]

Related to #80025 (and possibly the referenced #91185 and #31913)

[geryogam]

@JelleZijlstra, could you also close this issue now that it has been fixed in PR #31913?