socketserver: Add an opt-in option to get Python 3.6 behaviour on server_close()

[vstinner]

BPO	33540
Nosy	@pitrou, @vstinner, @tiran, @ned-deily, @miss-islington
PRs	bpo-33540: Add block_on_close attr to socketserver #6911 [3.7] bpo-33540: Add block_on_close attr to socketserver (GH-6911) #7088 [3.7] bpo-33540: Fix socketserver.ThreadingMixIn if block_on_close=False #7309 bpo-33540: Fix socketserver.ThreadingMixIn if block_on_close=False #7310 [3.6] bpo-33540, socketserver: Add _block_on_close for tests #7317

^{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 = <Date 2018-05-24.02:03:27.765>
created_at = <Date 2018-05-16.15:59:33.584>
labels = ['3.8', '3.7', 'library', 'release-blocker']
title = 'socketserver: Add an opt-in option to get Python 3.6 behaviour on server_close()'
updated_at = <Date 2018-06-01.14:24:53.121>
user = 'https://github.com/vstinner'

bugs.python.org fields:

activity = <Date 2018-06-01.14:24:53.121>
actor = 'vstinner'
assignee = 'none'
closed = True
closed_date = <Date 2018-05-24.02:03:27.765>
closer = 'vstinner'
components = ['Library (Lib)']
creation = <Date 2018-05-16.15:59:33.584>
creator = 'vstinner'
dependencies = []
files = []
hgrepos = []
issue_num = 33540
keywords = ['patch']
message_count = 9.0
messages = ['316825', '316833', '317059', '317486', '317494', '317502', '318403', '318404', '318418']
nosy_count = 5.0
nosy_names = ['pitrou', 'vstinner', 'christian.heimes', 'ned.deily', 'miss-islington']
pr_nums = ['6911', '7088', '7309', '7310', '7317']
priority = 'release blocker'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue33540'
versions = ['Python 3.7', 'Python 3.8']

[vstinner]

bpo-31151 changed ForkingMixIn and bpo-31233 changed ThreadingMixIn to block on server_closer() for processes/threads.

I propose to add a new opt-in option to get the Python 3.6 behaviour. The old behaviour was wrong, but I expect that some people rely on it.

[vstinner]

I consider that it's a regression, so I pick the release blocker priority. Sorry Ned!

[ned-deily]

It would be great if someone(s) could give PR 6911 a review, since it is a release blocker for 3.7.0rc1. Thanks!

[vstinner]

New changeset 453bd0b by Victor Stinner in branch 'master':
bpo-33540: Add block_on_close attr to socketserver (GH-6911)
453bd0b

[miss-islington]

New changeset fa286ed by Miss Islington (bot) in branch '3.7':
bpo-33540: Add block_on_close attr to socketserver (GH-6911)
fa286ed

[vstinner]

I added the block_on_close class attribute to Python 3.7 and master. I'm not really proud of it, but it's the price of the backward compatibility!

[vstinner]

New changeset 29ae9dc by Victor Stinner in branch '3.7':
bpo-33540: Fix socketserver.ThreadingMixIn if block_on_close=False (GH-7309)
29ae9dc

[vstinner]

New changeset 829fcd0 by Victor Stinner in branch 'master':
bpo-33540: Fix socketserver.ThreadingMixIn if block_on_close=False (GH-7310)
829fcd0

[vstinner]

New changeset 1381bfe by Victor Stinner in branch '3.6':
bpo-33540, socketserver: Add _block_on_close for tests (GH-7317)
1381bfe