C implementation of parts of copy.deepcopy

[villemoes]

BPO	28607
Nosy	@Villemoes, @pablogsal
Files	deepcopy.patch deepcopy.patch: patch v2

^{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 2016-11-03.23:11:36.119>
labels = ['extension-modules', '3.7', 'performance']
title = 'C implementation of parts of copy.deepcopy'
updated_at = <Date 2019-01-20.21:43:41.055>
user = 'https://github.com/villemoes'

bugs.python.org fields:

activity = <Date 2019-01-20.21:43:41.055>
actor = 'pablogsal'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Extension Modules']
creation = <Date 2016-11-03.23:11:36.119>
creator = 'villemoes'
dependencies = []
files = ['45344', '45400']
hgrepos = []
issue_num = 28607
keywords = ['patch']
message_count = 3.0
messages = ['280032', '280360', '334105']
nosy_count = 2.0
nosy_names = ['villemoes', 'pablogsal']
pr_nums = []
priority = 'normal'
resolution = None
stage = 'test needed'
status = 'open'
superseder = None
type = 'performance'
url = 'https://bugs.python.org/issue28607'
versions = ['Python 3.7']

• PR: gh-72793: C implementation of parts of copy.deepcopy #91610

[villemoes]

This is mostly an RFC patch. It compiles and passes the test suite. A somewhat silly microbenchmark such as

./python -m timeit -s 'import copy; x = dict([(str(x), x) for x in range(10000)]);' 'copy.deepcopy(x)'

runs about 30x faster. In the (2.7 only) application which motivated this, the part of its initialization that does a lot of deepcopying drops from 11s to 3s. That it's so much less is presumably because the application holds on to the deepcopies, so there's much more allocation going on than in the microbenchmark, but I haven't investigated thoroughly. In any case, a 3.5x speedup is also nice.

[villemoes]

New version, addressing (hopefully) all review comments.

[pablogsal]

Hy Rasmus,

would you like to make a PR with your patch and adding some tests?

Thanks!

[peendebak]

@pablogsal The issue has been stale for a long time, but I created a PR and provided benchmarks.

The functionality of deepcopy has not changed, so I am not sure whether more tests are required. If so, can you let me know which specific tests you had in mind?