Remove unneeded ZipFile IO

[data-ux]

BPO	45505
Nosy	@data-ux
PRs	bpo-45505: ZipFile file header optimization #25966

^{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 2021-10-18.05:30:14.320>
labels = ['library', 'performance']
title = 'Remove unneeded ZipFile IO'
updated_at = <Date 2021-10-18.05:30:14.320>
user = 'https://github.com/data-ux'

bugs.python.org fields:

activity = <Date 2021-10-18.05:30:14.320>
actor = 'data-ux'
assignee = 'none'
closed = False
closed_date = None
closer = None
components = ['Library (Lib)']
creation = <Date 2021-10-18.05:30:14.320>
creator = 'data-ux'
dependencies = []
files = []
hgrepos = []
issue_num = 45505
keywords = []
message_count = 1.0
messages = ['404155']
nosy_count = 1.0
nosy_names = ['data-ux']
pr_nums = ['25966']
priority = 'normal'
resolution = None
stage = None
status = 'open'
superseder = None
type = 'performance'
url = 'https://bugs.python.org/issue45505'
versions = []

[data-ux]

Currently in the ZipFile class implementation, when processing the zip file headers "extra" field, a .read() call is used without using the returned value in any way. This call could be replaced with a .seek() to avoid actually doing the IO.

The change would improve performance in use cases where the fileobject given to ZipFile is backed by high latency IO, eg. HTTP range requests.