Update zipimport to support importlib.invalidate_caches()

[brettcannon]

BPO	14678
Nosy	@brettcannon, @gpshead, @ncoghlan, @ericvsmith, @merwok, @ericsnowcurrently, @serhiy-storchaka, @crwilcox, @miss-islington, @desmondcheongzx
PRs	bpo-14678: Update zipimport to support importlib.invalidate_caches() #24159

^{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 2021-03-08.20:06:43.292>
created_at = <Date 2012-04-26.17:41:49.735>
labels = ['easy', 'type-feature', 'library', '3.10']
title = 'Update zipimport to support importlib.invalidate_caches()'
updated_at = <Date 2021-03-08.20:06:46.608>
user = 'https://github.com/brettcannon'

bugs.python.org fields:

activity = <Date 2021-03-08.20:06:46.608>
actor = 'brett.cannon'
assignee = 'none'
closed = True
closed_date = <Date 2021-03-08.20:06:43.292>
closer = 'brett.cannon'
components = ['Library (Lib)']
creation = <Date 2012-04-26.17:41:49.735>
creator = 'brett.cannon'
dependencies = []
files = []
hgrepos = []
issue_num = 14678
keywords = ['patch', 'easy']
message_count = 12.0
messages = ['159404', '159406', '159948', '159979', '160022', '197551', '325715', '361711', '362842', '362889', '362905', '388294']
nosy_count = 11.0
nosy_names = ['brett.cannon', 'gregory.p.smith', 'ncoghlan', 'eric.smith', 'eric.araujo', 'eric.snow', 'serhiy.storchaka', 'superluser', 'crwilcox', 'miss-islington', 'desmondcheongzx']
pr_nums = ['24159']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'enhancement'
url = 'https://bugs.python.org/issue14678'
versions = ['Python 3.10']

[brettcannon]

zipimport's finders that get cached in sys.path_importer_cache should probably be updated to support importlib.invalidate_caches() (else the module should get re-implemented in pure Python in importlib and then be tossed).

[merwok]

zipimport in Python sounds good to me.

[brettcannon]

I should mention I have a version from importers that is probably out-of-date: http://code.google.com/p/importers/source/browse/importers/zip.py

[ncoghlan]

Moving zipimporter to Python code is harder than it sounds: we don't want to break the ability to ship the standard library itself inside a zipfile.

If you try to move zipimporter to pure Python, you could easily end up with a *very* ugly bootstrapping problem, on par with that already encountered when hacking on importlib._bootstrap.

In fact, the path of least resistance here might actually be to implement zipimporter directly *in* importlib._bootstrap.

[brettcannon]

The real problem becomes the issue of what zipfile depends on, which complicates bootstrapping. I mean things could go as far as to write a script that takes in anchor points in the stdlib and freezes all code they depend on, but that seems like potential overkill and executable bloat.

But something needs to happen as zipimporter has major shortcomings thanks to to its re-implementation of zip handling (e.g. no ZIP64 support, etc.). Plus no one ever wants to touch that code.

[gpshead]

Brett wrote a pure python zipimporter in http://bugs.python.org/issue17630 :)

FWIW, the zipimport.c implementation (in 2.7) causes us serious pain when we've got the stdlib in a .zip file and need to update that while there are running python processes (which have cached the central directory of that .zip file but do not also keep a fd to the zip file open or even attempt to stat the file to see if it has changed [racy but it'd still be better than the nothing that is done now]).

[serhiy-storchaka]

zipimport have been rewritten in pure Python (bpo-25711).

[crwilcox]

What work remains to be done for this issue?

[brettcannon]

@crwilcox Don't know what's left. That's probably part of the work that's left. ;) This issue is so old that it's possibly out-of-date since zipimport got rewritten in pure Python.

[crwilcox]

That is my thinking as well after rooting around a bit. I unfortunately am not knowledgable enough here to be 100% sure it is complete.

[brettcannon]

It looks like it hasn't been done, else https://github.com/python/cpython/blob/master/Lib/zipimport.py would have some definition of a invalidate_caches() method. And specifically the finder lacks that method so it can be called by

cpython/Lib/importlib/_bootstrap_external.py

Lines 1245 to 1257 in c705fd1

	class PathFinder:
	"""Meta path finder for sys.path and package __path__ attributes."""
	@classmethod
	def invalidate_caches(cls):
	"""Call the invalidate_caches() method on all path entry finders
	stored in sys.path_importer_caches (where implemented)."""
	for name, finder in list(sys.path_importer_cache.items()):
	if finder is None:
	del sys.path_importer_cache[name]
	elif hasattr(finder, 'invalidate_caches'):
	finder.invalidate_caches()

.

[miss-islington]

New changeset 3abf6f0 by Desmond Cheong in branch 'master':
bpo-14678: Update zipimport to support importlib.invalidate_caches() (GH-24159)
3abf6f0