summaryrefslogtreecommitdiffstats
path: root/meta/recipes-devtools/python/python3
diff options
context:
space:
mode:
authorAlexander Kanavin <alex@linutronix.de>2024-06-28 17:01:16 +0200
committerRichard Purdie <richard.purdie@linuxfoundation.org>2024-07-01 13:56:03 +0100
commit058c3a6e2b30103202f3995930825fc2f366274f (patch)
tree3653f37a14086baf10af6e010b6e2e3534e0abd8 /meta/recipes-devtools/python/python3
parent0703fd6c6c336b5e676a463772e026dd65f5a20d (diff)
downloadpoky-058c3a6e2b30103202f3995930825fc2f366274f.tar.gz
python3: drop deterministic_imports.patch
https://github.com/python/importlib_metadata/issues/489 Upstream discussion revealed that: - the issue was happening due to a bug in setuptools - the bug was fixed in setuptools 63 (I confirmed this: empty dist-info directories no longer seem to trigger any issues if the patch is dropped) - the patch would obscure any further issues of this kind instead of exposing them for easy fixing, so they suggest we drop it and report if anything else pops up (From OE-Core rev: 1a4f25abf29b47949782641c68cc7d3e4136988a) Signed-off-by: Alexander Kanavin <alex@linutronix.de> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Diffstat (limited to 'meta/recipes-devtools/python/python3')
-rw-r--r--meta/recipes-devtools/python/python3/deterministic_imports.patch40
1 files changed, 0 insertions, 40 deletions
diff --git a/meta/recipes-devtools/python/python3/deterministic_imports.patch b/meta/recipes-devtools/python/python3/deterministic_imports.patch
deleted file mode 100644
index 2de6ae2e98..0000000000
--- a/meta/recipes-devtools/python/python3/deterministic_imports.patch
+++ /dev/null
@@ -1,40 +0,0 @@
1From 039d5e652796b55f1132afa568c7432b6ed89afd Mon Sep 17 00:00:00 2001
2From: Richard Purdie <richard.purdie@linuxfoundation.org>
3Date: Fri, 27 May 2022 17:05:44 +0100
4Subject: [PATCH] python3: Ensure stale empty python module directories don't
5
6There are two issues here. Firstly, the modules are accessed in on disk order. This
7means behaviour seen on one system might not reproduce on another and is a real headache.
8
9Secondly, empty directories left behind by previous modules might be looked at. This
10has caused a long string of different issues for us.
11
12As a result, patch this to a behaviour which works for us.
13
14Upstream-Status: Submitted [https://github.com/python/cpython/issues/120492; need to first talk to upstream to see if they'll take one or both fixes]
15Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
16
17---
18 Lib/importlib/metadata/__init__.py | 9 ++++++++-
19 1 file changed, 8 insertions(+), 1 deletion(-)
20
21diff --git a/Lib/importlib/metadata/__init__.py b/Lib/importlib/metadata/__init__.py
22index 82e0ce1..969cac4 100644
23--- a/Lib/importlib/metadata/__init__.py
24+++ b/Lib/importlib/metadata/__init__.py
25@@ -710,7 +710,14 @@ class Lookup:
26 self.infos = FreezableDefaultDict(list)
27 self.eggs = FreezableDefaultDict(list)
28
29- for child in path.children():
30+ for child in sorted(path.children()):
31+ childpath = pathlib.Path(path.root, child)
32+ try:
33+ if childpath.is_dir() and not any(childpath.iterdir()):
34+ # Empty directories aren't interesting
35+ continue
36+ except PermissionError:
37+ continue
38 low = child.lower()
39 if low.endswith((".dist-info", ".egg-info")):
40 # rpartition is faster than splitext and suitable for this purpose.