Repository navigation
Compatibility problem with Python 3.9? #102
Description
Hello,
I'm the maintainer of git-up in Fedora. Recently we did a mass archive against Python 3.9 and git-up's test started failing. Here's the relevant log:
+ rm -rfv /builddir/build/BUILDROOT/git-up-1.6.1-5.fc33.x86_64/usr/bin/__pycache__
+ /usr/lib/rpm/find-debuginfo.sh -j8 --strict-build-id -m -i --build-id-seed 1.6.1-5.fc33 --unique-debug-suffix -1.6.1-5.fc33.x86_64 --unique-debug-src-base git-up-1.6.1-5.fc33.x86_64 --run-dwz --dwz-low-mem-die-limit 10000000 --dwz-max-die-limit 110000000 -S debugsourcefiles.list /builddir/build/BUILD/git-up-1.6.1
find: 'debug': No such file or directory
+ /usr/lib/rpm/check-buildroot
+ /usr/lib/rpm/redhat/brp-ldconfig
+ /usr/lib/rpm/brp-compress
+ /usr/lib/rpm/redhat/brp-strip-lto /usr/bin/strip
+ /usr/lib/rpm/brp-strip-static-archive /usr/bin/strip
+ /usr/lib/rpm/redhat/brp-python-bytecompile '' 1 0
Bytecompiling .py files below /builddir/build/BUILDROOT/git-up-1.6.1-5.fc33.x86_64/usr/lib/python3.9 using /usr/bin/python3.9
+ /usr/lib/rpm/brp-python-hardlink
+ /usr/lib/rpm/redhat/brp-mangle-shebangs
Executing(%check): /bin/sh -e /var/tmp/rpm-tmp.YFhh0w
+ umask 022
+ cd /builddir/build/BUILD
+ cd git-up-1.6.1
+ git config --global user.email koji@fedoraproject.org
+ git config --global user.name 'Koji Build System'
+ PYTHONPATH=/builddir/build/BUILDROOT/git-up-1.6.1-5.fc33.x86_64/usr/lib/python3.9/site-packages
+ nosetests-3.9
.S..S...S........................E..S
======================================================================
ERROR: test suite for <module 'PyGitUp.tests.test_up_to_date' from '/builddir/build/BUILD/git-up-1.6.1/PyGitUp/tests/test_up_to_date.py'>
----------------------------------------------------------------------
Traceback (most recent call last):
File "/usr/lib/python3.9/site-packages/nose/suite.py", line 209, in run
self.setUp()
File "/usr/lib/python3.9/site-packages/nose/suite.py", line 292, in setUp
self.setupContext(ancestor)
File "/usr/lib/python3.9/site-packages/nose/suite.py", line 315, in setupContext
try_run(context, names)
File "/usr/lib/python3.9/site-packages/nose/util.py", line 471, in try_run
return func()
File "/builddir/build/BUILD/git-up-1.6.1/PyGitUp/tests/test_up_to_date.py", line 32, in setup
repo.remotes.origin.pull()
File "/usr/lib/python3.9/site-packages/git/remote.py", line 811, in pull
res = self._get_fetch_info_from_stderr(proc, progress)
File "/usr/lib/python3.9/site-packages/git/remote.py", line 708, in _get_fetch_info_from_stderr
output.extend(FetchInfo._from_line(self.repo, err_line, fetch_line)
File "/usr/lib/python3.9/site-packages/git/remote.py", line 708, in <genexpr>
output.extend(FetchInfo._from_line(self.repo, err_line, fetch_line)
File "/usr/lib/python3.9/site-packages/git/remote.py", line 292, in _from_line
raise ValueError("Failed to parse line: %r" % line)
ValueError: Failed to parse line: ' git config pull.rebase false # merge (the default strategy)'
-------------------- >> begin captured logging << --------------------
git.cmd: DEBUG: Popen(['git', 'init'], cwd=/tmp/PyGitUp.ufk366oy/master.up-to-date, universal_newlines=False, shell=None, istream=None)
git.cmd: DEBUG: Popen(['git', 'cat-file', '--batch-check'], cwd=/tmp/PyGitUp.ufk366oy/master.up-to-date, universal_newlines=False, shell=None, istream=<valid stream>)
git.cmd: DEBUG: Popen(['git', 'cat-file', '--batch'], cwd=/tmp/PyGitUp.ufk366oy/master.up-to-date, universal_newlines=False, shell=None, istream=<valid stream>)
git.cmd: DEBUG: Popen(['git', 'checkout', '-b', 'initial'], cwd=/tmp/PyGitUp.ufk366oy/master.up-to-date, universal_newlines=False, shell=None, istream=None)
git.cmd: DEBUG: Popen(['git', 'checkout', '-b', 'up-to-date'], cwd=/tmp/PyGitUp.ufk366oy/master.up-to-date, universal_newlines=False, shell=None, istream=None)
git.cmd: DEBUG: Popen(['git', 'clone', '-b', 'up-to-date', '-v', '/tmp/PyGitUp.ufk366oy/master.up-to-date/.git', '/tmp/PyGitUp.ufk366oy/up-to-date'], cwd=/tmp/PyGitUp.ufk366oy/master.up-to-date, universal_newlines=True, shell=None, istream=None)
git.repo.base: DEBUG: Cmd(['git', 'clone', '-b', 'up-to-date', '-v', '/tmp/PyGitUp.ufk366oy/master.up-to-date/.git', '/tmp/PyGitUp.ufk366oy/up-to-date'])'s unused stdout:
git.cmd: DEBUG: Popen(['git', 'pull', '-v', 'origin'], cwd=/tmp/PyGitUp.ufk366oy/up-to-date, universal_newlines=True, shell=None, istream=None)
git.remote: DEBUG: Fetch head lines do not match lines provided via progress information
length of progress lines 6 should be equal to lines in FETCH_HEAD file 3
Will ignore extra progress lines or fetch head lines.
git.remote: DEBUG: info lines: [' git config pull.rebase false # merge (the default strategy)', ' git config pull.rebase true # rebase', ' git config pull.ff only # fast-forward only', ' eb2d67d..6adb676 up-to-date -> origin/up-to-date', ' = [up to date] initial -> origin/initial', ' = [up to date] master -> origin/master']
git.remote: DEBUG: head info : ["6adb676af33e7f95cf91d1fc3841adbbbcbc97f7\t\tbranch 'up-to-date' of /tmp/PyGitUp.ufk366oy/master.up-to-date/\n", "eb2d67d81794f11739aa098c22b41e129a2ce70d\tnot-for-merge\tbranch 'initial' of /tmp/PyGitUp.ufk366oy/master.up-to-date/\n", "eb2d67d81794f11739aa098c22b41e129a2ce70d\tnot-for-merge\tbranch 'master' of /tmp/PyGitUp.ufk366oy/master.up-to-date/\n"]
--------------------- >> end captured logging << ---------------------
----------------------------------------------------------------------
Ran 36 tests in 3.301s
FAILED (SKIP=4, errors=1)
I'm not sure why there would be a different hash here. Do you think this might be due to a change in Python 3.9?
Activity
msiemens commentedon Aug 15, 2020
Interesting. I suspect this is more due to a Git update than caused by the Python upgrade. I use Python 3.8 locally and tests started to fail for me too. I'll look into this…
msiemens commentedon Aug 15, 2020
Seems like this is related to gitpython-developers/GitPython#1014 which in turn is caused by git/git@d18c950 which was published in Git v2.27.0. The issue has been fixed in GitPython v3.1.3 but we're using GitPython 2.x. Updating to GitPython 3.x means dropping Python 2 support. So I'll publish a release that drops Python 2 support in order to fix this issue.
mhjacks commentedon Aug 15, 2020
Thank you! That will work well for us as we don't support Python 2 in Fedora anymore.
msiemens commentedon Aug 15, 2020
Should be fixed with v2.0.0 🙃
mhjacks commentedon Aug 16, 2020
I don't see setup.py in the new release tarball?
msiemens commentedon Aug 25, 2020
Oh, weird. I forgot to publish the release to PyPI. The release there should contain the
setup.py. I switched the build to Poetry which generates thesetup.pywhen building the release artifacts.mhjacks commentedon Aug 25, 2020
Ah, that makes sense. Thanks! I was unaware that Poetry did that. I'll work on packaging 2.0.0 for Fedora.
mhjacks commentedon Aug 25, 2020
OK, I'm working on packaging, and I've run a super-minor test problem (init.py in tests/ doesn't seem to have
import osand the test suite complains about that). But there's also this:% nosetests-3
Is this another case of "the output changed" or is it something more significant?
msiemens commentedon Aug 26, 2020
That's strange. Seems like this is the exact issue that you had at the start. Could you post the output of running
python -c 'import git; print(git.__version__)'in your build environment where PyGitUp is installed?mhjacks commentedon Aug 26, 2020
Which is interesting, because the package claims to be 3.1.0:
python3-GitPython-3.1.0-3.fc33.noarch
Clearly the package is installing 3.0.9 though:
% rpm -ql python3-GitPython-3.1.0-3.fc33.noarch
/usr/lib/python3.9/site-packages/GitPython-3.0.9-py3.9.egg-info
/usr/lib/python3.9/site-packages/GitPython-3.0.9-py3.9.egg-info/PKG-INFO
/usr/lib/python3.9/site-packages/GitPython-3.0.9-py3.9.egg-info/SOURCES.txt
/usr/lib/python3.9/site-packages/GitPython-3.0.9-py3.9.egg-info/dependency_links.txt
/usr/lib/python3.9/site-packages/GitPython-3.0.9-py3.9.egg-info/not-zip-safe
/usr/lib/python3.9/site-packages/GitPython-3.0.9-py3.9.egg-info/requires.txt
/usr/lib/python3.9/site-packages/GitPython-3.0.9-py3.9.egg-info/top_level.txt
mhjacks commentedon Aug 30, 2020
It appears we're past the initially reported issue, so I'm ready to close this (Fedora mislabeling a package version isn't really your problem). I see you also loosened the dependency on colorama. Thanks! Would you please consider adding the
import osto the test suite as well?msiemens commentedon Sep 2, 2020
Where is it missing? As far as I can tell, the tests run without errors
mhjacks commentedon Sep 2, 2020
When I run
nosetestson f32I get this error:
I do not see an
import osin this file:https://github.com/msiemens/PyGitUp/blob/master/PyGitUp/tests/__init__.py so I suspect that's why I'm seeing this error. I could be wrong, though.
msiemens commentedon Sep 3, 2020
Funny! Since PyGitUp 2.0 I switched to PyTest for running the tests and there everything runs fine. But running with nosetests throws an exception as in your example. I'll publish a post-release with a fix
msiemens commentedon Sep 3, 2020
This should be fixed now with v2.0.1.post1
mhjacks commentedon Sep 3, 2020
Excellent! Thanks.