Broken GIT_SORT_TIME ?
Open
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 1.7k
- Forks
- 408
- Avg merge
- 2d 57m
- Merged PRs (30d)
- 7
Description
This is a script to reproduce the behavior I'm seeing:
#!/usr/bin/env python
import tempfile
import time
import os
import pygit2
path = tempfile.mkdtemp(prefix='pygit2_test_')
repo = pygit2.init_repository(path)
author = pygit2.Signature('Alice Author', 'alice@authors.tld')
committer = pygit2.Signature('Cecil Committer', 'cecil@committers.tld')
# Commit #1
# Create a file in that git repo
with open(os.path.join(path, 'sources'), 'w') as stream:
stream.write('foo\n bar')
repo.index.add('sources')
repo.index.write()
parents = []
# Commits the files added
tree = repo.index.write_tree()
print(repo.create_commit(
'refs/heads/master', # the name of the reference to update
author,
committer,
'Commit #1',
# binary string representing the tree object ID
tree,
# list of binary strings representing parents of the new commit
parents,
))
commit1 = repo.revparse_single('HEAD')
time.sleep(0.1)
# Commit #2
parents = [commit1.oid.hex]
subfolder = os.path.join('folder1', 'folder2')
os.makedirs(os.path.join(path, subfolder))
# Create a file in that git repo
with open(os.path.join(path, subfolder, 'file'), 'w') as stream:
stream.write('foo\n bar\nbaz')
repo.index.add(os.path.join(subfolder, 'file'))
repo.index.write()
# Commits the files added
tree = repo.index.write_tree()
print(repo.create_commit(
'refs/heads/master', # the name of the reference to update
author,
committer,
'Commit #2',
# binary string representing the tree object ID
tree,
# list of binary strings representing parents of the new commit
parents
))
commit2 = repo.revparse_single('HEAD')
time.sleep(0.1)
# Commit #3
parents = [commit2.oid.hex]
# Update the sources file
with open(os.path.join(path, 'sources'), 'w') as stream:
stream.write('foo\n bar\nbaz!')
repo.index.add('sources')
repo.index.write()
# Commits the files added
tree = repo.index.write_tree()
print(repo.create_commit(
'refs/heads/master', # the name of the reference to update
author,
committer,
'Commit #3',
# binary string representing the tree object ID
tree,
# list of binary strings representing parents of the new commit
parents
))
commit3 = repo.revparse_single('HEAD')
print('')
main_walker = repo.walk(commit3.oid.hex, pygit2.GIT_SORT_TIME)
commits_msgs = []
while 1:
try:
com = main_walker.next()
print(com.message)
print(com.oid.hex)
except StopIteration:
break
It gives me the following output:
7335ffcfdb2039a9631ef5c891ec6a12feaeb342
50ea7d4691ee78b60323ac1d4174a6609190fbaf
47e61d2eaa8c5d3d1562475a43aa99c0e7fb9b8a
Commit #3
47e61d2eaa8c5d3d1562475a43aa99c0e7fb9b8a
Commit #1
7335ffcfdb2039a9631ef5c891ec6a12feaeb342
Commit #2
50ea7d4691ee78b60323ac1d4174a6609190fbaf
The first three lines are the commit as they are committed.
The following lines are the commit message and hash as repo.walk(commit3, pygit2.GIT_SORT_TIME) is returning them.
Something looks broken to me :)
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by running the reproduction script and inspecting the pygit2 binding for repo.walk with GIT_SORT_TIME. Compare the observed order with the documented sorting behavior and existing walker tests; done means the ordering is corrected or the behavior is clearly documented with a regression test.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- git, python
- Domain
- devtools
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100