def test_get_commits_modified_file(repo: GitRepository):
commits = repo.get_commits_modified_file('file2.java')
assert len(commits) == 3
assert '09f6182cef737db02a085e1d018963c7a29bde5a' in commits
assert '6411e3096dd2070438a17b225f44475136e54e3a' in commits
assert 'a88c84ddf42066611e76e6cb690144e5357d132c' in commits
def traverse_commits(self) -> Generator[Commit, None, None]:
"""
Analyze all the specified commits (all of them by default), returning
a generator of commits.
"""
for path_repo in self._conf.get('path_to_repos'):
# if it is a remote repo, clone it first in a temporary folder!
if self._is_remote(path_repo):
if self._conf.get('clone_repo_to'):
clone_folder = str(Path(self._conf.get('clone_repo_to')))
if not os.path.isdir(clone_folder):
raise Exception(
"Not a directory: {0}".format(clone_folder))
path_repo = self._clone_remote_repos(
clone_folder, path_repo)
else:
tmp_folder = tempfile.TemporaryDirectory()
path_repo = self._clone_remote_repos(
tmp_folder.name, path_repo)
git_repo = GitRepository(path_repo, self._conf)
self._conf.set_value("git_repo", git_repo)
self._conf.sanity_check_filters()
logger.info('Analyzing git repository in %s', git_repo.path)
# Get the commits that modified the filepath. In this case, we can not use
# git rev-list since it doesn't have the option --follow, necessary to follow
# the renames. Hence, we manually call git log instead
if self._conf.get('filepath') is not None:
self._conf.set_value(
'filepath_commits',
git_repo.get_commits_modified_file(
self._conf.get('filepath')))
# Gets only the commits that are tagged
if self._conf.get('only_releases'):
self._conf.set_value('tagged_commits',
git_repo.get_tagged_commits())
# Build the arguments to pass to git rev-list.
rev, kwargs = self._conf.build_args()
# Iterate over all the commits returned by git rev-list
for commit in git_repo.get_list_commits(rev, **kwargs):
logger.info('Commit #%s in %s from %s', commit.hash,
commit.committer_date, commit.author.name)
if self._conf.is_commit_filtered(commit):
logger.info('Commit #%s filtered', commit.hash)
continue
yield commit
# cleaning, this is necessary since GitPython issues on memory leaks
self._conf.set_value("git_repo", None)
git_repo.clear()
def traverse_commits(self) -> Generator[Commit, None, None]:
"""
Analyze all the specified commits (all of them by default), returning
a generator of commits.
"""
for path_repo in self._conf.get('path_to_repos'):
if self._is_remote(path_repo):
path_repo = self._clone_remote_repos(self._clone_folder(),
path_repo)
git_repo = GitRepository(path_repo, self._conf)
# saving the GitRepository object for further use
self._conf.set_value("git_repo", git_repo)
# when multiple repos are given in input, this variable will serve as a reminder
# of which one we are currently analyzing
self._conf.set_value('path_to_repo', path_repo)
# checking that the filters are set correctly
self._conf.sanity_check_filters()
logger.info('Analyzing git repository in %s', git_repo.path)
# Get the commits that modified the filepath. In this case, we can not use
# git rev-list since it doesn't have the option --follow, necessary to follow
# the renames. Hence, we manually call git log instead
if self._conf.get('filepath') is not None:
self._conf.set_value(
'filepath_commits',
git_repo.get_commits_modified_file(
self._conf.get('filepath')))
# Gets only the commits that are tagged
if self._conf.get('only_releases'):
self._conf.set_value('tagged_commits',
git_repo.get_tagged_commits())
# Build the arguments to pass to git rev-list.
rev, kwargs = self._conf.build_args()
# Iterate over all the commits returned by git rev-list
for commit in git_repo.get_list_commits(rev, **kwargs):
logger.info('Commit #%s in %s from %s', commit.hash,
commit.committer_date, commit.author.name)
if self._conf.is_commit_filtered(commit):
logger.info('Commit #%s filtered', commit.hash)
continue
yield commit
# cleaning, this is necessary since GitPython issues on memory leaks
self._conf.set_value("git_repo", None)
git_repo.clear()
def traverse_commits(self) -> Generator[Commit, None, None]:
"""
Analyze all the specified commits (all of them by default), returning
a generator of commits.
"""
for path_repo in self._conf.get('path_to_repos'):
# if it is a remote repo, clone it first in a temporary folder!
if self._is_remote(path_repo):
if self._conf.get('clone_repo_to'):
clone_folder = str(Path(self._conf.get('clone_repo_to')))
if not os.path.isdir(clone_folder):
raise Exception("Not a directory: " \
"{0}".format(clone_folder))
path_repo = self._clone_remote_repos(
clone_folder, path_repo)
else:
tmp_folder = tempfile.TemporaryDirectory()
path_repo = self._clone_remote_repos(
tmp_folder.name, path_repo)
git_repo = GitRepository(path_repo, self._conf)
self._conf.set_value("git_repo", git_repo)
self._conf.sanity_check_filters()
logger.info('Analyzing git repository in %s', git_repo.path)
if self._conf.get('filepath') is not None:
self._conf.set_value(
'filepath_commits',
git_repo.get_commits_modified_file(
self._conf.get('filepath')))
if self._conf.get('only_releases'):
self._conf.set_value('tagged_commits',
git_repo.get_tagged_commits())
for commit in git_repo.get_list_commits(
self._conf.get('only_in_branch'),
not self._conf.get('reversed_order')):
logger.info('Commit #%s in %s from %s', commit.hash,
commit.committer_date, commit.author.name)
if self._conf.is_commit_filtered(commit):
logger.info('Commit #%s filtered', commit.hash)
continue
yield commit
# cleaning
self._conf.set_value("git_repo", None)
git_repo.clear()
def traverse_commits(self) -> Generator[Commit, None, None]:
"""
Analyze all the specified commits (all of them by default), returning
a generator of commits.
"""
if isinstance(self._path_to_repo, str):
self._path_to_repo = [self._path_to_repo]
for path_repo in self._path_to_repo:
# if it is a remote repo, clone it first in a temporary folder!
if self._isremote(path_repo):
tmp_folder = tempfile.TemporaryDirectory()
path_repo = self._clone_remote_repos(tmp_folder.name,
path_repo)
git_repo = GitRepository(path_repo)
self._sanity_check_filters(git_repo)
self._check_timezones()
logger.info('Analyzing git repository in %s', git_repo.path)
if self._filepath is not None:
self._filepath_commits = git_repo.get_commits_modified_file(
self._filepath)
if self._only_releases:
self._tagged_commits = git_repo.get_tagged_commits()
for commit in git_repo.get_list_commits(self._only_in_branch,
not self._reversed_order):
logger.info('Commit #%s in %s from %s', commit.hash,
commit.committer_date, commit.author.name)
if self._is_commit_filtered(commit):
logger.info('Commit #%s filtered', commit.hash)
continue
yield commit
def test_get_commits_modified_file_missing_file():
gr = GitRepository('test-repos/test1/')
commits = gr.get_commits_modified_file('non-existing-file.java')
assert len(commits) == 0
def test_get_commits_modified_file_missing_file(repo: GitRepository):
commits = repo.get_commits_modified_file('non-existing-file.java')
assert len(commits) == 0
def traverse_commits(self) -> Generator[Commit, None, None]:
"""
Analyze all the specified commits (all of them by default), returning
a generator of commits.
"""
for path_repo in self._conf.get('path_to_repos'):
local_path_repo = path_repo
if self._is_remote(path_repo):
local_path_repo = self._clone_remote_repo(
self._clone_folder(), path_repo)
# Get absolute path
local_path_repo = str(Path(local_path_repo).expanduser().resolve())
# when multiple repos are given in input, this variable will serve as a reminder
# of which one we are currently analyzing
self._conf.set_value('path_to_repo', local_path_repo)
git_repo = GitRepository(local_path_repo, self._conf)
# saving the GitRepository object for further use
self._conf.set_value("git_repo", git_repo)
# checking that the filters are set correctly
self._conf.sanity_check_filters()
logger.info('Analyzing git repository in %s', git_repo.path)
# Get the commits that modified the filepath. In this case, we can not use
# git rev-list since it doesn't have the option --follow, necessary to follow
# the renames. Hence, we manually call git log instead
if self._conf.get('filepath') is not None:
self._conf.set_value(
'filepath_commits',
git_repo.get_commits_modified_file(
self._conf.get('filepath')))
# Gets only the commits that are tagged
if self._conf.get('only_releases'):
self._conf.set_value('tagged_commits',
git_repo.get_tagged_commits())
# Build the arguments to pass to git rev-list.
rev, kwargs = self._conf.build_args()
# Iterate over all the commits returned by git rev-list
for commit in git_repo.get_list_commits(rev, **kwargs):
logger.info('Commit #%s in %s from %s', commit.hash,
commit.committer_date, commit.author.name)
if self._conf.is_commit_filtered(commit):
logger.info('Commit #%s filtered', commit.hash)
continue
yield commit
# cleaning, this is necessary since GitPython issues on memory leaks
self._conf.set_value("git_repo", None)
git_repo.clear()
# delete the temporary directory if created
if self._is_remote(path_repo) and self._cleanup is True:
assert self._tmp_dir is not None
try:
self._tmp_dir.cleanup()
except PermissionError:
# on Windows, Python 3.5, 3.6, 3.7 are not able to delete
# git directories because of read-only files. This is now fixed
# in python 3.8. In this case, we need to use an
# onerror callback to clear the read-only bit.
# see https://docs.python.org/3/library/shutil.html?highlight=shutil#rmtree-example
def remove_readonly(func, path, _):
os.chmod(path, stat.S_IWRITE)
func(path)
shutil.rmtree(self._tmp_dir.name, onerror=remove_readonly)
