def get_first_parents(repo_path):
repo = dulwich.repo.Repo(repo_path)
#: these are the commits that are parents of more than one other commit
first_parents = []
on_merge = False
for entry in repo.get_walker(order=dulwich.walk.ORDER_TOPO):
commit = entry.commit
# In order to properly work on python 2 and 3 we need some utf magic
parents = commit.parents and [_to_str(i) for i in commit.parents]
if not parents:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
elif len(parents) == 1 and not on_merge:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if parents[0] not in first_parents:
first_parents.append(parents[0])
elif len(parents) > 1 and not on_merge:
on_merge = True
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if parents[0] not in first_parents:
first_parents.append(parents[0])
elif parents and commit.sha().hexdigest() in first_parents:
if parents[0] not in first_parents:
first_parents.append(parents[0])
return first_parents
def get_first_parents(repo_path):
repo = dulwich.repo.Repo(repo_path)
#: these are the commits that are parents of more than one other commit
first_parents = []
on_merge = False
for entry in repo.get_walker(order=dulwich.walk.ORDER_TOPO):
commit = entry.commit
if not commit.parents:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
elif len(commit.parents) == 1 and not on_merge:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
elif len(commit.parents) > 1 and not on_merge:
on_merge = True
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
elif commit.parents and commit.sha().hexdigest() in first_parents:
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
return first_parents
def get_first_parents(repo_path):
repo = dulwich.repo.Repo(repo_path)
#: these are the commits that are parents of more than one other commit
first_parents = []
on_merge = False
for entry in repo.get_walker(order=dulwich.walk.ORDER_TOPO):
commit = entry.commit
# In order to properly work on python 2 and 3 we need some utf magic
parents = commit.parents and [i.decode('utf-8') for i in
commit.parents]
if not parents:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
elif len(parents) == 1 and not on_merge:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if parents[0] not in first_parents:
first_parents.append(parents[0])
elif len(parents) > 1 and not on_merge:
on_merge = True
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if parents[0] not in first_parents:
first_parents.append(parents[0])
elif parents and commit.sha().hexdigest() in first_parents:
if parents[0] not in first_parents:
first_parents.append(parents[0])
return first_parents
def get_first_parents(repo_path):
repo = dulwich.repo.Repo(repo_path)
#: these are the commits that are parents of more than one other commit
first_parents = []
on_merge = False
for entry in repo.get_walker(order=dulwich.walk.ORDER_TOPO):
commit = entry.commit
if not commit.parents:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
elif len(commit.parents) == 1 and not on_merge:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
elif len(commit.parents) > 1 and not on_merge:
on_merge = True
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
elif commit.parents and commit.sha().hexdigest() in first_parents:
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
return first_parents
def get_first_parents(repo_path):
repo = dulwich.repo.Repo(repo_path)
#: these are the commits that are parents of more than one other commit
first_parents = []
on_merge = False
for entry in repo.get_walker(order=dulwich.walk.ORDER_TOPO):
commit = entry.commit
if not commit.parents:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
elif len(commit.parents) == 1 and not on_merge:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
elif len(commit.parents) > 1 and not on_merge:
on_merge = True
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
elif commit.parents and commit.sha().hexdigest() in first_parents:
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
if commit.parents:
# If this is the case, we have a shallow git clone
# which means that we don't have the metadata of the
# first's commit parent.
LOGGER.debug('This is a shallow git clone,'
' removing the first\'s commit parent.')
first_parents.pop()
return first_parents
def get_children_per_parent(repo_path):
repo = dulwich.repo.Repo(repo_path)
children_per_parent = defaultdict(set)
for entry in repo.get_walker(order=dulwich.walk.ORDER_TOPO):
for parent in entry.commit.parents:
children_per_parent[parent].add(entry.commit.sha().hexdigest())
return children_per_parent
def get_children_per_parent(repo_path):
repo = dulwich.repo.Repo(repo_path)
children_per_parent = defaultdict(set)
for entry in repo.get_walker(order=dulwich.walk.ORDER_TOPO):
for parent in entry.commit.parents:
children_per_parent[parent].add(entry.commit.sha().hexdigest())
return children_per_parent
def get_changelog(repo_path, from_commit=None):
"""
Given a repo path and an option commit/tag/refspec to start from, will
get the rpm compatible changelog
Args:
repo_path (str): path to the git repo
from_commit (str): refspec (partial commit hash, tag, branch, full
refspec, partial refspec) to start the changelog from
Returns:
str: Rpm compatible changelog
"""
repo = dulwich.repo.Repo(repo_path)
tags = get_tags(repo)
refs = get_refs(repo)
changelog = []
maj_version = 0
feat_version = 0
fix_version = 0
start_including = False
if from_commit is None:
start_including = True
for entry in repo.get_walker(reverse=True):
commit = entry.commit
commit_sha = commit.sha().hexdigest()
if commit_sha in tags:
maj_version, feat_version = tags[commit_sha].split('.')
maj_version = int(maj_version)
feat_version = int(feat_version)
fix_version = 0
elif MAJOR_HEADER.search(commit.message):
maj_version += 1
feat_version = 0
fix_version = 0
elif FEAT_HEADER.search(commit.message):
feat_version += 1
fix_version = 0
else:
fix_version += 1
version = '%s.%s.%s' % (maj_version, feat_version, fix_version)
if (
not start_including and not commit_sha.startswith(from_commit)
and not fuzzy_matches_refs(from_commit, refs.get(commit_sha, []))
):
continue
start_including = True
changelog.append(pretty_commit(commit, version))
return '\n'.join(reversed(changelog))
def build_relatives(repo):
children = {}
parents = {}
heads = {n: s for n, s in repo.get_refs().items()
if n.startswith(b'refs/heads/')}
for entry in repo.get_walker(include=list(heads.values())):
for p in entry.commit.parents:
parents.setdefault(entry.commit.id, set()).add(p)
children.setdefault(p, set()).add(entry.commit.id)
return children, parents
def get_changelog(repo_path, from_commit=None):
"""
Given a repo path and an option commit/tag/refspec to start from, will
get the rpm compatible changelog
Args:
repo_path (str): path to the git repo
from_commit (str): refspec (partial commit hash, tag, branch, full
refspec, partial refspec) to start the changelog from
Returns:
str: Rpm compatible changelog
"""
repo = dulwich.repo.Repo(repo_path)
tags = get_tags(repo)
refs = get_refs(repo)
changelog = []
maj_version = 0
feat_version = 0
fix_version = 0
start_including = False
if from_commit is None:
start_including = True
for entry in repo.get_walker(reverse=True):
commit = entry.commit
commit_sha = commit.sha().hexdigest()
if commit_sha in tags:
maj_version, feat_version = tags[commit_sha].split('.')
maj_version = int(maj_version)
feat_version = int(feat_version)
fix_version = 0
elif MAJOR_HEADER.search(commit.message):
maj_version += 1
feat_version = 0
fix_version = 0
elif FEAT_HEADER.search(commit.message):
feat_version += 1
fix_version = 0
else:
fix_version += 1
version = '%s.%s.%s' % (maj_version, feat_version, fix_version)
if (not start_including and not commit_sha.startswith(from_commit) and
not fuzzy_matches_refs(from_commit, refs.get(commit_sha, []))):
continue
start_including = True
changelog.append(pretty_commit(commit, version))
return '\n'.join(reversed(changelog))
def get_links(repo):
for walk_entry in repo.get_walker():
commit = walk_entry.commit
trace = TraceInfo(
author = commit.author
, committer = commit.committer
, commit_id = commit.id
, date = commit.commit_time # + commit.commit_time_zone ?
, message = commit.message
, bug_ids = detect(commit.message)
)
yield trace
def faasm_main():
if os.environ.get("PYTHONWASM") == "1":
repo_path = "/lib/python3.7/site-packages/pyperformance/benchmarks/data/asyncio.git"
else:
repo_path = "/usr/local/code/faasm/venv/lib/python3.6/site-packages/pyperformance/benchmarks/data/asyncio.git"
if not exists(repo_path):
repo_path = "/usr/local/lib/python3.6/dist-packages/pyperformance/benchmarks/data/asyncio.git"
repo = dulwich.repo.Repo(repo_path)
head = repo.head()
# Iterate on all changes on the Git repository
for _ in repo.get_walker(head):
pass
repo.close()
def get_first_parents(repo_path):
repo = dulwich.repo.Repo(repo_path)
#: these are the commits that are parents of more than one other commit
first_parents = []
on_merge = False
for entry in repo.get_walker(order=dulwich.walk.ORDER_TOPO):
commit = entry.commit
commit_sha = commit.sha().hexdigest().encode('utf-8')
if not commit.parents:
append_parent(first_parents, commit, add_parent=False)
elif len(commit.parents) == 1 and not on_merge:
append_parent(first_parents, commit)
elif len(commit.parents) > 1 and not on_merge:
on_merge = True
append_parent(first_parents, commit)
elif commit.parents and commit_sha in first_parents:
append_parent(first_parents, commit, add_digest=False)
return first_parents
def get_root(repo):
""" Retrieve the hash of the repo root to uniquely identify the git
repository
"""
# Check if the repository is empty
if len(repo.get_refs()) == 0:
return None
# Get walker needs at least the HEAD ref to be present
walker = repo.get_walker()
entry = None
# Iterate on the lazy iterator to get to the last one
for entry in walker:
pass
assert entry is not None
# SHA should always be valid utf-8
return to_utf8(entry.commit.id)
def get_version(repo_path):
"""
Given a repo will return the version string, according to semantic
versioning, counting as non-backwards compatible commit any one with a
message header that matches (case insensitive)::
sem-ver: .*break.*
And as features any commit with a header matching::
sem-ver: feature
And counting any other as a bugfix
"""
repo = dulwich.repo.Repo(repo_path)
tags = get_tags(repo)
maj_version = 0
feat_version = 0
fix_version = 0
for entry in repo.get_walker(reverse=True):
commit = entry.commit
commit_sha = commit.sha().hexdigest()
if commit_sha in tags:
maj_version, feat_version = tags[commit_sha].split('.')
maj_version = int(maj_version)
feat_version = int(feat_version)
fix_version = 0
elif MAJOR_HEADER.search(commit.message):
maj_version += 1
feat_version = 0
fix_version = 0
elif FEAT_HEADER.search(commit.message):
feat_version += 1
fix_version = 0
else:
fix_version += 1
return '%s.%s.%s' % (maj_version, feat_version, fix_version)
def get_version(repo_path):
"""
Given a repo will return the version string, according to semantic
versioning, counting as non-backwards compatible commit any one with a
message header that matches (case insensitive)::
sem-ver: .*break.*
And as features any commit with a header matching::
sem-ver: feature
And counting any other as a bugfix
"""
repo = dulwich.repo.Repo(repo_path)
tags = get_tags(repo)
maj_version = 0
feat_version = 0
fix_version = 0
for entry in repo.get_walker(reverse=True):
commit = entry.commit
commit_sha = commit.sha().hexdigest()
if commit_sha in tags:
maj_version, feat_version = tags[commit_sha].split('.')
maj_version = int(maj_version)
feat_version = int(feat_version)
fix_version = 0
elif MAJOR_HEADER.search(commit.message):
maj_version += 1
feat_version = 0
fix_version = 0
elif FEAT_HEADER.search(commit.message):
feat_version += 1
fix_version = 0
else:
fix_version += 1
return '%s.%s.%s' % (maj_version, feat_version, fix_version)
def releases(git_dir):
repo = dulwich.repo.Repo(git_dir)
releases = []
tags = [
repo.get_object(repo.refs['refs/tags/%s' % tag])
for tag in repo.refs.keys(base='refs/tags/')
]
tag_commits = {}
for tag in tags:
if tag.type_name == 'tag':
tag_commits.setdefault(tag.object[1], []).append(tag)
for walker in repo.get_walker(repo.refs['HEAD']):
if walker.commit.id not in tag_commits:
continue
for tag in tag_commits[walker.commit.id]:
m = re.match(r'beaker-([\d.]*)$', tag.name)
if m:
break
# also check for tito tags, used up to 0.14.1
m = re.match(r'beaker-([\d.]*)-1$', tag.name)
if m:
break
if not m:
continue
version = m.group(1)
name, email = re.match(r'(.*) <(.*)>', tag.tagger).groups()
timestamp = datetime.datetime.fromtimestamp(
tag.tag_time, tzoffset(None, tag.tag_timezone))
releases.append(
Release(version=m.group(1),
timestamp=timestamp,
name=name.decode('utf-8'),
email=email,
tag=tag.name))
releases = sorted(releases, key=lambda r: r.timestamp, reverse=True)
# skip anything prior to 0.9
releases = list(takewhile(lambda r: r.version != '0.8.99', releases))
return sorted(releases, key=lambda r: r.version_tuple, reverse=True)
def get_first_parents(repo_path):
repo = dulwich.repo.Repo(repo_path)
#: these are the commits that are parents of more than one other commit
first_parents = []
on_merge = False
for entry in repo.get_walker(order=dulwich.walk.ORDER_TOPO):
commit = entry.commit
if not commit.parents:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
elif len(commit.parents) == 1 and not on_merge:
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
elif len(commit.parents) > 1 and not on_merge:
on_merge = True
if commit.sha().hexdigest() not in first_parents:
first_parents.append(commit.sha().hexdigest())
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
elif commit.parents and commit.sha().hexdigest() in first_parents:
if commit.parents[0] not in first_parents:
first_parents.append(commit.parents[0])
if commit.parents:
# If this is the case, we have a shallow git clone
# which means that we don't have the metadata of the
# first's commit parent.
LOGGER.debug(
'This is a shallow git clone,'
' removing the first\'s commit parent.'
)
first_parents.pop()
return first_parents
def iter_all_commits(repo):
# iterate on all changes on the Git repository
for entry in repo.get_walker(head):
pass
projects = load_projects()
for project in projects:
dest_fn = os.path.join(project.full_path, 'issue2git.csv')
if os.path.exists(dest_fn):
continue
if project.name == 'eclipse':
continue
repos = load_repos(project)
golds = load_goldsets(project)
ids = set(i for i,g in golds)
i2g = dict.fromkeys(ids)
for k in i2g:
i2g[k] = set()
for repo in repos:
#b = re.compile('BOOKKEEPER-([\d]+):')
#b = re.compile('ZOOKEEPER-([\d]+)')
b = re.compile('%s-([\d]+)' % project.name.upper())
for entry in repo.get_walker():
a = entry.commit
for issue in b.findall(a.message):
if issue in i2g:
i2g[issue].add(a.id)
with open(dest_fn, 'w') as f:
w = csv.writer(f)
for issue, gits in i2g.items():
if gits:
w.writerow([issue] + list(gits))
projects = load_projects()
for project in projects:
dest_fn = os.path.join(project.full_path, 'issue2git.csv')
if os.path.exists(dest_fn):
continue
if project.name == 'eclipse':
continue
repos = load_repos(project)
golds = load_goldsets(project)
ids = set(i for i, g in golds)
i2g = dict.fromkeys(ids)
for k in i2g:
i2g[k] = set()
for repo in repos:
#b = re.compile('BOOKKEEPER-([\d]+):')
#b = re.compile('ZOOKEEPER-([\d]+)')
b = re.compile('%s-([\d]+)' % project.name.upper())
for entry in repo.get_walker():
a = entry.commit
for issue in b.findall(a.message):
if issue in i2g:
i2g[issue].add(a.id)
with open(dest_fn, 'w') as f:
w = csv.writer(f)
for issue, gits in i2g.items():
if gits:
w.writerow([issue] + list(gits))
