def test_simple_local(self):
f1_1 = make_object(Blob, data=b'f1')
commit_spec = [[1], [2, 1], [3, 1, 2]]
trees = {
1: [(b'f1', f1_1), (b'f2', f1_1)],
2: [(b'f1', f1_1), (b'f2', f1_1)],
3: [(b'f1', f1_1), (b'f2', f1_1)],
}
c1, c2, c3 = build_commit_graph(self.repo.object_store, commit_spec,
trees)
self.repo.refs[b"refs/heads/master"] = c3.id
self.repo.refs[b"refs/tags/foo"] = c3.id
target_path = tempfile.mkdtemp()
errstream = BytesIO()
self.addCleanup(shutil.rmtree, target_path)
r = porcelain.clone(self.repo.path,
target_path,
checkout=False,
errstream=errstream)
self.assertEqual(r.path, target_path)
target_repo = Repo(target_path)
self.assertEqual(target_repo.head(), c3.id)
self.assertEqual(c3.id, target_repo.refs[b'refs/tags/foo'])
self.assertTrue(b'f1' not in os.listdir(target_path))
self.assertTrue(b'f2' not in os.listdir(target_path))
c = r.get_config()
encoded_path = self.repo.path
if not isinstance(encoded_path, bytes):
encoded_path = encoded_path.encode('utf-8')
self.assertEqual(encoded_path, c.get((b'remote', b'origin'), b'url'))
self.assertEqual(b'+refs/heads/*:refs/remotes/origin/*',
c.get((b'remote', b'origin'), b'fetch'))
def get_git_versions(repos, get_dirty_status=False, verbose=0):
""" Returns the repository head guid and dirty status and package version number if installed via pip.
The version is only returned if the repo is installed as pip package without edit mode.
NOTE: currently the dirty status is not working correctly due to a bug in dulwich...
Args:
repos ([str]): a list with repositories, e.g. ['qtt', 'qcodes']
get_dirty_status (bool): selects whether to use the dulwich package and collect the local code
changes for the repositories.
Retuns:
r (dict): dictionary with repo names, head guid and (optionally) dirty status for each given repository.
"""
heads = dict()
dirty_stats = dict()
for repo in repos:
try:
package = importlib.import_module(repo)
init_location = os.path.split(package.__file__)[0]
repo_location = os.path.join(init_location, '..')
repository = Repo(repo_location)
heads[repo] = repository.head().decode('ascii')
if get_dirty_status:
status = porcelain.status(repository)
is_dirty = len(status.unstaged) == 0 or any(
len(item) != 0 for item in status.staged.values())
dirty_stats[repo] = is_dirty
except (AttributeError, ModuleNotFoundError, NotGitRepository):
heads[repo] = 'none'
if get_dirty_status:
dirty_stats[repo] = 'none'
if verbose:
print('{0}: {1}'.format(repo, heads[repo]))
return (heads, dirty_stats)
def checkout(repo_path='.'):
repo = Repo(repo_path)
indexfile = repo.index_path()
obj_sto = repo.object_store
tree_id = repo[repo.head()].tree
build_index_from_tree(repo_path, indexfile, obj_sto, tree_id)
return [obj_sto.iter_tree_contents(tree_id)]
def dulwichCommit(self, filePath, fullPath, kind):
git = Repo(AUTOGIT_PATH)
staged = map(str, [filePath])
git.stage(staged)
index = git.open_index()
try:
committer = git._get_user_identity()
except ValueError:
committer = "autogit"
try:
head = git.head()
except KeyError:
return git.do_commit('%s - autogit commit (via dulwich)' % kind,
committer=committer)
changes = list(
tree_changes(git, index.commit(git.object_store),
git['HEAD'].tree))
if changes and len(changes) > 0:
return git.do_commit('%s - autogit commit (via dulwich)' % kind,
committer=committer)
return None
async def info(self, ctx: Context):
"""
Get information about the bot
"""
embed = Embed(
description=
"A utility bot designed just for the Python server! Try `bot.help()` for more info.",
url="https://github.com/discord-python/bot")
repo = Repo(".")
sha = repo[repo.head()].sha().hexdigest()
embed.add_field(name="Total Users",
value=str(len(
self.bot.get_guild(PYTHON_GUILD).members)))
embed.add_field(name="Git SHA", value=str(sha)[:7])
embed.set_author(name="Python Bot",
url="https://github.com/discord-python/bot",
icon_url=BOT_AVATAR_URL)
log.info(
f"{ctx.author} called bot.about(). Returning information about the bot."
)
await ctx.send(embed=embed)
async def about_command(self, ctx: Context):
"""
Get information about the bot
"""
embed = Embed(
description=
"A utility bot designed just for the Python server! Try `!help` for more info.",
url="https://gitlab.com/discord-python/projects/bot")
repo = Repo(".")
sha = repo[repo.head()].sha().hexdigest()
embed.add_field(name="Total Users",
value=str(len(self.bot.get_guild(Guild.id).members)))
embed.add_field(name="Git SHA", value=str(sha)[:7])
embed.set_author(name="Python Bot",
url="https://gitlab.com/discord-python/projects/bot",
icon_url=URLs.bot_avatar)
log.info(
f"{ctx.author} called !about. Returning information about the bot."
)
await ctx.send(embed=embed)
def test_simple_local(self):
f1_1 = make_object(Blob, data=b'f1')
commit_spec = [[1], [2, 1], [3, 1, 2]]
trees = {
1: [(b'f1', f1_1), (b'f2', f1_1)],
2: [(b'f1', f1_1), (b'f2', f1_1)],
3: [(b'f1', f1_1), (b'f2', f1_1)],
}
c1, c2, c3 = build_commit_graph(self.repo.object_store, commit_spec,
trees)
self.repo.refs[b"refs/heads/master"] = c3.id
self.repo.refs[b"refs/tags/foo"] = c3.id
target_path = tempfile.mkdtemp()
errstream = BytesIO()
self.addCleanup(shutil.rmtree, target_path)
r = porcelain.clone(self.repo.path,
target_path,
checkout=False,
errstream=errstream)
self.assertEqual(r.path, target_path)
target_repo = Repo(target_path)
self.assertEqual(target_repo.head(), c3.id)
self.assertEquals(c3.id, target_repo.refs[b'refs/tags/foo'])
self.assertTrue(b'f1' not in os.listdir(target_path))
self.assertTrue(b'f2' not in os.listdir(target_path))
def commit_style(repo: Repo, format_rule_name: str) -> Tuple[str, str]:
"""
Call bash script which commit all changes to `style_name` branch and checkout master back.
:param repo: Repo instance to the repository for which style were applied.
:param format_rule_name: Applied format rule name.
:return: Two commit hashes: where style was applied and where style was disrupt.
"""
def commit(repo: Repo, msg: str) -> str:
"""Commit everything."""
for tree_path, entry in repo.open_index().items():
full_path = os.path.join(repo.path.encode(), tree_path)
blob = blob_from_path_and_stat(full_path, os.lstat(full_path))
if blob.id != entry.sha:
repo.stage(tree_path)
return repo.do_commit(msg.encode(), b"Source{d} ML Team <*****@*****.**>")
repopath = repo.path
base = repo.head()
branch_create(repopath, format_rule_name, force=True)
update_head(repopath, format_rule_name)
style_commit_sha = commit(repo, format_rule_name)
build_index_from_tree(repo.path, repo.index_path(), repo.object_store, repo[base].tree)
revert_style_commit_sha = commit(repo, "Revert " + format_rule_name)
update_head(repopath, b"master")
return style_commit_sha.decode(), revert_style_commit_sha.decode()
async def info(self, ctx: Context):
"""
Get information about the bot
"""
embed = Embed(
description=
"A utility bot designed just for the Python server! Try `bot.help()` for more info.",
url="https://github.com/discord-python/bot")
repo = Repo(".")
sha = repo[repo.head()].sha().hexdigest()
embed.add_field(name="Total Users",
value=str(len(
self.bot.get_guild(PYTHON_GUILD).members)))
embed.add_field(name="Git SHA", value=str(sha)[:7])
embed.set_author(
name="Python Bot",
url="https://github.com/discord-python/bot",
icon_url=
"https://raw.githubusercontent.com/discord-python/branding/master/logos/logo_circle.png"
)
await ctx.send(embed=embed)
def test_simple_local(self):
f1_1 = make_object(Blob, data=b'f1')
commit_spec = [[1], [2, 1], [3, 1, 2]]
trees = {1: [(b'f1', f1_1), (b'f2', f1_1)],
2: [(b'f1', f1_1), (b'f2', f1_1)],
3: [(b'f1', f1_1), (b'f2', f1_1)], }
c1, c2, c3 = build_commit_graph(self.repo.object_store,
commit_spec, trees)
self.repo.refs[b"refs/heads/master"] = c3.id
self.repo.refs[b"refs/tags/foo"] = c3.id
target_path = tempfile.mkdtemp()
errstream = BytesIO()
self.addCleanup(shutil.rmtree, target_path)
r = porcelain.clone(self.repo.path, target_path,
checkout=False, errstream=errstream)
self.assertEqual(r.path, target_path)
target_repo = Repo(target_path)
self.assertEqual(target_repo.head(), c3.id)
self.assertEqual(c3.id, target_repo.refs[b'refs/tags/foo'])
self.assertTrue(b'f1' not in os.listdir(target_path))
self.assertTrue(b'f2' not in os.listdir(target_path))
c = r.get_config()
encoded_path = self.repo.path
if not isinstance(encoded_path, bytes):
encoded_path = encoded_path.encode('utf-8')
self.assertEqual(encoded_path, c.get((b'remote', b'origin'), b'url'))
self.assertEqual(
b'+refs/heads/*:refs/remotes/origin/*',
c.get((b'remote', b'origin'), b'fetch'))
def clone_role(url, branch, clone_root_path, clone_folder=None, depth=None):
""" Git clone
:param url: Source of the git repo
:param branch: Branch of the git repo
:param clone_root_path: The main folder in which the repo will be cloned.
:param clone_folder: The relative folder name of the git clone to the clone_root_path
:param depth(str): The git shallow clone depth
:returns: latest sha of the clone and its location
"""
gitcall = ["git", "clone"]
if depth and depth.isdigit():
gitcall.extend(["--depth", depth, "--no-single-branch"])
gitcall.append(url)
gitcall.extend(["-b", branch])
if not clone_folder:
clone_folder = url.split("/")[-1]
dirpath = os.path.join(clone_root_path, clone_folder)
gitcall.append(dirpath)
subprocess.check_call(gitcall)
repo = Repo(dirpath)
return repo.head(), dirpath
class Git():
"""
object that holds the git repository
"""
def __init__(self):
self.repo_path = user_data_dir(appname, appauthor)
self.files_under_version_controll = ['config.json', 'data.json']
# initialize repo if it doesn't exist
try:
self.repo = Repo(self.repo_path)
except NotGitRepository:
# create repo
if not os.path.exists(self.repo_path):
try:
os.makedirs(self.repo_path)
except OSError as exc: # Guard against race condition
if exc.errno != errno.EEXIST:
raise
Repo.init(self.repo_path)
self.repo = Repo(self.repo_path)
self.commit('initial commit')
def commit(self, message):
"""
commits the current status of files_under_version_controll
:param message: str; commit message
"""
self.repo.stage(self.files_under_version_controll)
self.repo.do_commit(str.encode(message), str.encode('nextSongs'))
def get_current_head(self):
"""
get sha as bytes of current head
:return: bytes; sha1 checksum of current head
"""
return self.repo.head()
def get_commits(self):
"""
generates a list of last commits
:return: list-of-dulwich.objects.Commit
"""
commits = []
for i in self.repo.get_walker():
commits.append(i.commit)
return reversed(
sorted(
commits,
key=lambda x: datetime.datetime.fromtimestamp(x.author_time)))
def restore(self, commit):
"""
does a hard reset to a given commit
:param commit: list-of-dulwich.objects.Commit; commit to reset to
"""
porcelain.reset(self.repo, 'hard',
str.encode(commit.sha().hexdigest()))
self.commit("Restored setting and data.")
Config.read_config()
def test_submodule(self):
temp_dir = tempfile.mkdtemp()
repo_dir = os.path.join(os.path.dirname(__file__), "data", "repos")
shutil.copytree(os.path.join(repo_dir, "a.git"), os.path.join(temp_dir, "a.git"), symlinks=True)
rel = os.path.relpath(os.path.join(repo_dir, "submodule"), temp_dir)
os.symlink(os.path.join(rel, "dotgit"), os.path.join(temp_dir, ".git"))
r = Repo(temp_dir)
self.assertEqual(r.head(), "a90fa2d900a17e99b433217e988c4eb4a2e9a097")
def get_last_commit(self):
try:
repo = Repo(os.path.dirname(self.path))
except NotGitRepository:
repo = Repo.init(os.path.dirname(self.path))
head = repo.head()
head_commit = repo.get_object(head)
return head_commit
class ArticleList(object):
def __init__(self):
self.repo = Repo('wiki')
self.head = self.repo.get_object(self.repo.head())
self.tree = self.repo.get_object(self.head.tree)
def get_article_titles(self):
return [a for a in self.tree]
def test_sprouted_tags(self):
path, gitsha = self.make_onerev_branch()
r = GitRepo(path)
r.refs[b"refs/tags/lala"] = r.head()
oldrepo = Repository.open(path)
revid = oldrepo.get_mapping().revision_id_foreign_to_bzr(gitsha)
newbranch = self.clone_git_branch(path, "f")
self.assertEqual({"lala": revid}, newbranch.tags.get_tag_dict())
self.assertEqual([revid], newbranch.repository.all_revision_ids())
def test_submodule(self):
temp_dir = tempfile.mkdtemp()
repo_dir = os.path.join(os.path.dirname(__file__), 'data', 'repos')
shutil.copytree(os.path.join(repo_dir, 'a.git'),
os.path.join(temp_dir, 'a.git'), symlinks=True)
rel = os.path.relpath(os.path.join(repo_dir, 'submodule'), temp_dir)
os.symlink(os.path.join(rel, 'dotgit'), os.path.join(temp_dir, '.git'))
r = Repo(temp_dir)
self.assertEqual(r.head(), 'a90fa2d900a17e99b433217e988c4eb4a2e9a097')
def _dulwich_commit(self, author, message=DEFAULT_COMMIT_MSG):
"""
Commit staged files in the repo
"""
_repo = Repo(self.config['top_dir'])
commit_id = _repo.do_commit(message, committer=author)
if not _repo.head() == commit_id:
raise SartorisError(message=exit_codes[14], exit_code=14)
def get_revision(self):
"""Read git revision information for the bcfg2 repository"""
try:
repo = Repo(self.datastore)
revision = repo.head()
except:
logger.error("Failed to read git repository; disabling git support")
raise Bcfg2.Server.Plugin.PluginInitError
return revision
def git_commit_info():
git = Repo('.')
commit = git.get_object(git.head())
return {
'id': commit.id.decode("utf-8")[0:7],
'id_full': commit.id.decode("utf-8"),
'author': regex.findall("(.*?) <(.*?)>",
commit.author.decode("utf-8"))[0],
'message': commit.message.decode("utf-8").strip('\r\n').split('\n')[0]
}
def test_submodule(self):
temp_dir = tempfile.mkdtemp()
repo_dir = os.path.join(os.path.dirname(__file__), 'data', 'repos')
shutil.copytree(os.path.join(repo_dir, 'a.git'),
os.path.join(temp_dir, 'a.git'),
symlinks=True)
rel = os.path.relpath(os.path.join(repo_dir, 'submodule'), temp_dir)
os.symlink(os.path.join(rel, 'dotgit'), os.path.join(temp_dir, '.git'))
r = Repo(temp_dir)
self.assertEqual(r.head(), 'a90fa2d900a17e99b433217e988c4eb4a2e9a097')
def get_revision(self):
"""Read git revision information for the Bcfg2 repository."""
try:
repo = Repo(self.datastore)
revision = repo.head()
except:
logger.error(
"Failed to read git repository; disabling git support")
raise Bcfg2.Server.Plugin.PluginInitError
return revision
class Article(object):
def __init__(self, title):
self.title = title.encode('UTF-8')
self.repo = Repo('wiki')
self.head = self.repo.get_object(self.repo.head())
self.tree = self.repo.get_object(self.head.tree)
try:
sha = self.tree[self.title][1]
self.content = self.repo[sha].data
except KeyError:
self.content = u'This space intentionally left blank.'
def __str__(self):
return self.title
def get_content(self):
return self.content
def get_title(self):
return self.title
def update_content(self, new_content, author, email, message):
new_content = new_content.encode('UTF-8')
author = author.encode('UTF-8')
message = message.encode('UTF-8')
email = email.encode('UTF-8')
# create blob, add to existing tree
blob = Blob.from_string(new_content)
self.tree[self.title] = (0100644, blob.id)
# commit
commit = Commit()
commit.tree = self.tree.id
commit.parents = [self.head.id]
commit.author = commit.committer = "%s <%s>" % (author, email)
commit.commit_time = commit.author_time = int(time())
tz = parse_timezone('+0100')[0] # FIXME: get proper timezone
commit.commit_timezone = commit.author_timezone = tz
commit.encoding = 'UTF-8'
commit.message = message
# save everything
object_store = self.repo.object_store
object_store.add_object(blob)
object_store.add_object(self.tree)
object_store.add_object(commit)
self.repo.refs['refs/heads/master'] = commit.id
def update_title(self, new_title):
pass
def get_source_revision_from_git(folder: Path) -> str:
dot_git: Path = folder / '.git'
if not dot_git.is_dir():
raise OSError('Not a Git working copy')
repo = Repo(folder)
# We truncate the hash string to 7 characters,
# that is also the way Git represents in short form.
try:
return repo.head()[:REVISION_LENGTH].decode()
except KeyError:
# Folder has just been "git init", no data yet
return '-'
def get_hash(self, _id, _user):
"""
Returns the hashes
:param _id: The _id of the repository
:return: The hash of the repository
"""
_result = self.node.find(
{"_id": ObjectId(_id)},
_user,
_error_prefix_if_not_allowed=
"This user doesn't have permissions for this repository.")
_repo = Repo(os.path.join(_result["folder"], str(_id)))
return _repo.head()
def test_submodule(self):
temp_dir = self.mkdtemp()
repo_dir = os.path.join(os.path.dirname(__file__), 'data', 'repos')
if isinstance(temp_dir, bytes):
temp_dir_str = temp_dir.decode(sys.getfilesystemencoding())
else:
temp_dir_str = temp_dir
shutil.copytree(os.path.join(repo_dir, 'a.git'),
os.path.join(temp_dir_str, 'a.git'), symlinks=True)
rel = os.path.relpath(os.path.join(repo_dir, 'submodule'), temp_dir_str)
os.symlink(os.path.join(rel, 'dotgit'), os.path.join(temp_dir_str, '.git'))
r = Repo(temp_dir)
self.assertEqual(r.head(), b'a90fa2d900a17e99b433217e988c4eb4a2e9a097')
def test_submodule(self):
temp_dir = self.mkdtemp()
repo_dir = os.path.join(os.path.dirname(__file__), 'data', 'repos')
if isinstance(temp_dir, bytes):
temp_dir_str = temp_dir.decode(sys.getfilesystemencoding())
else:
temp_dir_str = temp_dir
shutil.copytree(os.path.join(repo_dir, 'a.git'),
os.path.join(temp_dir_str, 'a.git'), symlinks=True)
rel = os.path.relpath(os.path.join(repo_dir, 'submodule'), temp_dir_str)
os.symlink(os.path.join(rel, 'dotgit'), os.path.join(temp_dir_str, '.git'))
r = Repo(temp_dir)
self.assertEqual(r.head(), b'a90fa2d900a17e99b433217e988c4eb4a2e9a097')
def teardown(self):
folder = Path.cwd()
try:
(folder / '.version').unlink()
except FileNotFoundError:
pass
if self.original_revision is None:
# Remove the Git repo created in test
try:
rp = Repo(str(folder))
head = rp[rp.head()]
if not len(head.parents):
shutil.rmtree(folder / '.git')
except NotGitRepository:
pass
def _dulwich_reset_to_tag(self, tag=None):
"""
Resets the HEAD to the commit
"""
_repo = Repo(self.config['top_dir'])
if not tag:
sha = _repo.head()
else:
sha = self._get_commit_sha_for_tag(tag)
try:
_repo.refs['HEAD'] = sha
except AttributeError:
raise GitMethodsError(message=exit_codes[7], exit_code=7)
def _dulwich_reset_to_tag(self, tag=None):
"""
Resets the HEAD to the commit
"""
_repo = Repo(self.config['top_dir'])
if not tag:
sha = _repo.head()
else:
sha = self._get_commit_sha_for_tag(tag)
try:
_repo.refs['HEAD'] = sha
except AttributeError:
raise GitMethodsError(message=exit_codes[7], exit_code=7)
def read_submodule_head(path):
"""Read the head commit of a submodule.
:param path: path to the submodule
:return: HEAD sha, None if not a valid head/repository
"""
from dulwich.errors import NotGitRepository
from dulwich.repo import Repo
try:
repo = Repo(path)
except NotGitRepository:
return None
try:
return repo.head()
except KeyError:
return None
def test_normal_from_repo(self):
# Create repo
folder = Path.cwd()
try:
rp = Repo(str(folder))
except NotGitRepository:
rp = Repo.init(str(folder))
try:
version = rp.head().decode()
self.original_revision = version
except KeyError:
FILE_NAME_TEST = 'file_test.txt'
test_file = folder / FILE_NAME_TEST
test_file.touch()
rp.stage(FILE_NAME_TEST.encode())
version = rp.do_commit(b'Test commit').decode()
v = get_source_revision()
assert v == version[:10]
def project_detail(request, project_slug):
"""Show commits for this project."""
repo, project = _get_repo(project_slug)
try:
repo = Repo(project.path_to_repo)
except:
return HttpResponse("Invalid repo path %s" % project.path_to_repo)
commits = repo.revision_history(repo.head())
branches = repo.refs.as_dict('refs/heads').keys()
return render_to_response('project/project_detail.html', {
'project': project,
'commits': commits,
'branches': branches,
}, context_instance=RequestContext(request))
def _identify(self,scribblearea):
top = self.select.module_name+"."+self.select.currentText()
try:
import os
if (os.getcwd()).endswith('stopeight'):
if (self.select.module_name=='stopeight.analyzer.file'):
if hasattr(scribblearea,'tablet_id'):
sub = str(scribblearea.tablet_id)
return (top,sub)
else:
return (top,'MouseData')
else:
from dulwich.repo import Repo
clibs_repo = Repo('../stopeight/')
if (self.select.module_name == ('stopeight_clibs_legacy')) or (self.select.module_name == ('stopeight_clibs_analyzer')):
clibs_repo = Repo('../stopeight-clibs/')
sub = (clibs_repo.head().decode('utf-8'))
return (top,sub)
except:
return (top)
def read_submodule_head(path):
"""Read the head commit of a submodule.
:param path: path to the submodule
:return: HEAD sha, None if not a valid head/repository
"""
from dulwich.errors import NotGitRepository
from dulwich.repo import Repo
# Repo currently expects a "str", so decode if necessary.
# TODO(jelmer): Perhaps move this into Repo() ?
if not isinstance(path, str):
path = path.decode(sys.getfilesystemencoding())
try:
repo = Repo(path)
except NotGitRepository:
return None
try:
return repo.head()
except KeyError:
return None
def read_submodule_head(path):
"""Read the head commit of a submodule.
:param path: path to the submodule
:return: HEAD sha, None if not a valid head/repository
"""
from dulwich.errors import NotGitRepository
from dulwich.repo import Repo
# Repo currently expects a "str", so decode if necessary.
# TODO(jelmer): Perhaps move this into Repo() ?
if not isinstance(path, str):
path = path.decode(sys.getfilesystemencoding())
try:
repo = Repo(path)
except NotGitRepository:
return None
try:
return repo.head()
except KeyError:
return None
def readme(self, date=True, git_commit=False, git_path=".", info=()):
"""Adds a README.md with useful info.
The README consists of the name of the directory (passed in at init
time), followed by a bulleted list with various pieces of info.
Args:
date (bool): Add the date and time in the bulleted list of info.
git_commit (bool): Add the current git commit hash in the bulleted
list of info.
git_path (str or pathlib.Path): The path to the git repo (i.e. a
directory that contains `.git`). Only applicable if `git_commit`
is True. If the path given is not to a Git repo, a warning is
issued, and the Git Commit is replaced with `"(no repo found)"`
info (list of str): A list of additional bullets to add in the
README.
Returns:
pathlib.path: Full path to the README.
"""
readme_path = self.pfile("README.md")
with readme_path.open("w") as file:
lines = [f"# {self._name}", ""]
if date:
date_str = self._datetime.strftime("%Y-%m-%d %H:%M:%S")
lines.append(f"- Date: {date_str}")
if git_commit:
git_path = Path(git_path)
if (git_path / ".git").exists():
repo = Repo(str(git_path))
commit_hash = repo.head().decode("utf-8")
else:
warnings.warn("No Git repo found")
commit_hash = "(no repo found)"
lines.append(f"- Git Commit: {commit_hash}")
lines.extend(map(lambda s: f"- {s}", info))
file.write("\n".join(lines) + "\n")
return readme_path
def git_prelod(abbr):
if not hasattr(settings, "ENABLE_GIT") or not settings.ENABLE_GIT:
return
global git_active_repo
global git_active_commit
global git_active_tree
global git_old_tree
global HEAD
gitdir = "%s/%s.git" % (settings.GIT_PATH, abbr)
if not os.path.exists(gitdir):
git_repo_init(gitdir)
git_active_repo = Repo(gitdir)
git_active_commit = Commit()
HEAD = git_active_repo.head()
commit = git_active_repo.commit(HEAD)
tree = git_active_repo.tree(commit.tree)
git_old_tree = tree.id
git_active_tree = tree
def __init__(self, path, init = False):
if init :
os.makedirs(path)
repo = Repo.init_bare(path)
else:
repo = Repo(path)
self._repo = repo
self.path = path
self.branches = self._getBranches()
try:
head = repo.head()
commitLast = repo.commit(head)
self.head = {
'sha' : head,
'tree' : commitLast.tree
}
except KeyError:
self.head = None
def read_submodule_head(path):
"""Read the head commit of a submodule.
Args:
path: path to the submodule
Returns: HEAD sha, None if not a valid head/repository
"""
from dulwich.errors import NotGitRepository
from dulwich.repo import Repo
# Repo currently expects a "str", so decode if necessary.
# TODO(user): Perhaps move this into Repo() ?
if not isinstance(path, str):
path = os.fsdecode(path)
try:
repo = Repo(path)
except NotGitRepository:
return None
try:
return repo.head()
except KeyError:
return None
def dulwichCommit(self, filePath, fullPath, kind): git = Repo(AUTOGIT_PATH) staged = map(str,[filePath]) git.stage( staged ) index = git.open_index() try: committer = git._get_user_identity() except ValueError: committer = "autogit" try: head = git.head() except KeyError: return git.do_commit( '%s - autogit commit (via dulwich)' % kind, committer=committer) changes = list(tree_changes(git, index.commit(git.object_store), git['HEAD'].tree)) if changes and len(changes) > 0: return git.do_commit( '%s - autogit commit (via dulwich)' % kind, committer=committer) return None
def git_prelod(abbr):
if not hasattr(settings, "ENABLE_GIT") or not settings.ENABLE_GIT:
return
global git_active_repo
global git_active_commit
global git_active_tree
global git_old_tree
global HEAD
gitdir = "%s/%s.git" % (settings.GIT_PATH, abbr)
if not os.path.exists(gitdir):
git_repo_init(gitdir)
git_active_repo = Repo(gitdir)
git_active_commit = Commit()
HEAD = git_active_repo.head()
commit = git_active_repo.commit(HEAD)
tree = git_active_repo.tree(commit.tree)
git_old_tree = tree.id
git_active_tree = tree
def test_simple_local(self):
f1_1 = make_object(Blob, data=b"f1")
commit_spec = [[1], [2, 1], [3, 1, 2]]
trees = {
1: [(b"f1", f1_1), (b"f2", f1_1)],
2: [(b"f1", f1_1), (b"f2", f1_1)],
3: [(b"f1", f1_1), (b"f2", f1_1)],
}
c1, c2, c3 = build_commit_graph(self.repo.object_store, commit_spec, trees)
self.repo.refs[b"refs/heads/master"] = c3.id
self.repo.refs[b"refs/tags/foo"] = c3.id
target_path = tempfile.mkdtemp()
errstream = BytesIO()
self.addCleanup(shutil.rmtree, target_path)
r = porcelain.clone(self.repo.path, target_path, checkout=False, errstream=errstream)
self.assertEqual(r.path, target_path)
target_repo = Repo(target_path)
self.assertEqual(target_repo.head(), c3.id)
self.assertEqual(c3.id, target_repo.refs[b"refs/tags/foo"])
self.assertTrue(b"f1" not in os.listdir(target_path))
self.assertTrue(b"f2" not in os.listdir(target_path))
class GitAPI(GitAPIBase):
""" API for :class:`Git` using :mod:`dulwich` """
def __init__(self, path):
GitAPIBase.__init__(self, path)
self.repo = Repo(self.path)
self.client, self.origin_path = get_transport_and_path(
self.repo.get_config().get(("remote", "origin"),
"url"))
def revision(self):
return self.repo.head()
def pull(self):
try:
remote_refs = self.client.fetch(
self.origin_path, self.repo,
determine_wants=self.repo.object_store.determine_wants_all)
except KeyError:
etype, err = sys.exc_info()[:2]
# try to work around bug
# https://bugs.launchpad.net/dulwich/+bug/1025886
try:
# pylint: disable=W0212
self.client._fetch_capabilities.remove('thin-pack')
# pylint: enable=W0212
except KeyError:
raise etype(err)
remote_refs = self.client.fetch(
self.origin_path, self.repo,
determine_wants=self.repo.object_store.determine_wants_all)
tree_id = self.repo[remote_refs['HEAD']].tree
# iterate over tree content, giving path and blob sha.
for entry in self.repo.object_store.iter_tree_contents(tree_id):
entry_in_path = entry.in_path(self.repo.path)
ensure_dir_exists(os.path.split(entry_in_path.path)[0])
GitFile(entry_in_path.path,
'wb').write(self.repo[entry.sha].data)
def get_sha_and_dirtiness(
prompt_on_dirty: bool = True) -> Optional[Tuple[str, bool]]:
try:
git_status = status()
except NotGitRepository:
return None
dirty = False
def to_str(string: Union[str, bytes]) -> str:
if isinstance(string, str):
return string
return string.decode("utf8")
def print_files(filenames: Iterable[Union[str, bytes]]) -> None:
print("\n".join(f" - {to_str(filename)}" for filename in filenames))
if git_status.untracked:
print("Those files are untracked:", file=stderr)
print_files(git_status.untracked)
dirty = True
if git_status.unstaged:
print("Those files are unstaged:", file=stderr)
print_files(git_status.unstaged)
dirty = True
if any(git_status.staged.values()):
print("Those files are uncommited:", file=stderr)
print_files(chain(*git_status.staged.values()))
dirty = True
if dirty:
print("Are you sure you want to continue [y/n]? ", end="")
answer = input()
if answer != "y":
exit(1)
repo = Repo(".")
sha = to_str(repo.head())
return sha, dirty
class GitStorageHistory(object):
def __init__(self, repodir, storage_root_subdir=os.curdir):
self.repo = Repo(repodir)
self.storage_root_subdir = storage_root_subdir
def __getitem__(self, commit_sha):
commit = self.repo.commit(commit_sha)
tree = self.repo.tree(commit.tree)
return GitTreeStorage(self.repo, tree, self.storage_root_subdir)
@property
def latest_revision(self):
return self.repo.head()
def eod_revisions(self, max_revision=None):
if max_revision is None:
max_revision = self.latest_revision
results = {}
commit_shas = set([max_revision])
while commit_shas:
commit_sha = commit_shas.pop()
commit = self.repo.commit(commit_sha)
commit_date = date_of(commit)
commit_info = (time_of(commit), commit_sha)
if commit_date in results:
results[commit_date] = max(results[commit_date], commit_info)
else:
results[commit_date] = commit_info
commit_shas.update(commit.parents)
return dict((date, sha) for (date, (time, sha)) in results.iteritems())
class Gittle(object):
"""All paths used in Gittle external methods must be paths relative to the git repository
"""
DEFAULT_COMMIT = "HEAD"
DEFAULT_BRANCH = "master"
DEFAULT_REMOTE = "origin"
DEFAULT_MESSAGE = "**No Message**"
DEFAULT_USER_INFO = {"name": None, "email": None}
DIFF_FUNCTIONS = {
"classic": utils.git.classic_tree_diff,
"dict": utils.git.dict_tree_diff,
"changes": utils.git.dict_tree_diff,
}
DEFAULT_DIFF_TYPE = "dict"
HIDDEN_REGEXES = [
# Hide git directory
r".*\/\.git\/.*"
]
# References
REFS_BRANCHES = "refs/heads/"
REFS_REMOTES = "refs/remotes/"
REFS_TAGS = "refs/tags/"
# Name pattern truths
# Used for detecting if files are :
# - deleted
# - added
# - changed
PATTERN_ADDED = (False, True)
PATTERN_REMOVED = (True, False)
PATTERN_MODIFIED = (True, True)
# Permissions
MODE_DIRECTORY = 040000 # Used to tell if a tree entry is a directory
# Tree depth
MAX_TREE_DEPTH = 1000
# Acceptable Root paths
ROOT_PATHS = (os.path.curdir, os.path.sep)
def __init__(self, repo_or_path, origin_uri=None, auth=None, report_activity=None, *args, **kwargs):
if isinstance(repo_or_path, DulwichRepo):
self.repo = repo_or_path
elif isinstance(repo_or_path, Gittle):
self.repo = DulwichRepo(repo_or_path.path)
elif isinstance(repo_or_path, basestring):
path = os.path.abspath(repo_or_path)
self.repo = DulwichRepo(path)
else:
logging.warning("Repo is of type %s" % type(repo_or_path))
raise Exception("Gittle must be initialized with either a dulwich repository or a string to the path")
# Set path
self.path = self.repo.path
# The remote url
self.origin_uri = origin_uri
# Report client activty
self._report_activity = report_activity
# Build ignore filter
self.hidden_regexes = copy.copy(self.HIDDEN_REGEXES)
self.hidden_regexes.extend(self._get_ignore_regexes())
self.ignore_filter = utils.paths.path_filter_regex(self.hidden_regexes)
self.filters = [self.ignore_filter]
# Get authenticator
if auth:
self.authenticator = auth
else:
self.auth(*args, **kwargs)
def report_activity(self, *args, **kwargs):
if not self._report_activity:
return
return self._report_activity(*args, **kwargs)
def _format_author(self, name, email):
return "%s <%s>" % (name, email)
def _format_userinfo(self, userinfo):
name = userinfo.get("name")
email = userinfo.get("email")
if name and email:
return self._format_author(name, email)
return None
def _format_ref(self, base, extra):
return "".join([base, extra])
def _format_ref_branch(self, branch_name):
return self._format_ref(self.REFS_BRANCHES, branch_name)
def _format_ref_remote(self, remote_name):
return self._format_ref(self.REFS_REMOTES, remote_name)
def _format_ref_tag(self, tag_name):
return self._format_ref(self.REFS_TAGS, tag_name)
@property
def head(self):
"""Return SHA of the current HEAD
"""
return self.repo.head()
@property
def is_bare(self):
"""Bare repositories have no working directories or indexes
"""
return self.repo.bare
@property
def is_working(self):
return not (self.is_bare)
def has_index(self):
"""Opposite of is_bare
"""
return self.repo.has_index()
@property
def has_commits(self):
"""
If the repository has no HEAD we consider that is has no commits
"""
try:
self.repo.head()
except KeyError:
return False
return True
def ref_walker(self, ref=None):
"""
Very simple, basic walker
"""
ref = ref or "HEAD"
sha = self._commit_sha(ref)
return self.repo.revision_history(sha)
def branch_walker(self, branch):
branch = branch or self.DEFAULT_BRANCH
ref = self._format_ref_branch(branch)
return self.ref_walker(ref)
def commit_info(self, start=0, end=None, branch=None):
"""Return a generator of commits with all their attached information
"""
if not self.has_commits:
return []
commits = [utils.git.commit_info(entry) for entry in self.branch_walker(branch)]
if not end:
return commits
return commits[start:end]
@funky.uniquify
def recent_contributors(self, n=None, branch=None):
n = n or 10
return funky.pluck(self.commit_info(end=n, branch=branch), "author")
@property
def commit_count(self):
try:
return len(self.ref_walker())
except KeyError:
return 0
def commits(self):
"""Return a list of SHAs for all the concerned commits
"""
return [commit["sha"] for commit in self.commit_info()]
@property
def git_dir(self):
return self.repo.controldir()
def auth(self, *args, **kwargs):
self.authenticator = GittleAuth(*args, **kwargs)
return self.authenticator
# Generate a branch selector (used for pushing)
def _wants_branch(self, branch_name=None):
branch_name = branch_name or self.DEFAULT_BRANCH
refs_key = self._format_ref_branch(branch_name)
sha = self.branches[branch_name]
def wants_func(old):
refs_key = self._format_ref_branch(branch_name)
return {refs_key: sha}
return wants_func
def _get_ignore_regexes(self):
gitignore_filename = os.path.join(self.path, ".gitignore")
if not os.path.exists(gitignore_filename):
return []
lines = open(gitignore_filename).readlines()
globers = map(lambda line: line.rstrip(), lines)
return utils.paths.globers_to_regex(globers)
# Get the absolute path for a file in the git repo
def abspath(self, repo_file):
return os.path.abspath(os.path.join(self.path, repo_file))
# Get the relative path from the absolute path
def relpath(self, abspath):
return os.path.relpath(abspath, self.path)
@property
def last_commit(self):
return self[self.repo.head()]
@property
def index(self):
return self.repo.open_index()
@classmethod
def init(cls, path, bare=None, *args, **kwargs):
"""Initialize a repository"""
mkdir_safe(path)
# Constructor to use
if bare:
constructor = DulwichRepo.init_bare
else:
constructor = DulwichRepo.init
# Create dulwich repo
repo = constructor(path)
# Create Gittle repo
return cls(repo, *args, **kwargs)
@classmethod
def init_bare(cls, *args, **kwargs):
kwargs.setdefault("bare", True)
return cls.init(*args, **kwargs)
def get_client(self, origin_uri=None, **kwargs):
# Get the remote URL
origin_uri = origin_uri or self.origin_uri
# Fail if inexistant
if not origin_uri:
raise InvalidRemoteUrl()
client_kwargs = {}
auth_kwargs = self.authenticator.kwargs()
client_kwargs.update(auth_kwargs)
client_kwargs.update(kwargs)
client_kwargs.update({"report_activity": self.report_activity})
client, remote_path = get_transport_and_path(origin_uri, **client_kwargs)
return client, remote_path
def push_to(self, origin_uri, branch_name=None, progress=None, progress_stderr=None):
selector = self._wants_branch(branch_name=branch_name)
client, remote_path = self.get_client(origin_uri, progress_stderr=progress_stderr)
return client.send_pack(remote_path, selector, self.repo.object_store.generate_pack_contents, progress=progress)
# Like: git push
def push(self, origin_uri=None, branch_name=None, progress=None, progress_stderr=None):
return self.push_to(origin_uri, branch_name, progress, progress_stderr)
# Not recommended at ALL ... !!!
def dirty_pull_from(self, origin_uri, branch_name=None):
# Remove all previously existing data
rmtree(self.path)
mkdir_safe(self.path)
self.repo = DulwichRepo.init(self.path)
# Fetch brand new copy from remote
return self.pull_from(origin_uri, branch_name)
def pull_from(self, origin_uri, branch_name=None):
return self.fetch(origin_uri)
# Like: git pull
def pull(self, origin_uri=None, branch_name=None):
return self.pull_from(origin_uri, branch_name)
def fetch_remote(self, origin_uri=None):
# Get client
client, remote_path = self.get_client(origin_uri=origin_uri)
# Fetch data from remote repository
remote_refs = client.fetch(remote_path, self.repo)
return remote_refs
def _setup_fetched_refs(self, refs, origin, bare):
remote_tags = utils.git.subrefs(refs, "refs/tags")
remote_heads = utils.git.subrefs(refs, "refs/heads")
# Filter refs
clean_remote_tags = utils.git.clean_refs(remote_tags)
clean_remote_heads = utils.git.clean_refs(remote_heads)
# Base of new refs
heads_base = "refs/remotes/" + origin
if bare:
heads_base = "refs/heads"
# Import branches
self.import_refs(heads_base, clean_remote_heads)
# Import tags
self.import_refs("refs/tags", clean_remote_tags)
# Update HEAD
self["HEAD"] = refs["HEAD"]
def fetch(self, origin_uri=None, bare=None, origin=None):
bare = bare or False
origin = origin or self.DEFAULT_REMOTE
# Remote refs
remote_refs = self.fetch_remote(origin_uri)
# Update head
# Hit repo because head doesn't yet exist so
# print("REFS = %s" % remote_refs)
# Update refs (branches, tags, HEAD)
self._setup_fetched_refs(remote_refs, origin, bare)
# Checkout working directories
if not bare:
self.checkout_all()
else:
self.update_server_info()
@classmethod
def clone(cls, origin_uri, local_path, auth=None, mkdir=True, bare=False, *args, **kwargs):
"""Clone a remote repository"""
mkdir_safe(local_path)
# Initialize the local repository
if bare:
local_repo = cls.init_bare(local_path)
else:
local_repo = cls.init(local_path)
repo = cls(local_repo, origin_uri=origin_uri, auth=auth, *args, **kwargs)
repo.fetch(bare=bare)
# Add origin
# TODO
return repo
@classmethod
def clone_bare(cls, *args, **kwargs):
"""Same as .clone except clones to a bare repository by default
"""
kwargs.setdefault("bare", True)
return cls.clone(*args, **kwargs)
def _commit(self, committer=None, author=None, message=None, files=None, tree=None, *args, **kwargs):
if not tree:
# If no tree then stage files
modified_files = files or self.modified_files
logging.warning("STAGING : %s" % modified_files)
self.add(modified_files)
# Messages
message = message or self.DEFAULT_MESSAGE
author_msg = self._format_userinfo(author)
committer_msg = self._format_userinfo(committer)
return self.repo.do_commit(
message=message, author=author_msg, committer=committer_msg, encoding="UTF-8", tree=tree, *args, **kwargs
)
def _tree_from_structure(self, structure):
# TODO : Support directories
tree = Tree()
for file_info in structure:
# str only
try:
data = file_info["data"].encode("ascii")
name = file_info["name"].encode("ascii")
mode = file_info["mode"]
except:
# Skip file on encoding errors
continue
blob = Blob()
blob.data = data
# Store file's contents
self.repo.object_store.add_object(blob)
# Add blob entry
tree.add(name, mode, blob.id)
# Store tree
self.repo.object_store.add_object(tree)
return tree.id
# Like: git commmit -a
def commit(self, name=None, email=None, message=None, files=None, *args, **kwargs):
user_info = {"name": name, "email": email}
return self._commit(committer=user_info, author=user_info, message=message, files=files, *args, **kwargs)
def commit_structure(self, name=None, email=None, message=None, structure=None, *args, **kwargs):
"""Main use is to do commits directly to bare repositories
For example doing a first Initial Commit so the repo can be cloned and worked on right away
"""
if not structure:
return
tree = self._tree_from_structure(structure)
user_info = {"name": name, "email": email}
return self._commit(committer=user_info, author=user_info, message=message, tree=tree, *args, **kwargs)
# Push all local commits
# and pull all remote commits
def sync(self, origin_uri=None):
self.push(origin_uri)
return self.pull(origin_uri)
def lookup_entry(self, relpath, trackable_files=set()):
if not relpath in trackable_files:
raise KeyError
abspath = self.abspath(relpath)
with open(abspath, "rb") as git_file:
data = git_file.read()
s = sha1()
s.update("blob %u\0" % len(data))
s.update(data)
return (s.hexdigest(), os.stat(abspath).st_mode)
@property
@funky.transform(set)
def tracked_files(self):
return list(self.index)
@property
@funky.transform(set)
def raw_files(self):
return utils.paths.subpaths(self.path)
@property
@funky.transform(set)
def ignored_files(self):
return utils.paths.subpaths(self.path, filters=self.filters)
@property
@funky.transform(set)
def trackable_files(self):
return self.raw_files - self.ignored_files
@property
@funky.transform(set)
def untracked_files(self):
return self.trackable_files - self.tracked_files
"""
@property
@funky.transform(set)
def modified_staged_files(self):
"Checks if the file has changed since last commit"
timestamp = self.last_commit.commit_time
index = self.index
return [
f
for f in self.tracked_files
if index[f][1][0] > timestamp
]
"""
# Return a list of tuples
# representing the changed elements in the git tree
def _changed_entries(self, ref=None):
ref = ref or self.DEFAULT_COMMIT
if not self.has_commits:
return []
obj_sto = self.repo.object_store
tree_id = self[ref].tree
names = self.trackable_files
lookup_func = partial(self.lookup_entry, trackable_files=names)
# Format = [((old_name, new_name), (old_mode, new_mode), (old_sha, new_sha)), ...]
tree_diff = changes_from_tree(names, lookup_func, obj_sto, tree_id, want_unchanged=False)
return list(tree_diff)
@funky.transform(set)
def _changed_entries_by_pattern(self, pattern):
changed_entries = self._changed_entries()
filtered_paths = [
funky.first_true(names)
for names, modes, sha in changed_entries
if tuple(map(bool, names)) == pattern and funky.first_true(names)
]
return filtered_paths
@property
@funky.transform(set)
def removed_files(self):
return self._changed_entries_by_pattern(self.PATTERN_REMOVED) - self.ignored_files
@property
@funky.transform(set)
def added_files(self):
return self._changed_entries_by_pattern(self.PATTERN_ADDED) - self.ignored_files
@property
@funky.transform(set)
def modified_files(self):
modified_files = self._changed_entries_by_pattern(self.PATTERN_MODIFIED) - self.ignored_files
return modified_files
@property
@funky.transform(set)
def modified_unstaged_files(self):
timestamp = self.last_commit.commit_time
return [f for f in self.tracked_files if os.stat(self.abspath(f)).st_mtime > timestamp]
@property
def pending_files(self):
"""
Returns a list of all files that could be possibly staged
"""
# Union of both
return self.modified_files | self.added_files | self.removed_files
@property
def pending_files_by_state(self):
files = {"modified": self.modified_files, "added": self.added_files, "removed": self.removed_files}
# "Flip" the dictionary
return {path: state for state, paths in files.items() for path in paths}
"""
@property
@funky.transform(set)
def modified_files(self):
return self.modified_staged_files | self.modified_unstaged_files
"""
# Like: git add
@funky.arglist_method
def stage(self, files):
return self.repo.stage(files)
def add(self, *args, **kwargs):
return self.stage(*args, **kwargs)
# Like: git rm
@funky.arglist_method
def rm(self, files, force=False):
index = self.index
index_files = filter(lambda f: f in index, files)
for f in index_files:
del self.index[f]
return index.write()
def mv_fs(self, file_pair):
old_name, new_name = file_pair
os.rename(old_name, new_name)
# Like: git mv
@funky.arglist_method
def mv(self, files_pair):
index = self.index
files_in_index = filter(lambda f: f[0] in index, files_pair)
map(self.mv_fs, files_in_index)
old_files = map(funky.first, files_in_index)
new_files = map(funky.last, files_in_index)
self.add(new_files)
self.rm(old_files)
self.add(old_files)
return
@working_only
def _checkout_tree(self, tree):
return build_index_from_tree(self.repo.path, self.repo.index_path(), self.repo.object_store, tree)
def checkout_all(self, commit_sha=None):
commit_sha = commit_sha or self.head
commit_tree = self._commit_tree(commit_sha)
# Rebuild index from the current tree
return self._checkout_tree(commit_tree)
def checkout(self, commit_sha=None, files=None):
"""Checkout only a select amount of files
"""
commit_sha = commit_sha or self.head
files = files or []
return self
@funky.arglist_method
def reset(self, files, commit="HEAD"):
pass
def rm_all(self):
self.index.clear()
return self.index.write()
def _to_commit(self, commit_obj):
"""Allows methods to accept both SHA's or dulwich Commit objects as arguments
"""
if isinstance(commit_obj, basestring):
return self.repo[commit_obj]
return commit_obj
def _commit_sha(self, commit_obj):
"""Extracts a Dulwich commits SHA
"""
if utils.git.is_sha(commit_obj):
return commit_obj
elif isinstance(commit_obj, basestring):
# Can't use self[commit_obj] to avoid infinite recursion
commit_obj = self.repo[commit_obj]
return commit_obj.id
def _blob_data(self, sha):
"""Return a blobs content for a given SHA
"""
return self[sha].data
# Get the nth parent back for a given commit
def get_parent_commit(self, commit, n=None):
""" Recursively gets the nth parent for a given commit
Warning: Remember that parents aren't the previous commits
"""
if n is None:
n = 1
commit = self._to_commit(commit)
parents = commit.parents
if n <= 0 or not parents:
# Return a SHA
return self._commit_sha(commit)
parent_sha = parents[0]
parent = self[parent_sha]
# Recur
return self.get_parent_commit(parent, n - 1)
def get_previous_commit(self, commit_ref, n=None):
commit_sha = self._parse_reference(commit_ref)
n = n or 1
commits = self.commits()
return funky.next(commits, commit_sha, n=n, default=commit_sha)
def _parse_reference(self, ref_string):
# COMMIT_REF~x
if "~" in ref_string:
ref, count = ref_string.split("~")
count = int(count)
commit_sha = self._commit_sha(ref)
return self.get_previous_commit(commit_sha, count)
return self._commit_sha(ref_string)
def _commit_tree(self, commit_sha):
"""Return the tree object for a given commit
"""
return self[commit_sha].tree
def diff(self, commit_sha, compare_to=None, diff_type=None, filter_binary=True):
diff_type = diff_type or self.DEFAULT_DIFF_TYPE
diff_func = self.DIFF_FUNCTIONS[diff_type]
if not compare_to:
compare_to = self.get_previous_commit(commit_sha)
return self._diff_between(compare_to, commit_sha, diff_function=diff_func)
def diff_working(self, ref=None, filter_binary=True):
"""Diff between the current working directory and the HEAD
"""
return utils.git.diff_changes_paths(
self.repo.object_store, self.path, self._changed_entries(ref=ref), filter_binary=filter_binary
)
def get_commit_files(self, commit_sha, parent_path=None, is_tree=None, paths=None):
"""Returns a dict of the following Format :
{
"directory/filename.txt": {
'name': 'filename.txt',
'path': "directory/filename.txt",
"sha": "xxxxxxxxxxxxxxxxxxxx",
"data": "blablabla",
"mode": 0xxxxx",
},
...
}
"""
# Default values
context = {}
is_tree = is_tree or False
parent_path = parent_path or ""
if is_tree:
tree = self[commit_sha]
else:
tree = self[self._commit_tree(commit_sha)]
for mode, path, sha in tree.entries():
# Check if entry is a directory
if mode == self.MODE_DIRECTORY:
context.update(
self.get_commit_files(sha, parent_path=os.path.join(parent_path, path), is_tree=True, paths=paths)
)
continue
subpath = os.path.join(parent_path, path)
# Only add the files we want
if not (paths is None or subpath in paths):
continue
# Add file entry
context[subpath] = {"name": path, "path": subpath, "mode": mode, "sha": sha, "data": self._blob_data(sha)}
return context
def file_versions(self, path):
"""Returns all commits where given file was modified
"""
versions = []
commits_info = self.commit_info()
seen_shas = set()
for commit in commits_info:
try:
files = self.get_commit_files(commit["sha"], paths=[path])
file_path, file_data = files.items()[0]
except IndexError:
continue
file_sha = file_data["sha"]
if file_sha in seen_shas:
continue
else:
seen_shas.add(file_sha)
# Add file info
commit["file"] = file_data
versions.append(file_data)
return versions
def _diff_between(self, old_commit_sha, new_commit_sha, diff_function=None, filter_binary=True):
"""Internal method for getting a diff between two commits
Please use .diff method unless you have very speciic needs
"""
# If commit is first commit (new_commit_sha == old_commit_sha)
# then compare to an empty tree
if new_commit_sha == old_commit_sha:
old_tree = Tree()
else:
old_tree = self._commit_tree(old_commit_sha)
new_tree = self._commit_tree(new_commit_sha)
return diff_function(self.repo.object_store, old_tree, new_tree, filter_binary=filter_binary)
def changes(self, *args, **kwargs):
""" List of changes between two SHAs
Returns a list of lists of tuples :
[
[
(oldpath, newpath), (oldmode, newmode), (oldsha, newsha)
],
...
]
"""
kwargs["diff_type"] = "changes"
return self.diff(*args, **kwargs)
def changes_count(self, *args, **kwargs):
return len(self.changes(*args, **kwargs))
def _refs_by_pattern(self, pattern):
refs = self.refs
def item_filter(key_value):
"""Filter only concered refs"""
key, value = key_value
return key.startswith(pattern)
def item_map(key_value):
"""Rewrite keys"""
key, value = key_value
new_key = key[len(pattern) :]
return (new_key, value)
return dict(map(item_map, filter(item_filter, refs.items())))
@property
def refs(self):
return self.repo.get_refs()
def set_refs(refs_dict):
for k, v in refs_dict.items():
self.repo[k] = v
def import_refs(self, base, other):
return self.repo.refs.import_refs(base, other)
@property
def branches(self):
return self._refs_by_pattern(self.REFS_BRANCHES)
def _active_branch(self, refs=None, head=None):
head = head or self.head
refs = refs or self.branches
try:
return {branch: branch_head for branch, branch_head in refs.items() if branch_head == head}.items()[0]
except IndexError:
pass
return (None, None)
@property
def active_branch(self):
return self._active_branch()[0]
@property
def active_sha(self):
return self._active_branch()[1]
@property
def remote_branches(self):
return self._refs_by_pattern(self.REFS_REMOTES)
@property
def tags(self):
return self._refs_by_pattern(self.REFS_TAGS)
@property
def remotes(self):
""" Dict of remotes
{
'origin': 'http://friendco.de/some_user/repo.git',
...
}
"""
config = self.repo.get_config()
return {keys[1]: values["url"] for keys, values in config.items() if keys[0] == "remote"}
def add_ref(self, new_ref, old_ref):
self.repo.refs[new_ref] = self.repo.refs[old_ref]
self.update_server_info()
def remove_ref(self, ref_name):
# Returns False if ref doesn't exist
if not ref_name in self.repo.refs:
return False
del self.repo.refs[ref_name]
self.update_server_info()
return True
def create_branch(self, base_branch, new_branch, tracking=None):
"""Try creating a new branch which tracks the given remote
if such a branch does not exist then branch off a local branch
"""
# The remote to track
tracking = self.DEFAULT_REMOTE
# Already exists
if new_branch in self.branches:
raise Exception("branch %s already exists" % new_branch)
# Get information about remote_branch
remote_branch = os.path.sep.join([tracking, base_branch])
# Fork Local
if base_branch in self.branches:
base_ref = self._format_ref_branch(base_branch)
# Fork remote
elif remote_branch in self.remote_branches:
base_ref = self._format_ref_remote(remote_branch)
# TODO : track
else:
raise Exception(
"Can not find the branch named '%s' to fork either locally or in '%s'" % (base_branch, tracking)
)
# Reference of new branch
new_ref = self._format_ref_branch(new_branch)
# Copy reference to create branch
self.add_ref(new_ref, base_ref)
return new_ref
def remove_branch(self, branch_name):
ref = self._format_ref_branch(branch_name)
return self.remove_ref(ref)
def switch_branch(self, branch_name, tracking=None, create=None):
"""Changes the current branch
"""
if create is None:
create = True
# Check if branch exists
if not branch_name in self.branches:
self.create_branch(branch_name, branch_name, tracking=tracking)
# Get branch reference
branch_ref = self._format_ref_branch(branch_name)
# Change main branch
self.repo.refs.set_symbolic_ref("HEAD", branch_ref)
if self.is_working:
# Remove all files
self.clean_working()
# Add files for the current branch
self.checkout_all()
def clean(self, force=None, directories=None):
untracked_files = self.untracked_files
map(os.remove, untracked_files)
return untracked_files
def clean_working(self):
"""Purges all the working (removes everything except .git)
used by checkout_all to get clean branch switching
"""
return self.clean()
def _get_fs_structure(self, tree_sha, depth=None, parent_sha=None):
tree = self[tree_sha]
structure = {}
if depth is None:
depth = self.MAX_TREE_DEPTH
elif depth == 0:
return structure
for mode, path, sha in tree.entries():
# tree
if mode == self.MODE_DIRECTORY:
# Recur
structure[path] = self._get_fs_structure(sha, depth=depth - 1, parent_sha=tree_sha)
# commit
else:
structure[path] = sha
structure["."] = tree_sha
structure[".."] = parent_sha or tree_sha
return structure
def _get_fs_structure_by_path(self, tree_sha, path):
parts = path.split(os.path.sep)
depth = len(parts) + 1
structure = self._get_fs_structure(tree_sha, depth=depth)
return funky.subkey(structure, parts)
def commit_ls(self, ref, subpath=None):
"""List a "directory" for a given commit
using the tree of thqt commit
"""
tree_sha = self._commit_tree(ref)
# Root path
if subpath in self.ROOT_PATHS or not subpath:
return self._get_fs_structure(tree_sha, depth=1)
# Any other path
return self._get_fs_structure_by_path(tree_sha, subpath)
def commit_file(self, ref, path):
"""Return info on a given file for a given commit
"""
name, info = self.get_commit_files(ref, paths=[path]).items()[0]
return info
def commit_tree(self, ref, *args, **kwargs):
tree_sha = self._commit_tree(ref)
return self._get_fs_structure(tree_sha, *args, **kwargs)
def update_server_info(self):
if not self.is_bare:
return
update_server_info(self.repo)
def _is_fast_forward(self):
pass
def _merge_fast_forward(self):
pass
def __hash__(self):
"""This is required otherwise the memoize function will just mess it up
"""
return hash(self.path)
def __getitem__(self, key):
sha = self._parse_reference(key)
return self.repo[sha]
def __setitem__(self, key, value):
self.repo[key] = value
# Alias to clone_bare
fork = clone_bare
log = commit_info
diff_count = changes_count
comtributors = recent_contributors
#!/usr/bin/env python2
import os.path
import urlparse
from email.utils import formatdate
from dulwich.repo import Repo
from dulwich.objects import Blob, Tree, Commit
from docutils import io, nodes
from docutils.core import publish_doctree, publish_from_doctree
from render import MyWriter
repo = Repo(".")
commit_sha = repo.head()
commit = repo.get_object(commit_sha)
index = repo.open_index()
assert not list(index.changes_from_tree(repo.object_store, commit.tree)), "uncommited changes"
store = repo.object_store
def render_rst(blob, path):
doc = publish_doctree(blob.as_raw_string())
for node in doc.traverse(nodes.reference):
uri = urlparse.urlparse(node['refuri'])
if not uri.netloc and os.path.basename(uri.path) == "README.rst":
node['refuri'] = urlparse.urlunparse(
(uri.scheme, uri.netloc, uri.path[:-10] or "./", uri.params, uri.query, uri.fragment))
output = publish_from_doctree(
doc,
destination_path=path,
def get_version(cls):
from dulwich.repo import Repo
repo = Repo(cls.src_path)
return cls.version + "git" + repo.head()
class Repo(object):
"""
Wrapper around a libgit Repository that knows:
* How to get all the files in the repository
* How to get the oid of HEAD
* How to get the commit times of the files we want commit times for
It's written with speed in mind, given the constraints of making
performant code in python!
"""
def __init__(self, root_folder):
self.git = Repository(root_folder)
def all_files(self):
"""Return a set of all the files under git control"""
return set([entry.decode() for entry, _ in self.git.open_index().items()])
@property
def first_commit(self):
"""Return the oid of HEAD"""
return self.git.head().decode()
def file_commit_times(self, use_files_paths, debug=False):
"""
Traverse the commits in the repository, starting from HEAD until we have
found the commit times for all the files we care about.
Yield each file once, only when it is found to be changed in some commit.
If self.debug is true, also output log.debug for the speed we are going
through commits (output commits/second every 1000 commits and every
100000 commits)
"""
prefixes = PrefixTree()
prefixes.fill(use_files_paths)
for entry in self.git.get_walker():
# Commit time taking into account the timezone
commit_time = entry.commit.commit_time - entry.commit.commit_timezone
# Get us the two different tree structures between parents and current
cf_and_pf, changes = self.tree_structures_for(()
, entry.commit.tree
, [self.git.get_object(oid).tree for oid in entry.commit.parents]
, prefixes
)
# Deep dive into any differences
difference = []
if changes:
cfs_and_pfs = [(cf_and_pf, changes)]
while cfs_and_pfs:
nxt, changes = cfs_and_pfs.pop(0)
for thing, changes, is_path in self.differences_between(nxt[0], nxt[1], changes, prefixes):
if is_path:
found = prefixes.remove(thing[:-1], thing[-1])
if found:
difference.append('/'.join(thing))
else:
cfs_and_pfs.append((thing, changes))
# Only yield if there was a difference
if difference:
yield entry.commit.sha().hexdigest(), commit_time, difference
# If nothing remains, then break!
if not prefixes:
break
def entries_in_tree_oid(self, prefix, tree_oid):
"""Find the tree at this oid and return entries prefixed with ``prefix``"""
try:
tree = self.git.get_object(tree_oid)
except KeyError:
log.warning("Couldn't find object {0}".format(tree_oid))
return empty
else:
return frozenset(self.entries_in_tree(prefix, tree))
def entries_in_tree(self, prefix, tree):
"""
Traverse the entries in this tree and yield (prefix, is_tree, oid)
Where prefix is a tuple of the given prefix and the name of the entry.
"""
for entry in tree.items():
if prefix:
new_prefix = prefix + (entry.path.decode(), )
else:
new_prefix = (entry.path.decode(), )
yield (new_prefix, stat.S_ISDIR(entry.mode), entry.sha)
def tree_structures_for(self, prefix, current_oid, parent_oids, prefixes):
"""
Return the entries for this commit, the entries of the parent commits,
and the difference between the two (current_files - parent_files)
"""
if prefix and prefixes and prefix not in prefixes:
return empty, empty
parent_files = set()
for oid in parent_oids:
parent_files.update(self.entries_in_tree_oid(prefix, oid))
current_files = self.entries_in_tree_oid(prefix, current_oid)
return (current_files, parent_files), (current_files - parent_files)
def differences_between(self, current_files, parent_files, changes, prefixes):
"""
yield (thing, changes, is_path)
If is_path is true, changes is None and thing is the path as a tuple.
If is_path is false, thing is the current_files and parent_files for
that changed treeentry and changes is the difference between current_files
and parent_files.
The code here is written to squeeze as much performance as possible out
of this operation.
"""
parent_oid = None
if any(is_tree for _, is_tree, _ in changes):
if len(changes) == 1:
wanted_path = list(changes)[0][0]
parent_oid = frozenset([oid for path, is_tree, oid in parent_files if path == wanted_path and is_tree])
else:
parent_values = defaultdict(set)
parent_changes = parent_files - current_files
for path, is_tree, oid in parent_changes:
if is_tree:
parent_values[path].add(oid)
for path, is_tree, oid in changes:
if is_tree and path not in prefixes:
continue
if not is_tree:
yield path, None, True
else:
parent_oids = parent_oid if parent_oid is not None else parent_values.get(path, empty)
cf_and_pf, changes = self.tree_structures_for(path, oid, parent_oids, prefixes)
if changes:
yield cf_and_pf, changes, False
def git_commit_info():
git = Repo('.')
commit = git.get_object(git.head())
return {'id': commit.id.decode("utf-8")[0:7], 'id_full': commit.id.decode("utf-8"),
'author': regex.findall("(.*?) <(.*?)>", commit.author.decode("utf-8"))[0],
'message': commit.message.decode("utf-8").strip('\r\n').split('\n')[0]}
def clone(
cls,
url: str,
name: str | None = None,
branch: str | None = None,
tag: str | None = None,
revision: str | None = None,
source_root: Path | None = None,
clean: bool = False,
) -> Repo:
source_root = source_root or cls.get_default_source_root()
source_root.mkdir(parents=True, exist_ok=True)
name = name or cls.get_name_from_source_url(url=url)
target = source_root / name
refspec = GitRefSpec(branch=branch, revision=revision, tag=tag)
if target.exists():
if clean:
# force clean the local copy if it exists, do not reuse
remove_directory(target, force=True)
else:
# check if the current local copy matches the requested ref spec
try:
current_repo = Repo(str(target)) # type: ignore[no-untyped-call]
with current_repo:
current_sha = current_repo.head().decode("utf-8")
except (NotGitRepository, AssertionError, KeyError):
# something is wrong with the current checkout, clean it
remove_directory(target, force=True)
else:
if not is_revision_sha(revision=current_sha):
# head is not a sha, this will cause issues later, lets reset
remove_directory(target, force=True)
elif (
refspec.is_sha
and refspec.revision is not None
and current_sha.startswith(refspec.revision)
):
# if revision is used short-circuit remote fetch head matches
return current_repo
try:
if not cls.is_using_legacy_client():
local = cls._clone(url=url, refspec=refspec, target=target)
cls._clone_submodules(repo=local)
return local
except HTTPUnauthorized:
# we do this here to handle http authenticated repositories as dulwich
# does not currently support using credentials from git-credential helpers.
# upstream issue: https://github.com/jelmer/dulwich/issues/873
#
# this is a little inefficient, however preferred as this is transparent
# without additional configuration or changes for existing projects that
# use http basic auth credentials.
logger.debug(
"Unable to fetch from private repository '%s', falling back to"
" system git",
url,
)
# fallback to legacy git client
return cls._clone_legacy(url=url, refspec=refspec, target=target)
if n == 0: break
return ''.join(reversed(s))
CACHE_PREFIX = num_encode(os.getpid()) + '.'
CACHE_PREFIX += '.'.join(str(num_encode(int(p))) for p in str(time.time()).split('.'))
try:
BACKLOG_ARCHIVE = settings.BACKLOG_ARCHIVE
if not os.path.exists(os.path.join(BACKLOG_ARCHIVE, '.git')):
BACKLOG_ARCHIVE = None
except AttributeError:
BACKLOG_ARCHIVE = None
except ImportError:
BACKLOG_ARCHIVE = None
RELEASE = None
__projectdir__ = os.path.abspath(os.path.dirname(__file__))
while __projectdir__ != '/' and not os.path.exists(os.path.join(__projectdir__, 'settings.py')):
__projectdir__ = os.path.dirname(__projectdir__)
if __projectdir__ != '/' and os.path.exists(os.path.join(__projectdir__, '.git')):
repo = Repo(__projectdir__)
for commit in repo.revision_history(repo.head()):
RELEASE = 'Git: ' + commit.tree
break
if RELEASE is None:
try:
import agilitorelease
RELEASE = agilitorelease.RELEASE
except:
pass
class GitStorage():
def _ignoreFile(self, dirName, fileName):
"""
used for the copTree stuff
``dirName``
the working directory
``fileName``
list of files inside the directory (dirName)
"""
result = []
for i in fileName:
path = dirName + i
if path not in fileToIgnore:
result.append(path)
return result
def _commit(self, tree):
"""
commit a tree used only by the init
``tree``
tree to commit
"""
commit = Commit()
commit.tree = tree.id
commit.encoding = "UTF-8"
commit.committer = commit.author = 'debexpo <%s>' % (pylons.config['debexpo.email'])
commit.commit_time = commit.author_time = int(time())
tz = parse_timezone('-0200')[0]
commit.commit_timezone = commit.author_timezone = tz
commit.message = " "
self.repo.object_store.add_object(tree)
self.repo.object_store.add_object(commit)
self.repo.refs["HEAD"] = commit.id
log.debug('commiting')
return commit.id
def __init__(self, path):
#creating the repository
if os.path.isdir(path):
log.debug("directory exist, taking it as a git repository")
self.repo = Repo(path)
else:
log.debug("directory doesn't exist, creating")
os.makedirs(path)
log.debug("initiate the repo")
self.repo = Repo.init(path)
log.debug("adding an empty tree to the repository")
self._commit(Tree())
#only this function will be used on upload
def change(self, files):
"""
used to change afile in the git storage can be called for the first upload we don't care
``files``
a list of file to change
"""
if len(files) == 0:
log.debug("trying to change nothing will do... nothing")
else:
log.debug("this will change %i files" % (len(files)))
for f in files:
self.repo.stage(str(f))
log.debug("stages dones")
self.repo.do_commit("this is so awesome that nobody will never see it",
committer="same here <*****@*****.**>")
def buildTreeDiff(self, dest, tree=None, originalTree=None):
"""
creating files from the diff between 2 trees, it will be used in the code browser to get older version
(walking on history)
``tree``
the tree that you want to compare to
``dest``
the destination folder to build sources
``originalTree``
the original Tree, by default it's the last one
by default it's retun the last changed files
"""
if tree is None:
head = self.repo.commit(self.repo.commit(self.repo.head()).parents[0])
tree = self.repo.tree(head.tree)
if originalTree is None:
originalTree = self.repo.tree(self.repo.commit(self.repo.head()).tree)
blobToBuild = []
#getting blob that have changed
for blob in self.repo.object_store.iter_tree_contents(tree.id):
if blob not in originalTree:
blobToBuild.append(blob)
fileToIgnore.append(blob.path)
repoLocation = os.path.join(str(self.repo).split("'")[1])
#creating the folder with link to older files
if os.path.exists(repoLocation + dest):
log.warning("%s already exist, copy will not work")
else:
log.debug("copying files")
shutil.copytree(repoLocation, repoLocation + dest, symlinks=True, ignore=self._ignoreFile)
for b in blobToBuild:
fileDirectory = os.path.split(b.path)
fileDirectory.pop()
if not os.path.exists(os.path.join(repoLocation + dest, os.path.join(fileDirectory))):
os.makedirs(os.path.join(repoLocation + dest, os.path.join(fileDirectory)))
file = open(os.path.join(repoLocation + dest, b.path), 'w')
file.write(self.repo.get_object(b.sha).as_raw_string())
file.close()
tree = None
originalTree = None
#get*
def getLastTree(self):
"""
return the last tree
"""
return self.repo.tree(self.repo._commit(self.repo.head()).tree)
def getAllTrees(self):
"""
return trees
"""
result = []
commit = self.repo._commit(self.repo.head())
for c in commit._get_parents():
result.append(c.tree)
return result
def getOlderFileContent(self, file):
"""
return the first file's content that changed from the file
``file``
the file to work on
"""
with open(file) as f:
originalBlob = Blob.from_string("".join(f.readlines()))
trees = self.getAllTrees()
for t in trees:
#parsing tree in order to find the tree where the file change
if originalBlob not in t:
tree = t
break
#tree must be existent, other way file is not correct
if tree is None:
log.error(
"there is no tree that contain this blob this souldn't happen, other way this file does not appear to come from this package")
else:
if self.repo._commit(self.repo.head()).tree == tree:
olderTree = self.repo.commit(self.repo.head())._get_parents()[0].tree
else:
for c in self.repo._commit(self.repo.head())._get_parents():
if c.tree == tree:
try:
olderTree = c.get_parents()[0]
except IndexError:
log.debug("file is the last version")
olderTree = tree
if olderTree != tree:
#we must check here the blob that contains the older file
for b in self.repo.object_store.iter_tree_contents(olderTree.id):
if originalBlob.path == b.path:
#older blob find! awesome, in the first loop we already test if they are the same
# that's why we can now return the content of the file
return self.repo.get_object(b.sha).as_raw_string()
return ""
def getOlderCommits(self):
"""
return a list of all commits
"""
return self.repo.commit(self.repo.head())._get_parents()
# -*- coding: utf-8 -*-
from distutils.core import setup
try:
from dulwich.repo import Repo
repo = Repo('.')
version = repo.head()
except:
version = 'unknown'
datafiles = []
setup(name='vcsstats',
version=version,
author='Henrik Stuart, Alexander Færøy',
author_email='*****@*****.**',
url='http://github.com/ahf/vcsstats/tree/master',
description='Version control system statistics',
license='MIT',
scripts=[],
data_files=datafiles,
packages=['vcsstatslib', 'vcsstatslib.charts']
)
def edit(self, content=None, username=None, *virtual_path, **keywords):
'''id: the git id of the blob before it was edited
branch: master (default)'''
#setup the defaults
branch = "master"
url = ManagedPath(virtual_path)
if "branch" in keywords: branch = keywords["branch"]
sha = self.sha
print "sha is: ", sha
print "keywords: ", keywords
commit = sha
print "content is: ", content
print "self.filename is: ", self.filename
if content is None:
repo = Repo(self.package.path())
set_head(repo, branch)
if not sha:
print "repo.head() = ", repo.head()
sha = dulwich.object_store.tree_lookup_path(repo.get_object, repo.get_object(repo.head()).tree, self.filename)[1]
obj = repo.get_object(sha)
contents = obj.as_pretty_string()
return_contents = "<form action=\"" + cherrypy.url() + "\" method=\"POST\">"
return_contents = return_contents + "<textarea name=content rows='20' cols='120'>" + contents + "</textarea><br />"
#if the user isn't logged in ...
if not hasattr(cherrypy.session, "login"): return_contents = return_contents + "username: <input type=text name=username value=\"anonymous\"><br />"
if sha: return_contents = return_contents + "<input type=hidden name=id value=\"" + sha + "\">"
return_contents = return_contents + "<input type=hidden name=\"branch\" value=\"" + branch + "\">"
return_contents = return_contents + "<input type=submit name=submit value=edit></form>"
self.content = return_contents
self.branch = branch
return self.respond()
elif (sha or branch): #it's been edited
if username==None and hasattr(cherrypy.session, "login"):
if cherrypy.session.login==None: raise ValueError, "FileViewer.edit: no username supplied"
elif username==None or username=="anonymous":
anon = True #whether or not the user is anonymous
if SESSION_KEY in cherrypy.session.keys():
username = cherrypy.session[SESSION_KEY].username
anon = False
else: username = "******"
#at least until we get access control lists working
if anon:
if branch=="master": #don't let anonymous users modify branch "master"
branch = "anonymous"
branch = "anonymous"
#make the blob
blob = Blob.from_string(content)
repo = Repo(self.package.path())
#change to the right branch
last_head = repo.head()
set_head(repo, "master")
last_commit = repo.get_object(repo.head())
tree = repo.tree(repo.get_object(repo.head()).tree)
#set the file
tree[self.filename] = (0100644, blob.id)
#make the commit
commit = Commit()
commit.tree = tree.id
commit.parents = [last_head]
commit.author = commit.committer = username
commit.commit_time = commit.author_time = int(time.time())
commit.commit_timezone = commit.author_timezone = parse_timezone("-0600")
commit.encoding = "UTF-8"
commit.message = "not implemented yet"
repo.object_store.add_object(blob)
repo.object_store.add_object(tree)
repo.object_store.add_object(commit)
repo.refs["refs/heads/" + branch] = commit.id
repo.refs["HEAD"] = "ref: refs/heads/" + branch
new_link = "<a href=\"/package/" + self.package.name + ":" + branch + "/" + self.filename + "/" + blob.id + "\">go to the new version</a>"
self.new_link = new_link
self.content = add_newlines("edited (name=%s, branch=%s, sha=%s) new link: %s\n\n\n" % (username, branch, sha, new_link))
self.content = self.content + "<pre>" + content + "</pre>"
self.branch = branch
return self.respond()
class GitRepo(object):
def __init__(self, path):
if os.path.exists(path):
if not os.path.isdir(path):
raise IOError('Git repository "%s" must be a directory.' %
path)
try:
self.repo = Repo(path)
except NotGitRepository:
# repo does not exist
self.repo = Repo.init(path, not os.path.exists(path))
self.temp_persist_files = []
def _get_commit(self, version="HEAD"):
commit = self.repo[version]
if not isinstance(commit, Commit):
raise NotCommitError(commit)
return commit
def get_type(self, name, version="HEAD"):
commit = self._get_commit(version)
tree = self.repo.tree(commit.tree)
if name not in tree:
raise KeyError('Cannot find object "%s"' % name)
if tree[name][0] & stat.S_IFDIR:
return "tree"
else:
return "blob"
def get_path(self, name, version="HEAD", path_type=None, out_name=None,
out_suffix=''):
if path_type is None:
path_type = self.get_type(name, version)
if path_type == 'tree':
return self.get_dir(name, version, out_name, out_suffix)
elif path_type == 'blob':
return self.get_file(name, version, out_name, out_suffix)
raise TypeError("Unknown path type '%s'" % path_type)
def _write_blob(self, blob_sha, out_fname=None, out_suffix=''):
if out_fname is None:
# create a temporary file
(fd, out_fname) = tempfile.mkstemp(suffix=out_suffix,
prefix='vt_persist')
os.close(fd)
self.temp_persist_files.append(out_fname)
else:
out_dirname = os.path.dirname(out_fname)
if out_dirname and not os.path.exists(out_dirname):
os.makedirs(out_dirname)
blob = self.repo.get_blob(blob_sha)
with open(out_fname, "wb") as f:
for b in blob.as_raw_chunks():
f.write(b)
return out_fname
def get_file(self, name, version="HEAD", out_fname=None,
out_suffix=''):
commit = self._get_commit(version)
tree = self.repo.tree(commit.tree)
if name not in tree:
raise KeyError('Cannot find blob "%s"' % name)
blob_sha = tree[name][1]
out_fname = self._write_blob(blob_sha, out_fname, out_suffix)
return out_fname
def get_dir(self, name, version="HEAD", out_dirname=None,
out_suffix=''):
if out_dirname is None:
# create a temporary directory
out_dirname = tempfile.mkdtemp(suffix=out_suffix,
prefix='vt_persist')
self.temp_persist_files.append(out_dirname)
elif not os.path.exists(out_dirname):
os.makedirs(out_dirname)
commit = self._get_commit(version)
tree = self.repo.tree(commit.tree)
if name not in tree:
raise KeyError('Cannot find tree "%s"' % name)
subtree_id = tree[name][1]
# subtree = self.repo.tree(subtree_id)
for entry in self.repo.object_store.iter_tree_contents(subtree_id):
out_fname = os.path.join(out_dirname, entry.path)
self._write_blob(entry.sha, out_fname)
return out_dirname
def get_hash(self, name, version="HEAD", path_type=None):
commit = self._get_commit(version)
tree = self.repo.tree(commit.tree)
if name not in tree:
raise KeyError('Cannot find object "%s"' % name)
return tree[name][1]
@staticmethod
def compute_blob_hash(fname, chunk_size=1<<16):
obj_len = os.path.getsize(fname)
head = object_header(Blob.type_num, obj_len)
with open(fname, "rb") as f:
def read_chunk():
return f.read(chunk_size)
my_iter = chain([head], iter(read_chunk,''))
return iter_sha1(my_iter)
return None
@staticmethod
def compute_tree_hash(dirname):
tree = Tree()
for entry in sorted(os.listdir(dirname)):
fname = os.path.join(dirname, entry)
if os.path.isdir(fname):
thash = GitRepo.compute_tree_hash(fname)
mode = stat.S_IFDIR # os.stat(fname)[stat.ST_MODE]
tree.add(entry, mode, thash)
elif os.path.isfile(fname):
bhash = GitRepo.compute_blob_hash(fname)
mode = os.stat(fname)[stat.ST_MODE]
tree.add(entry, mode, bhash)
return tree.id
@staticmethod
def compute_hash(path):
if os.path.isdir(path):
return GitRepo.compute_tree_hash(path)
elif os.path.isfile(path):
return GitRepo.compute_blob_hash(path)
raise TypeError("Do not support this type of path")
def get_latest_version(self, path):
head = self.repo.head()
walker = Walker(self.repo.object_store, [head], max_entries=1,
paths=[path])
return iter(walker).next().commit.id
def _stage(self, filename):
fullpath = os.path.join(self.repo.path, filename)
if os.path.islink(fullpath):
debug.warning("Warning: not staging symbolic link %s" % os.path.basename(filename))
elif os.path.isdir(fullpath):
for f in os.listdir(fullpath):
self._stage(os.path.join(filename, f))
else:
if os.path.sep != '/':
filename = filename.replace(os.path.sep, '/')
self.repo.stage(filename)
def add_commit(self, filename):
self.setup_git()
self._stage(filename)
commit_id = self.repo.do_commit('Updated %s' % filename)
return commit_id
def setup_git(self):
config_stack = self.repo.get_config_stack()
try:
config_stack.get(('user',), 'name')
config_stack.get(('user',), 'email')
except KeyError:
from vistrails.core.system import current_user
from dulwich.config import ConfigFile
user = current_user()
repo_conf = self.repo.get_config()
repo_conf.set(('user',), 'name', user)
repo_conf.set(('user',), 'email', '%s@localhost' % user)
repo_conf.write_to_path()
import multiprocessing
bind = "0.0.0.0:5000"
workers = multiprocessing.cpu_count() * 2 + 1
worker_class = "eventlet"
reload = True
try:
from dulwich.repo import Repo
r = Repo('.')
# gittle
def active_branch(repo, SYMREF=b'ref: ', REFS_BRANCHES=b'refs/heads/'):
"""Returns the name of the active branch, or None, if HEAD is detached
"""
x = repo.refs.read_ref('HEAD')
if not x.startswith(SYMREF):
return b''
else:
symref = x[len(SYMREF):]
if not symref.startswith(REFS_BRANCHES):
return b''
else:
return symref[len(REFS_BRANCHES):]
import gunicorn
gunicorn.SERVER_SOFTWARE = 'MarryBird %s(%s)' % (active_branch(r).decode('ascii'), r.head().decode('ascii')[:11])
except:
pass