bazarr/libs/git/objects/submodule/base.py

1204 lines
53 KiB
Python
Raw Normal View History

# need a dict to set bloody .name field
from io import BytesIO
import logging
import os
import stat
from unittest import SkipTest
import uuid
import git
from git.cmd import Git
from git.compat import (
string_types,
defenc,
is_win,
)
from git.config import (
SectionConstraint,
GitConfigParser,
cp
)
from git.exc import (
InvalidGitRepositoryError,
NoSuchPathError,
RepositoryDirtyError
)
from git.objects.base import IndexObject, Object
from git.objects.util import Traversable
from git.util import (
Iterable,
join_path_native,
to_native_path_linux,
RemoteProgress,
rmtree,
unbare_repo
)
from git.util import HIDE_WINDOWS_KNOWN_ERRORS
import os.path as osp
from .util import (
mkhead,
sm_name,
sm_section,
SubmoduleConfigParser,
find_first_remote_branch
)
__all__ = ["Submodule", "UpdateProgress"]
log = logging.getLogger('git.objects.submodule.base')
log.addHandler(logging.NullHandler())
class UpdateProgress(RemoteProgress):
"""Class providing detailed progress information to the caller who should
derive from it and implement the ``update(...)`` message"""
CLONE, FETCH, UPDWKTREE = [1 << x for x in range(RemoteProgress._num_op_codes, RemoteProgress._num_op_codes + 3)]
_num_op_codes = RemoteProgress._num_op_codes + 3
__slots__ = ()
BEGIN = UpdateProgress.BEGIN
END = UpdateProgress.END
CLONE = UpdateProgress.CLONE
FETCH = UpdateProgress.FETCH
UPDWKTREE = UpdateProgress.UPDWKTREE
# IndexObject comes via util module, its a 'hacky' fix thanks to pythons import
# mechanism which cause plenty of trouble of the only reason for packages and
# modules is refactoring - subpackages shouldn't depend on parent packages
class Submodule(IndexObject, Iterable, Traversable):
"""Implements access to a git submodule. They are special in that their sha
represents a commit in the submodule's repository which is to be checked out
at the path of this instance.
The submodule type does not have a string type associated with it, as it exists
solely as a marker in the tree and index.
All methods work in bare and non-bare repositories."""
_id_attribute_ = "name"
k_modules_file = '.gitmodules'
k_head_option = 'branch'
k_head_default = 'master'
k_default_mode = stat.S_IFDIR | stat.S_IFLNK # submodules are directories with link-status
# this is a bogus type for base class compatibility
type = 'submodule'
__slots__ = ('_parent_commit', '_url', '_branch_path', '_name', '__weakref__')
_cache_attrs = ('path', '_url', '_branch_path')
def __init__(self, repo, binsha, mode=None, path=None, name=None, parent_commit=None, url=None, branch_path=None):
"""Initialize this instance with its attributes. We only document the ones
that differ from ``IndexObject``
:param repo: Our parent repository
:param binsha: binary sha referring to a commit in the remote repository, see url parameter
:param parent_commit: see set_parent_commit()
:param url: The url to the remote repository which is the submodule
:param branch_path: full (relative) path to ref to checkout when cloning the remote repository"""
super(Submodule, self).__init__(repo, binsha, mode, path)
self.size = 0
self._parent_commit = parent_commit
if url is not None:
self._url = url
if branch_path is not None:
assert isinstance(branch_path, string_types)
self._branch_path = branch_path
if name is not None:
self._name = name
def _set_cache_(self, attr):
if attr in ('path', '_url', '_branch_path'):
reader = self.config_reader()
# default submodule values
try:
self.path = reader.get('path')
except cp.NoSectionError:
raise ValueError("This submodule instance does not exist anymore in '%s' file"
% osp.join(self.repo.working_tree_dir, '.gitmodules'))
# end
self._url = reader.get('url')
# git-python extension values - optional
self._branch_path = reader.get_value(self.k_head_option, git.Head.to_full_path(self.k_head_default))
elif attr == '_name':
raise AttributeError("Cannot retrieve the name of a submodule if it was not set initially")
else:
super(Submodule, self)._set_cache_(attr)
# END handle attribute name
def _get_intermediate_items(self, item):
""":return: all the submodules of our module repository"""
try:
return type(self).list_items(item.module())
except InvalidGitRepositoryError:
return []
# END handle intermediate items
@classmethod
def _need_gitfile_submodules(cls, git):
return git.version_info[:3] >= (1, 7, 5)
def __eq__(self, other):
"""Compare with another submodule"""
# we may only compare by name as this should be the ID they are hashed with
# Otherwise this type wouldn't be hashable
# return self.path == other.path and self.url == other.url and super(Submodule, self).__eq__(other)
return self._name == other._name
def __ne__(self, other):
"""Compare with another submodule for inequality"""
return not (self == other)
def __hash__(self):
"""Hash this instance using its logical id, not the sha"""
return hash(self._name)
def __str__(self):
return self._name
def __repr__(self):
return "git.%s(name=%s, path=%s, url=%s, branch_path=%s)"\
% (type(self).__name__, self._name, self.path, self.url, self.branch_path)
@classmethod
def _config_parser(cls, repo, parent_commit, read_only):
""":return: Config Parser constrained to our submodule in read or write mode
:raise IOError: If the .gitmodules file cannot be found, either locally or in the repository
at the given parent commit. Otherwise the exception would be delayed until the first
access of the config parser"""
parent_matches_head = True
if parent_commit is not None:
try:
parent_matches_head = repo.head.commit == parent_commit
except ValueError:
# We are most likely in an empty repository, so the HEAD doesn't point to a valid ref
pass
# end handle parent_commit
if not repo.bare and parent_matches_head:
fp_module = osp.join(repo.working_tree_dir, cls.k_modules_file)
else:
assert parent_commit is not None, "need valid parent_commit in bare repositories"
try:
fp_module = cls._sio_modules(parent_commit)
except KeyError:
raise IOError("Could not find %s file in the tree of parent commit %s" %
(cls.k_modules_file, parent_commit))
# END handle exceptions
# END handle non-bare working tree
if not read_only and (repo.bare or not parent_matches_head):
raise ValueError("Cannot write blobs of 'historical' submodule configurations")
# END handle writes of historical submodules
return SubmoduleConfigParser(fp_module, read_only=read_only)
def _clear_cache(self):
# clear the possibly changed values
for name in self._cache_attrs:
try:
delattr(self, name)
except AttributeError:
pass
# END try attr deletion
# END for each name to delete
@classmethod
def _sio_modules(cls, parent_commit):
""":return: Configuration file as BytesIO - we only access it through the respective blob's data"""
sio = BytesIO(parent_commit.tree[cls.k_modules_file].data_stream.read())
sio.name = cls.k_modules_file
return sio
def _config_parser_constrained(self, read_only):
""":return: Config Parser constrained to our submodule in read or write mode"""
try:
pc = self.parent_commit
except ValueError:
pc = None
# end handle empty parent repository
parser = self._config_parser(self.repo, pc, read_only)
parser.set_submodule(self)
return SectionConstraint(parser, sm_section(self.name))
@classmethod
def _module_abspath(cls, parent_repo, path, name):
if cls._need_gitfile_submodules(parent_repo.git):
return osp.join(parent_repo.git_dir, 'modules', name)
else:
return osp.join(parent_repo.working_tree_dir, path)
# end
@classmethod
def _clone_repo(cls, repo, url, path, name, **kwargs):
""":return: Repo instance of newly cloned repository
:param repo: our parent repository
:param url: url to clone from
:param path: repository-relative path to the submodule checkout location
:param name: canonical of the submodule
:param kwrags: additinoal arguments given to git.clone"""
module_abspath = cls._module_abspath(repo, path, name)
module_checkout_path = module_abspath
if cls._need_gitfile_submodules(repo.git):
kwargs['separate_git_dir'] = module_abspath
module_abspath_dir = osp.dirname(module_abspath)
if not osp.isdir(module_abspath_dir):
os.makedirs(module_abspath_dir)
module_checkout_path = osp.join(repo.working_tree_dir, path)
# end
clone = git.Repo.clone_from(url, module_checkout_path, **kwargs)
if cls._need_gitfile_submodules(repo.git):
cls._write_git_file_and_module_config(module_checkout_path, module_abspath)
# end
return clone
@classmethod
def _to_relative_path(cls, parent_repo, path):
""":return: a path guaranteed to be relative to the given parent-repository
:raise ValueError: if path is not contained in the parent repository's working tree"""
path = to_native_path_linux(path)
if path.endswith('/'):
path = path[:-1]
# END handle trailing slash
if osp.isabs(path):
working_tree_linux = to_native_path_linux(parent_repo.working_tree_dir)
if not path.startswith(working_tree_linux):
raise ValueError("Submodule checkout path '%s' needs to be within the parents repository at '%s'"
% (working_tree_linux, path))
path = path[len(working_tree_linux.rstrip('/')) + 1:]
if not path:
raise ValueError("Absolute submodule path '%s' didn't yield a valid relative path" % path)
# end verify converted relative path makes sense
# end convert to a relative path
return path
@classmethod
def _write_git_file_and_module_config(cls, working_tree_dir, module_abspath):
"""Writes a .git file containing a (preferably) relative path to the actual git module repository.
It is an error if the module_abspath cannot be made into a relative path, relative to the working_tree_dir
:note: will overwrite existing files !
:note: as we rewrite both the git file as well as the module configuration, we might fail on the configuration
and will not roll back changes done to the git file. This should be a non-issue, but may easily be fixed
if it becomes one
:param working_tree_dir: directory to write the .git file into
:param module_abspath: absolute path to the bare repository
"""
git_file = osp.join(working_tree_dir, '.git')
rela_path = osp.relpath(module_abspath, start=working_tree_dir)
if is_win:
if osp.isfile(git_file):
os.remove(git_file)
with open(git_file, 'wb') as fp:
fp.write(("gitdir: %s" % rela_path).encode(defenc))
with GitConfigParser(osp.join(module_abspath, 'config'),
read_only=False, merge_includes=False) as writer:
writer.set_value('core', 'worktree',
to_native_path_linux(osp.relpath(working_tree_dir, start=module_abspath)))
#{ Edit Interface
@classmethod
def add(cls, repo, name, path, url=None, branch=None, no_checkout=False):
"""Add a new submodule to the given repository. This will alter the index
as well as the .gitmodules file, but will not create a new commit.
If the submodule already exists, no matter if the configuration differs
from the one provided, the existing submodule will be returned.
:param repo: Repository instance which should receive the submodule
:param name: The name/identifier for the submodule
:param path: repository-relative or absolute path at which the submodule
should be located
It will be created as required during the repository initialization.
:param url: git-clone compatible URL, see git-clone reference for more information
If None, the repository is assumed to exist, and the url of the first
remote is taken instead. This is useful if you want to make an existing
repository a submodule of anotherone.
:param branch: name of branch at which the submodule should (later) be checked out.
The given branch must exist in the remote repository, and will be checked
out locally as a tracking branch.
It will only be written into the configuration if it not None, which is
when the checked out branch will be the one the remote HEAD pointed to.
The result you get in these situation is somewhat fuzzy, and it is recommended
to specify at least 'master' here.
Examples are 'master' or 'feature/new'
:param no_checkout: if True, and if the repository has to be cloned manually,
no checkout will be performed
:return: The newly created submodule instance
:note: works atomically, such that no change will be done if the repository
update fails for instance"""
if repo.bare:
raise InvalidGitRepositoryError("Cannot add submodules to bare repositories")
# END handle bare repos
path = cls._to_relative_path(repo, path)
# assure we never put backslashes into the url, as some operating systems
# like it ...
if url is not None:
url = to_native_path_linux(url)
# END assure url correctness
# INSTANTIATE INTERMEDIATE SM
sm = cls(repo, cls.NULL_BIN_SHA, cls.k_default_mode, path, name, url='invalid-temporary')
if sm.exists():
# reretrieve submodule from tree
try:
sm = repo.head.commit.tree[path]
sm._name = name
return sm
except KeyError:
# could only be in index
index = repo.index
entry = index.entries[index.entry_key(path, 0)]
sm.binsha = entry.binsha
return sm
# END handle exceptions
# END handle existing
# fake-repo - we only need the functionality on the branch instance
br = git.Head(repo, git.Head.to_full_path(str(branch) or cls.k_head_default))
has_module = sm.module_exists()
branch_is_default = branch is None
if has_module and url is not None:
if url not in [r.url for r in sm.module().remotes]:
raise ValueError(
"Specified URL '%s' does not match any remote url of the repository at '%s'" % (url, sm.abspath))
# END check url
# END verify urls match
mrepo = None
if url is None:
if not has_module:
raise ValueError("A URL was not given and existing repository did not exsit at %s" % path)
# END check url
mrepo = sm.module()
urls = [r.url for r in mrepo.remotes]
if not urls:
raise ValueError("Didn't find any remote url in repository at %s" % sm.abspath)
# END verify we have url
url = urls[0]
else:
# clone new repo
kwargs = {'n': no_checkout}
if not branch_is_default:
kwargs['b'] = br.name
# END setup checkout-branch
# _clone_repo(cls, repo, url, path, name, **kwargs):
mrepo = cls._clone_repo(repo, url, path, name, **kwargs)
# END verify url
## See #525 for ensuring git urls in config-files valid under Windows.
url = Git.polish_url(url)
# It's important to add the URL to the parent config, to let `git submodule` know.
# otherwise there is a '-' character in front of the submodule listing
# a38efa84daef914e4de58d1905a500d8d14aaf45 mymodule (v0.9.0-1-ga38efa8)
# -a38efa84daef914e4de58d1905a500d8d14aaf45 submodules/intermediate/one
with sm.repo.config_writer() as writer:
writer.set_value(sm_section(name), 'url', url)
# update configuration and index
index = sm.repo.index
with sm.config_writer(index=index, write=False) as writer:
writer.set_value('url', url)
writer.set_value('path', path)
sm._url = url
if not branch_is_default:
# store full path
writer.set_value(cls.k_head_option, br.path)
sm._branch_path = br.path
# we deliberately assume that our head matches our index !
sm.binsha = mrepo.head.commit.binsha
index.add([sm], write=True)
return sm
def update(self, recursive=False, init=True, to_latest_revision=False, progress=None, dry_run=False,
force=False, keep_going=False):
"""Update the repository of this submodule to point to the checkout
we point at with the binsha of this instance.
:param recursive: if True, we will operate recursively and update child-
modules as well.
:param init: if True, the module repository will be cloned into place if necessary
:param to_latest_revision: if True, the submodule's sha will be ignored during checkout.
Instead, the remote will be fetched, and the local tracking branch updated.
This only works if we have a local tracking branch, which is the case
if the remote repository had a master branch, or of the 'branch' option
was specified for this submodule and the branch existed remotely
:param progress: UpdateProgress instance or None if no progress should be shown
:param dry_run: if True, the operation will only be simulated, but not performed.
All performed operations are read-only
:param force:
If True, we may reset heads even if the repository in question is dirty. Additinoally we will be allowed
to set a tracking branch which is ahead of its remote branch back into the past or the location of the
remote branch. This will essentially 'forget' commits.
If False, local tracking branches that are in the future of their respective remote branches will simply
not be moved.
:param keep_going: if True, we will ignore but log all errors, and keep going recursively.
Unless dry_run is set as well, keep_going could cause subsequent/inherited errors you wouldn't see
otherwise.
In conjunction with dry_run, it can be useful to anticipate all errors when updating submodules
:note: does nothing in bare repositories
:note: method is definitely not atomic if recurisve is True
:return: self"""
if self.repo.bare:
return self
# END pass in bare mode
if progress is None:
progress = UpdateProgress()
# END handle progress
prefix = ''
if dry_run:
prefix = "DRY-RUN: "
# END handle prefix
# to keep things plausible in dry-run mode
if dry_run:
mrepo = None
# END init mrepo
try:
# ASSURE REPO IS PRESENT AND UPTODATE
#####################################
try:
mrepo = self.module()
rmts = mrepo.remotes
len_rmts = len(rmts)
for i, remote in enumerate(rmts):
op = FETCH
if i == 0:
op |= BEGIN
# END handle start
progress.update(op, i, len_rmts, prefix + "Fetching remote %s of submodule %r"
% (remote, self.name))
#===============================
if not dry_run:
remote.fetch(progress=progress)
# END handle dry-run
#===============================
if i == len_rmts - 1:
op |= END
# END handle end
progress.update(op, i, len_rmts, prefix + "Done fetching remote of submodule %r" % self.name)
# END fetch new data
except InvalidGitRepositoryError:
if not init:
return self
# END early abort if init is not allowed
# there is no git-repository yet - but delete empty paths
checkout_module_abspath = self.abspath
if not dry_run and osp.isdir(checkout_module_abspath):
try:
os.rmdir(checkout_module_abspath)
except OSError:
raise OSError("Module directory at %r does already exist and is non-empty"
% checkout_module_abspath)
# END handle OSError
# END handle directory removal
# don't check it out at first - nonetheless it will create a local
# branch according to the remote-HEAD if possible
progress.update(BEGIN | CLONE, 0, 1, prefix + "Cloning url '%s' to '%s' in submodule %r" %
(self.url, checkout_module_abspath, self.name))
if not dry_run:
mrepo = self._clone_repo(self.repo, self.url, self.path, self.name, n=True)
# END handle dry-run
progress.update(END | CLONE, 0, 1, prefix + "Done cloning to %s" % checkout_module_abspath)
if not dry_run:
# see whether we have a valid branch to checkout
try:
# find a remote which has our branch - we try to be flexible
remote_branch = find_first_remote_branch(mrepo.remotes, self.branch_name)
local_branch = mkhead(mrepo, self.branch_path)
# have a valid branch, but no checkout - make sure we can figure
# that out by marking the commit with a null_sha
local_branch.set_object(Object(mrepo, self.NULL_BIN_SHA))
# END initial checkout + branch creation
# make sure HEAD is not detached
mrepo.head.set_reference(local_branch, logmsg="submodule: attaching head to %s" % local_branch)
mrepo.head.ref.set_tracking_branch(remote_branch)
except (IndexError, InvalidGitRepositoryError):
log.warn("Failed to checkout tracking branch %s", self.branch_path)
# END handle tracking branch
# NOTE: Have to write the repo config file as well, otherwise
# the default implementation will be offended and not update the repository
# Maybe this is a good way to assure it doesn't get into our way, but
# we want to stay backwards compatible too ... . Its so redundant !
with self.repo.config_writer() as writer:
writer.set_value(sm_section(self.name), 'url', self.url)
# END handle dry_run
# END handle initialization
# DETERMINE SHAS TO CHECKOUT
############################
binsha = self.binsha
hexsha = self.hexsha
if mrepo is not None:
# mrepo is only set if we are not in dry-run mode or if the module existed
is_detached = mrepo.head.is_detached
# END handle dry_run
if mrepo is not None and to_latest_revision:
msg_base = "Cannot update to latest revision in repository at %r as " % mrepo.working_dir
if not is_detached:
rref = mrepo.head.ref.tracking_branch()
if rref is not None:
rcommit = rref.commit
binsha = rcommit.binsha
hexsha = rcommit.hexsha
else:
log.error("%s a tracking branch was not set for local branch '%s'", msg_base, mrepo.head.ref)
# END handle remote ref
else:
log.error("%s there was no local tracking branch", msg_base)
# END handle detached head
# END handle to_latest_revision option
# update the working tree
# handles dry_run
if mrepo is not None and mrepo.head.commit.binsha != binsha:
# We must assure that our destination sha (the one to point to) is in the future of our current head.
# Otherwise, we will reset changes that might have been done on the submodule, but were not yet pushed
# We also handle the case that history has been rewritten, leaving no merge-base. In that case
# we behave conservatively, protecting possible changes the user had done
may_reset = True
if mrepo.head.commit.binsha != self.NULL_BIN_SHA:
base_commit = mrepo.merge_base(mrepo.head.commit, hexsha)
if len(base_commit) == 0 or base_commit[0].hexsha == hexsha:
if force:
msg = "Will force checkout or reset on local branch that is possibly in the future of"
msg += "the commit it will be checked out to, effectively 'forgetting' new commits"
log.debug(msg)
else:
msg = "Skipping %s on branch '%s' of submodule repo '%s' as it contains un-pushed commits"
msg %= (is_detached and "checkout" or "reset", mrepo.head, mrepo)
log.info(msg)
may_reset = False
# end handle force
# end handle if we are in the future
if may_reset and not force and mrepo.is_dirty(index=True, working_tree=True, untracked_files=True):
raise RepositoryDirtyError(mrepo, "Cannot reset a dirty repository")
# end handle force and dirty state
# end handle empty repo
# end verify future/past
progress.update(BEGIN | UPDWKTREE, 0, 1, prefix +
"Updating working tree at %s for submodule %r to revision %s"
% (self.path, self.name, hexsha))
if not dry_run and may_reset:
if is_detached:
# NOTE: for now we force, the user is no supposed to change detached
# submodules anyway. Maybe at some point this becomes an option, to
# properly handle user modifications - see below for future options
# regarding rebase and merge.
mrepo.git.checkout(hexsha, force=force)
else:
mrepo.head.reset(hexsha, index=True, working_tree=True)
# END handle checkout
# if we may reset/checkout
progress.update(END | UPDWKTREE, 0, 1, prefix + "Done updating working tree for submodule %r"
% self.name)
# END update to new commit only if needed
except Exception as err:
if not keep_going:
raise
log.error(str(err))
# end handle keep_going
# HANDLE RECURSION
##################
if recursive:
# in dry_run mode, the module might not exist
if mrepo is not None:
for submodule in self.iter_items(self.module()):
submodule.update(recursive, init, to_latest_revision, progress=progress, dry_run=dry_run,
force=force, keep_going=keep_going)
# END handle recursive update
# END handle dry run
# END for each submodule
return self
@unbare_repo
def move(self, module_path, configuration=True, module=True):
"""Move the submodule to a another module path. This involves physically moving
the repository at our current path, changing the configuration, as well as
adjusting our index entry accordingly.
:param module_path: the path to which to move our module in the parent repostory's working tree,
given as repository-relative or absolute path. Intermediate directories will be created
accordingly. If the path already exists, it must be empty.
Trailing (back)slashes are removed automatically
:param configuration: if True, the configuration will be adjusted to let
the submodule point to the given path.
:param module: if True, the repository managed by this submodule
will be moved as well. If False, we don't move the submodule's checkout, which may leave
the parent repository in an inconsistent state.
:return: self
:raise ValueError: if the module path existed and was not empty, or was a file
:note: Currently the method is not atomic, and it could leave the repository
in an inconsistent state if a sub-step fails for some reason
"""
if module + configuration < 1:
raise ValueError("You must specify to move at least the module or the configuration of the submodule")
# END handle input
module_checkout_path = self._to_relative_path(self.repo, module_path)
# VERIFY DESTINATION
if module_checkout_path == self.path:
return self
# END handle no change
module_checkout_abspath = join_path_native(self.repo.working_tree_dir, module_checkout_path)
if osp.isfile(module_checkout_abspath):
raise ValueError("Cannot move repository onto a file: %s" % module_checkout_abspath)
# END handle target files
index = self.repo.index
tekey = index.entry_key(module_checkout_path, 0)
# if the target item already exists, fail
if configuration and tekey in index.entries:
raise ValueError("Index entry for target path did already exist")
# END handle index key already there
# remove existing destination
if module:
if osp.exists(module_checkout_abspath):
if len(os.listdir(module_checkout_abspath)):
raise ValueError("Destination module directory was not empty")
# END handle non-emptiness
if osp.islink(module_checkout_abspath):
os.remove(module_checkout_abspath)
else:
os.rmdir(module_checkout_abspath)
# END handle link
else:
# recreate parent directories
# NOTE: renames() does that now
pass
# END handle existence
# END handle module
# move the module into place if possible
cur_path = self.abspath
renamed_module = False
if module and osp.exists(cur_path):
os.renames(cur_path, module_checkout_abspath)
renamed_module = True
if osp.isfile(osp.join(module_checkout_abspath, '.git')):
module_abspath = self._module_abspath(self.repo, self.path, self.name)
self._write_git_file_and_module_config(module_checkout_abspath, module_abspath)
# end handle git file rewrite
# END move physical module
# rename the index entry - have to manipulate the index directly as
# git-mv cannot be used on submodules ... yeah
previous_sm_path = self.path
try:
if configuration:
try:
ekey = index.entry_key(self.path, 0)
entry = index.entries[ekey]
del(index.entries[ekey])
nentry = git.IndexEntry(entry[:3] + (module_checkout_path,) + entry[4:])
index.entries[tekey] = nentry
except KeyError:
raise InvalidGitRepositoryError("Submodule's entry at %r did not exist" % (self.path))
# END handle submodule doesn't exist
# update configuration
with self.config_writer(index=index) as writer: # auto-write
writer.set_value('path', module_checkout_path)
self.path = module_checkout_path
# END handle configuration flag
except Exception:
if renamed_module:
os.renames(module_checkout_abspath, cur_path)
# END undo module renaming
raise
# END handle undo rename
# Auto-rename submodule if it's name was 'default', that is, the checkout directory
if previous_sm_path == self.name:
self.rename(module_checkout_path)
# end
return self
@unbare_repo
def remove(self, module=True, force=False, configuration=True, dry_run=False):
"""Remove this submodule from the repository. This will remove our entry
from the .gitmodules file and the entry in the .git/config file.
:param module: If True, the module checkout we point to will be deleted
as well. If the module is currently on a commit which is not part
of any branch in the remote, if the currently checked out branch
working tree, or untracked files,
is ahead of its tracking branch, if you have modifications in the
In case the removal of the repository fails for these reasons, the
submodule status will not have been altered.
If this submodule has child-modules on its own, these will be deleted
prior to touching the own module.
:param force: Enforces the deletion of the module even though it contains
modifications. This basically enforces a brute-force file system based
deletion.
:param configuration: if True, the submodule is deleted from the configuration,
otherwise it isn't. Although this should be enabled most of the times,
this flag enables you to safely delete the repository of your submodule.
:param dry_run: if True, we will not actually do anything, but throw the errors
we would usually throw
:return: self
:note: doesn't work in bare repositories
:note: doesn't work atomically, as failure to remove any part of the submodule will leave
an inconsistent state
:raise InvalidGitRepositoryError: thrown if the repository cannot be deleted
:raise OSError: if directories or files could not be removed"""
if not (module or configuration):
raise ValueError("Need to specify to delete at least the module, or the configuration")
# END handle parameters
# Recursively remove children of this submodule
nc = 0
for csm in self.children():
nc += 1
csm.remove(module, force, configuration, dry_run)
del(csm)
# end
if configuration and not dry_run and nc > 0:
# Assure we don't leave the parent repository in a dirty state, and commit our changes
# It's important for recursive, unforced, deletions to work as expected
self.module().index.commit("Removed at least one of child-modules of '%s'" % self.name)
# end handle recursion
# DELETE REPOSITORY WORKING TREE
################################
if module and self.module_exists():
mod = self.module()
git_dir = mod.git_dir
if force:
# take the fast lane and just delete everything in our module path
# TODO: If we run into permission problems, we have a highly inconsistent
# state. Delete the .git folders last, start with the submodules first
mp = self.abspath
method = None
if osp.islink(mp):
method = os.remove
elif osp.isdir(mp):
method = rmtree
elif osp.exists(mp):
raise AssertionError("Cannot forcibly delete repository as it was neither a link, nor a directory")
# END handle brutal deletion
if not dry_run:
assert method
method(mp)
# END apply deletion method
else:
# verify we may delete our module
if mod.is_dirty(index=True, working_tree=True, untracked_files=True):
raise InvalidGitRepositoryError(
"Cannot delete module at %s with any modifications, unless force is specified"
% mod.working_tree_dir)
# END check for dirt
# figure out whether we have new commits compared to the remotes
# NOTE: If the user pulled all the time, the remote heads might
# not have been updated, so commits coming from the remote look
# as if they come from us. But we stay strictly read-only and
# don't fetch beforehand.
for remote in mod.remotes:
num_branches_with_new_commits = 0
rrefs = remote.refs
for rref in rrefs:
num_branches_with_new_commits += len(mod.git.cherry(rref)) != 0
# END for each remote ref
# not a single remote branch contained all our commits
if len(rrefs) and num_branches_with_new_commits == len(rrefs):
raise InvalidGitRepositoryError(
"Cannot delete module at %s as there are new commits" % mod.working_tree_dir)
# END handle new commits
# have to manually delete references as python's scoping is
# not existing, they could keep handles open ( on windows this is a problem )
if len(rrefs):
del(rref)
# END handle remotes
del(rrefs)
del(remote)
# END for each remote
# finally delete our own submodule
if not dry_run:
self._clear_cache()
wtd = mod.working_tree_dir
del(mod) # release file-handles (windows)
import gc
gc.collect()
try:
rmtree(wtd)
except Exception as ex:
if HIDE_WINDOWS_KNOWN_ERRORS:
raise SkipTest("FIXME: fails with: PermissionError\n %s", ex)
else:
raise
# END delete tree if possible
# END handle force
if not dry_run and osp.isdir(git_dir):
self._clear_cache()
try:
rmtree(git_dir)
except Exception as ex:
if HIDE_WINDOWS_KNOWN_ERRORS:
raise SkipTest("FIXME: fails with: PermissionError\n %s", ex)
else:
raise
# end handle separate bare repository
# END handle module deletion
# void our data not to delay invalid access
if not dry_run:
self._clear_cache()
# DELETE CONFIGURATION
######################
if configuration and not dry_run:
# first the index-entry
parent_index = self.repo.index
try:
del(parent_index.entries[parent_index.entry_key(self.path, 0)])
except KeyError:
pass
# END delete entry
parent_index.write()
# now git config - need the config intact, otherwise we can't query
# information anymore
with self.repo.config_writer() as writer:
writer.remove_section(sm_section(self.name))
with self.config_writer() as writer:
writer.remove_section()
# END delete configuration
return self
def set_parent_commit(self, commit, check=True):
"""Set this instance to use the given commit whose tree is supposed to
contain the .gitmodules blob.
:param commit:
Commit'ish reference pointing at the root_tree, or None to always point to the
most recent commit
:param check:
if True, relatively expensive checks will be performed to verify
validity of the submodule.
:raise ValueError: if the commit's tree didn't contain the .gitmodules blob.
:raise ValueError:
if the parent commit didn't store this submodule under the current path
:return: self"""
if commit is None:
self._parent_commit = None
return self
# end handle None
pcommit = self.repo.commit(commit)
pctree = pcommit.tree
if self.k_modules_file not in pctree:
raise ValueError("Tree of commit %s did not contain the %s file" % (commit, self.k_modules_file))
# END handle exceptions
prev_pc = self._parent_commit
self._parent_commit = pcommit
if check:
parser = self._config_parser(self.repo, self._parent_commit, read_only=True)
if not parser.has_section(sm_section(self.name)):
self._parent_commit = prev_pc
raise ValueError("Submodule at path %r did not exist in parent commit %s" % (self.path, commit))
# END handle submodule did not exist
# END handle checking mode
# update our sha, it could have changed
# If check is False, we might see a parent-commit that doesn't even contain the submodule anymore.
# in that case, mark our sha as being NULL
try:
self.binsha = pctree[self.path].binsha
except KeyError:
self.binsha = self.NULL_BIN_SHA
# end
self._clear_cache()
return self
@unbare_repo
def config_writer(self, index=None, write=True):
""":return: a config writer instance allowing you to read and write the data
belonging to this submodule into the .gitmodules file.
:param index: if not None, an IndexFile instance which should be written.
defaults to the index of the Submodule's parent repository.
:param write: if True, the index will be written each time a configuration
value changes.
:note: the parameters allow for a more efficient writing of the index,
as you can pass in a modified index on your own, prevent automatic writing,
and write yourself once the whole operation is complete
:raise ValueError: if trying to get a writer on a parent_commit which does not
match the current head commit
:raise IOError: If the .gitmodules file/blob could not be read"""
writer = self._config_parser_constrained(read_only=False)
if index is not None:
writer.config._index = index
writer.config._auto_write = write
return writer
@unbare_repo
def rename(self, new_name):
"""Rename this submodule
:note: This method takes care of renaming the submodule in various places, such as
* $parent_git_dir/config
* $working_tree_dir/.gitmodules
* (git >=v1.8.0: move submodule repository to new name)
As .gitmodules will be changed, you would need to make a commit afterwards. The changed .gitmodules file
will already be added to the index
:return: this submodule instance
"""
if self.name == new_name:
return self
# .git/config
with self.repo.config_writer() as pw:
# As we ourselves didn't write anything about submodules into the parent .git/config,
# we will not require it to exist, and just ignore missing entries.
if pw.has_section(sm_section(self.name)):
pw.rename_section(sm_section(self.name), sm_section(new_name))
# .gitmodules
with self.config_writer(write=True).config as cw:
cw.rename_section(sm_section(self.name), sm_section(new_name))
self._name = new_name
# .git/modules
mod = self.module()
if mod.has_separate_working_tree():
destination_module_abspath = self._module_abspath(self.repo, self.path, new_name)
source_dir = mod.git_dir
# Let's be sure the submodule name is not so obviously tied to a directory
if destination_module_abspath.startswith(mod.git_dir):
tmp_dir = self._module_abspath(self.repo, self.path, str(uuid.uuid4()))
os.renames(source_dir, tmp_dir)
source_dir = tmp_dir
# end handle self-containment
os.renames(source_dir, destination_module_abspath)
self._write_git_file_and_module_config(mod.working_tree_dir, destination_module_abspath)
# end move separate git repository
return self
#} END edit interface
#{ Query Interface
@unbare_repo
def module(self):
""":return: Repo instance initialized from the repository at our submodule path
:raise InvalidGitRepositoryError: if a repository was not available. This could
also mean that it was not yet initialized"""
# late import to workaround circular dependencies
module_checkout_abspath = self.abspath
try:
repo = git.Repo(module_checkout_abspath)
if repo != self.repo:
return repo
# END handle repo uninitialized
except (InvalidGitRepositoryError, NoSuchPathError):
raise InvalidGitRepositoryError("No valid repository at %s" % module_checkout_abspath)
else:
raise InvalidGitRepositoryError("Repository at %r was not yet checked out" % module_checkout_abspath)
# END handle exceptions
def module_exists(self):
""":return: True if our module exists and is a valid git repository. See module() method"""
try:
self.module()
return True
except Exception:
return False
# END handle exception
def exists(self):
"""
:return: True if the submodule exists, False otherwise. Please note that
a submodule may exist (in the .gitmodules file) even though its module
doesn't exist on disk"""
# keep attributes for later, and restore them if we have no valid data
# this way we do not actually alter the state of the object
loc = locals()
for attr in self._cache_attrs:
try:
if hasattr(self, attr):
loc[attr] = getattr(self, attr)
# END if we have the attribute cache
except (cp.NoSectionError, ValueError):
# on PY3, this can happen apparently ... don't know why this doesn't happen on PY2
pass
# END for each attr
self._clear_cache()
try:
try:
self.path
return True
except Exception:
return False
# END handle exceptions
finally:
for attr in self._cache_attrs:
if attr in loc:
setattr(self, attr, loc[attr])
# END if we have a cache
# END reapply each attribute
# END handle object state consistency
@property
def branch(self):
""":return: The branch instance that we are to checkout
:raise InvalidGitRepositoryError: if our module is not yet checked out"""
return mkhead(self.module(), self._branch_path)
@property
def branch_path(self):
"""
:return: full (relative) path as string to the branch we would checkout
from the remote and track"""
return self._branch_path
@property
def branch_name(self):
""":return: the name of the branch, which is the shortest possible branch name"""
# use an instance method, for this we create a temporary Head instance
# which uses a repository that is available at least ( it makes no difference )
return git.Head(self.repo, self._branch_path).name
@property
def url(self):
""":return: The url to the repository which our module-repository refers to"""
return self._url
@property
def parent_commit(self):
""":return: Commit instance with the tree containing the .gitmodules file
:note: will always point to the current head's commit if it was not set explicitly"""
if self._parent_commit is None:
return self.repo.commit()
return self._parent_commit
@property
def name(self):
""":return: The name of this submodule. It is used to identify it within the
.gitmodules file.
:note: by default, the name is the path at which to find the submodule, but
in git-python it should be a unique identifier similar to the identifiers
used for remotes, which allows to change the path of the submodule
easily
"""
return self._name
def config_reader(self):
"""
:return: ConfigReader instance which allows you to qurey the configuration values
of this submodule, as provided by the .gitmodules file
:note: The config reader will actually read the data directly from the repository
and thus does not need nor care about your working tree.
:note: Should be cached by the caller and only kept as long as needed
:raise IOError: If the .gitmodules file/blob could not be read"""
return self._config_parser_constrained(read_only=True)
def children(self):
"""
:return: IterableList(Submodule, ...) an iterable list of submodules instances
which are children of this submodule or 0 if the submodule is not checked out"""
return self._get_intermediate_items(self)
#} END query interface
#{ Iterable Interface
@classmethod
def iter_items(cls, repo, parent_commit='HEAD'):
""":return: iterator yielding Submodule instances available in the given repository"""
pc = repo.commit(parent_commit) # parent commit instance
try:
parser = cls._config_parser(repo, pc, read_only=True)
except IOError:
raise StopIteration
# END handle empty iterator
rt = pc.tree # root tree
for sms in parser.sections():
n = sm_name(sms)
p = parser.get(sms, 'path')
u = parser.get(sms, 'url')
b = cls.k_head_default
if parser.has_option(sms, cls.k_head_option):
b = str(parser.get(sms, cls.k_head_option))
# END handle optional information
# get the binsha
index = repo.index
try:
sm = rt[p]
except KeyError:
# try the index, maybe it was just added
try:
entry = index.entries[index.entry_key(p, 0)]
sm = Submodule(repo, entry.binsha, entry.mode, entry.path)
except KeyError:
raise InvalidGitRepositoryError(
"Gitmodule path %r did not exist in revision of parent commit %s" % (p, parent_commit))
# END handle keyerror
# END handle critical error
# fill in remaining info - saves time as it doesn't have to be parsed again
sm._name = n
if pc != repo.commit():
sm._parent_commit = pc
# end set only if not most recent !
sm._branch_path = git.Head.to_full_path(b)
sm._url = u
yield sm
# END for each section
#} END iterable interface