GitHub Viewer
# Copyright (C) 2008, 2009 Michael Trier (mtrier@gmail.com) and contributors
#
# This module is part of GitPython and is released under the
# 3-Clause BSD License: https://opensource.org/license/bsd-3-clause/
import sys
__all__ = [
"stream_copy",
"join_path",
"to_native_path_linux",
"join_path_native",
"Stats",
"IterableObj",
"IterableList",
"BlockingLockFile",
"LockFile",
"Actor",
"get_user_id",
"assure_directory_exists",
"RemoteProgress",
"CallableRemoteProgress",
"rmtree",
"unbare_repo",
]
if sys.platform == "win32":
__all__.append("to_native_path_windows")
import contextlib
import getpass
import logging
import ntpath
import os
import os.path as osp
import platform
import re
import shutil
import stat
import subprocess
import time
from abc import abstractmethod
from functools import wraps
from pathlib import Path
# typing ---------------------------------------------------------
from typing import (
TYPE_CHECKING,
Any,
AnyStr,
Callable,
Dict,
Generator,
Iterator,
List,
Optional,
Pattern,
Sequence,
Tuple,
Type,
TypeVar,
Union,
cast,
overload,
)
from urllib.parse import urlsplit, urlunsplit
# NOTE: Unused imports can be improved now that CI testing has fully resumed. Some of
# these be used indirectly through other GitPython modules, which avoids having to write
# gitdb all the time in their imports. They are not in __all__, at least currently,
# because they could be removed or changed at any time, and so should not be considered
# conceptually public to code outside GitPython. Linters of course do not like it.
from gitdb.util import (
LazyMixin, # noqa: F401
LockedFD, # noqa: F401
bin_to_hex, # noqa: F401
file_contents_ro, # noqa: F401
file_contents_ro_filepath, # noqa: F401
hex_to_bin, # noqa: F401
)
if TYPE_CHECKING:
from git.cmd import Git
from git.config import GitConfigParser, SectionConstraint
from git.remote import Remote
from git.repo.base import Repo
from git.compat import safe_decode
from git.types import (
HSH_TD,
Files_TD,
Has_id_attribute,
Literal,
PathLike,
Protocol,
SupportsIndex,
SupportsRead,
SupportsWrite,
Total_TD,
runtime_checkable,
)
# ---------------------------------------------------------------------
T_IterableObj = TypeVar("T_IterableObj", bound=Union["IterableObj", "Has_id_attribute"], covariant=True)
# So IterableList[Head] is subtype of IterableList[IterableObj].
T_Actor = TypeVar("T_Actor", bound="Actor")
_logger = logging.getLogger(__name__)
def to_hex_sha(sha: Union[str, bytes]) -> Union[str, bytes]:
"""Normalize legacy SHA-1 or SHA-256 IDs; use ``bin_to_hex`` for raw IDs."""
if len(sha) in (40, 64):
hex_to_bin(sha) # Validate already hexadecimal input.
return sha
if isinstance(sha, bytes) and len(sha) in (20, 32):
return bin_to_hex(sha)
raise ValueError("Expected a binary or hexadecimal SHA-1/SHA-256 object ID")
def to_bin_sha(sha: Union[str, bytes]) -> bytes:
"""Normalize legacy SHA-1 or SHA-256 IDs; use ``hex_to_bin`` for hex IDs."""
if isinstance(sha, bytes) and len(sha) in (20, 32):
return sha
if len(sha) in (40, 64):
return hex_to_bin(sha)
raise ValueError("Expected a binary or hexadecimal SHA-1/SHA-256 object ID")
# { Utility Methods
T = TypeVar("T")
def unbare_repo(func: Callable[..., T]) -> Callable[..., T]:
"""Methods with this decorator raise :exc:`~git.exc.InvalidGitRepositoryError` if
they encounter a bare repository."""
from .exc import InvalidGitRepositoryError
@wraps(func)
def wrapper(self: "Remote", *args: Any, **kwargs: Any) -> T:
if self.repo.bare:
raise InvalidGitRepositoryError("Method '%s' cannot operate on bare repositories" % func.__name__)
# END bare method
return func(self, *args, **kwargs)
# END wrapper
return wrapper
@contextlib.contextmanager
def cwd(new_dir: PathLike) -> Generator[PathLike, None, None]:
"""Context manager to temporarily change directory.
This is similar to :func:`contextlib.chdir` introduced in Python 3.11, but the
context manager object returned by a single call to this function is not reentrant.
"""
old_dir = os.getcwd()
os.chdir(new_dir)
try:
yield new_dir
finally:
os.chdir(old_dir)
@contextlib.contextmanager
def patch_env(name: str, value: str) -> Generator[None, None, None]:
"""Context manager to temporarily patch an environment variable."""
old_value = os.getenv(name)
os.environ[name] = value
try:
yield
finally:
if old_value is None:
del os.environ[name]
else:
os.environ[name] = old_value
def rmtree(path: PathLike) -> None:
"""Remove the given directory tree recursively.
:note:
We use :func:`shutil.rmtree` but adjust its behaviour to see whether files that
couldn't be deleted are read-only. Windows will not remove them in that case.
"""
def handler(function: Callable[[str], Any], path: str, _excinfo: Any) -> None:
"""Callback for :func:`shutil.rmtree`.
This works as either a ``onexc`` or ``onerror`` style callback.
"""
# Is the error an access error?
os.chmod(path, stat.S_IWUSR)
function(path)
if sys.platform != "win32":
shutil.rmtree(path)
elif sys.version_info >= (3, 12):
shutil.rmtree(path, onexc=handler)
else:
shutil.rmtree(path, onerror=handler)
def rmfile(path: PathLike) -> None:
"""Ensure file deleted also on *Windows* where read-only files need special
treatment."""
if osp.isfile(path):
if sys.platform == "win32":
os.chmod(path, 0o777)
os.remove(path)
def stream_copy(source: SupportsRead[AnyStr], destination: SupportsWrite[AnyStr], chunk_size: int = 512 * 1024) -> int:
"""Copy all data from the `source` stream into the `destination` stream in chunks
of size `chunk_size`.
:return:
Number of bytes written
"""
br = 0
while True:
chunk = source.read(chunk_size)
destination.write(chunk)
br += len(chunk)
if len(chunk) < chunk_size:
break
# END reading output stream
return br
def join_path(a: PathLike, *p: PathLike) -> PathLike:
R"""Join path tokens together similar to osp.join, but always use ``/`` instead of
possibly ``\`` on Windows."""
path = os.fspath(a)
for b in p:
b = os.fspath(b)
if not b:
continue
if b.startswith("/"):
path += b[1:]
elif path == "" or path.endswith("/"):
path += b
else:
path += "/" + b
# END for each path token to add
return path
if sys.platform == "win32":
def to_native_path_windows(path: PathLike) -> str:
path = os.fspath(path)
return path.replace("/", "\\")
def to_native_path_linux(path: PathLike) -> str:
path = os.fspath(path)
return path.replace("\\", "/")
to_native_path = to_native_path_windows
else:
# No need for any work on Linux.
def to_native_path_linux(path: PathLike) -> str:
return os.fspath(path)
to_native_path = to_native_path_linux
def join_path_native(a: PathLike, *p: PathLike) -> PathLike:
R"""Like :func:`join_path`, but makes sure an OS native path is returned.
This is only needed to play it safe on Windows and to ensure nice paths that only
use ``\``.
"""
return to_native_path(join_path(a, *p))
def _is_path_rooted(path: PathLike) -> bool:
r"""Whether ``path`` has a root, including one encoded in a UNC drive.
On Windows, ``\directory`` is rooted on the current drive without being
absolute, while ``C:\directory`` has both a drive and a root. In contrast,
``directory`` and the drive-relative ``C:directory`` have no root.
UNC paths are rooted: ``\\server\share`` stores the share in the drive
returned by :func:`os.path.splitdrive`, while ``\\server\share\directory``
additionally has a rooted tail.
On POSIX, which has no drive concept, this simply distinguishes absolute
paths such as ``/directory`` from relative paths such as ``directory``.
"""
drive, tail = osp.splitdrive(os.fspath(path))
separators = (os.sep,) if os.altsep is None else (os.sep, os.altsep)
return tail.startswith(separators) or drive.startswith(separators)
def _to_relative_path(root: PathLike, path: PathLike) -> str:
r"""Return a normalized Git-style path confined to ``root``.
A Windows path such as ``\directory`` is rooted but not absolute. Resolve it
against the drive of ``root`` rather than treating it as relative to ``root``.
Drive-relative paths such as ``C:directory`` are rejected because their meaning
depends on process-global per-drive state.
For example, with ``root`` set to ``C:\repo`` on Windows:
* ``directory\file`` -> ``directory/file``
* ``directory\`` -> ``directory/``
* ``C:\repo\directory\file`` -> ``directory/file``
* ``\repo\directory\file`` -> ``directory/file``
* ``C:directory\file`` -> :exc:`ValueError`
* ``C:\other\file`` -> :exc:`ValueError`
On POSIX, ``/repo/directory/file`` under ``/repo`` similarly becomes
``directory/file``. A trailing separator is preserved as a Git-style ``/``.
"""
path_str = os.fspath(path)
if not path_str:
return path_str
drive, _tail = osp.splitdrive(path_str)
rooted = _is_path_rooted(path_str)
if drive and not rooted:
raise ValueError("Drive-relative path %r is not supported" % path_str)
root_abs = osp.abspath(os.fspath(root))
path_abs = osp.abspath(osp.join(root_abs, path_str))
try:
common_path = osp.commonpath([root_abs, path_abs])
except ValueError as e:
raise ValueError("Path %r is not in repository at %r" % (path_str, root_abs)) from e
if common_path != root_abs:
raise ValueError("Path %r is not in repository at %r" % (path_str, root_abs))
relative_path = to_native_path_linux(osp.relpath(path_abs, root_abs))
separators = (os.sep,) if os.altsep is None else (os.sep, os.altsep)
if path_str.endswith(separators) and relative_path != "." and not relative_path.endswith("/"):
relative_path += "/"
return relative_path
_HFS_IGNORABLES = str.maketrans(
"", "", "\u200c\u200d\u200e\u200f\u202a\u202b\u202c\u202d\u202e\u206a\u206b\u206c\u206d\u206e\u206f\ufeff"
)
# Match Git's is_ntfs_dotgitmodules in path.c on a lowercased, trimmed name.
# Besides gitmod~1..4, fallback aliases have exactly eight ASCII characters:
# a shrinking prefix of "gi7eba", "~", and digits with no leading zero.
# Explicit digit counts avoid accepting shorter/longer names or Unicode digits.
_NTFS_DOTGITMODULES_SHORT_NAME = re.compile(
r"(?:gitmod~[1-4]|gi7eba~[1-9]|gi7eb~[1-9][0-9]|gi7e~[1-9][0-9]{2}|"
r"gi7~[1-9][0-9]{3}|gi~[1-9][0-9]{4}|g~[1-9][0-9]{5}|~[1-9][0-9]{6})"
)
def _validate_repo_path(path: PathLike, mode: Union[int, None] = None) -> None:
"""Reject unsafe tree/index paths without normalizing away their components.
Protect Git metadata aliases on NTFS and HFS even when writing on another
platform. Other POSIX filename characters, including newlines, remain valid.
:param mode:
Mode of the index or tree entry the path belongs to, where one is known.
Git refuses a symbolic link that aliases ``.gitmodules``, since the
submodule configuration would then be read through the link, so that
name is only rejected once the mode says the entry is a link.
"""
name = os.fspath(path)
if not name or "\0" in name or ntpath.splitdrive(name)[0] or name.startswith("/"):
raise ValueError("Invalid repository path %r" % name)
if os.name == "nt" and "\\" in name:
raise ValueError("Index paths must use '/' separators: %r" % name)
for component in name.split("/"):
if component in ("", ".", ".."):
raise ValueError("Invalid repository path %r" % name)
# NTFS recognizes backslashes and alternate data streams in metadata names.
for part in component.split("\\"):
ntfs_name = part.split(":", 1)[0].rstrip(" .").lower()
hfs_name = part.translate(_HFS_IGNORABLES).lower()
if ntfs_name in (".git", "git~1") or hfs_name == ".git":
raise ValueError("Repository path aliases Git metadata: %r" % name)
if mode is not None and stat.S_ISLNK(mode):
if (
ntfs_name == ".gitmodules"
or hfs_name == ".gitmodules"
or _NTFS_DOTGITMODULES_SHORT_NAME.fullmatch(ntfs_name) is not None
):
raise ValueError("Symbolic link aliases the submodule configuration: %r" % name)
def assure_directory_exists(path: PathLike, is_file: bool = False) -> bool:
"""Make sure that the directory pointed to by path exists.
:param is_file:
If ``True``, `path` is assumed to be a file and handled correctly.
Otherwise it must be a directory.
:return:
``True`` if the directory was created, ``False`` if it already existed.
"""
if is_file:
path = osp.dirname(path)
# END handle file
if not osp.isdir(path):
os.makedirs(path, exist_ok=True)
return True
return False
def _get_exe_extensions() -> Sequence[str]:
PATHEXT = os.environ.get("PATHEXT", None)
if PATHEXT:
return tuple(p.upper() for p in PATHEXT.split(os.pathsep))
elif sys.platform == "win32":
return (".BAT", ".COM", ".EXE")
else:
return ()
def py_where(program: str, path: Optional[PathLike] = None) -> List[str]:
"""Perform a path search to assist :func:`is_cygwin_git`.
This is not robust for general use. It is an implementation detail of
:func:`is_cygwin_git`. When a search following all shell rules is needed,
:func:`shutil.which` can be used instead.
:note:
Neither this function nor :func:`shutil.which` will predict the effect of an
executable search on a native Windows system due to a :class:`subprocess.Popen`
call without ``shell=True``, because shell and non-shell executable search on
Windows differ considerably.
"""
# From: http://stackoverflow.com/a/377028/548792
winprog_exts = _get_exe_extensions()
def is_exec(fpath: str) -> bool:
return (
osp.isfile(fpath)
and os.access(fpath, os.X_OK)
and (
sys.platform != "win32" or not winprog_exts or any(fpath.upper().endswith(ext) for ext in winprog_exts)
)
)
progs = []
if not path:
path = os.environ["PATH"]
for folder in os.fspath(path).split(os.pathsep):
folder = folder.strip('"')
if folder:
exe_path = osp.join(folder, program)
for f in [exe_path] + ["%s%s" % (exe_path, e) for e in winprog_exts]:
if is_exec(f):
progs.append(f)
return progs
def _cygexpath(drive: Optional[str], path: str) -> str:
if osp.isabs(path) and not drive:
# Invoked from `cygpath()` directly with `D:Apps\123`?
# It's an error, leave it alone just slashes)
p = path # convert to str if AnyPath given
else:
p = path and osp.normpath(osp.expanduser(path))
if osp.isabs(p):
if drive:
# Confusing, maybe a remote system should expand vars.
p = path
else:
p = cygpath(p)
elif drive:
p = "/proc/cygdrive/%s/%s" % (drive.lower(), p)
p_str = os.fspath(p) # ensure it is a str and not AnyPath
return p_str.replace("\\", "/")
_cygpath_parsers: Tuple[Tuple[Pattern[str], Callable[..., str], bool], ...] = (
# See: https://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx
# and: https://www.cygwin.com/cygwin-ug-net/using.html#unc-paths
(
re.compile(r"\\\\\?\\UNC\\([^\\]+)\\([^\\]+)(?:\\(.*))?"),
(lambda server, share, rest_path: "//%s/%s/%s" % (server, share, rest_path.replace("\\", "/"))),
False,
),
(re.compile(r"\\\\\?\\(\w):[/\\](.*)"), (_cygexpath), False),
(re.compile(r"(\w):[/\\](.*)"), (_cygexpath), False),
(re.compile(r"file:(.*)", re.I), (lambda rest_path: rest_path), True),
(re.compile(r"(\w{2,}:.*)"), (lambda url: url), False), # remote URL, do nothing
)
def cygpath(path: str) -> str:
"""Use :meth:`git.cmd.Git.polish_url` instead, that works on any environment."""
path = os.fspath(path) # Ensure is str and not AnyPath.
# Fix to use Paths when 3.5 dropped. Or to be just str if only for URLs?
if not path.startswith(("/cygdrive", "//", "/proc/cygdrive")):
for regex, parser, recurse in _cygpath_parsers:
match = regex.match(path)
if match:
if parser is _cygexpath:
path = parser(*match.groups())
else:
path = parser(*match.groups())
if recurse:
path = cygpath(path)
break
else:
path = _cygexpath(None, path)
return path
_decygpath_regex = re.compile(r"(?:/proc)?/cygdrive/(\w)(/.*)?")
def decygpath(path: PathLike) -> str:
path = os.fspath(path)
m = _decygpath_regex.match(path)
if m:
drive, rest_path = m.groups()
path = "%s:%s" % (drive.upper(), rest_path or "")
return path.replace("/", "\\")
#: Store boolean flags denoting if a specific Git executable
#: is from a Cygwin installation. TODO: use @functools.lru_cache(user_function)
_is_cygwin_cache: Dict[str, Optional[bool]] = {}
def _is_cygwin_git(git_executable: str) -> bool:
is_cygwin = _is_cygwin_cache.get(git_executable) # type: Optional[bool]
if is_cygwin is None:
is_cygwin = False
try:
git_dir = osp.dirname(git_executable)
if not git_dir:
res = py_where(git_executable)
git_dir = osp.dirname(res[0]) if res else ""
# Just a name given, not a real path.
uname_cmd = osp.join(git_dir, "uname")
if not (Path(uname_cmd).is_file() and os.access(uname_cmd, os.X_OK)):
_logger.debug(f"Failed checking if running in CYGWIN: {uname_cmd} is not an executable")
_is_cygwin_cache[git_executable] = is_cygwin
return is_cygwin
process = subprocess.Popen([uname_cmd], stdout=subprocess.PIPE, universal_newlines=True)
uname_out, _ = process.communicate()
# retcode = process.poll()
is_cygwin = "CYGWIN" in uname_out
except Exception as ex:
_logger.debug("Failed checking if running in CYGWIN due to: %r", ex)
_is_cygwin_cache[git_executable] = is_cygwin
return is_cygwin
@overload
def is_cygwin_git(git_executable: None) -> Literal[False]: ...
@overload
def is_cygwin_git(git_executable: PathLike) -> bool: ...
def is_cygwin_git(git_executable: Union[None, PathLike]) -> bool:
_logger.debug(f"{sys.platform=}, {git_executable=}")
if sys.platform != "cygwin":
return False
elif git_executable is None:
return False
else:
return _is_cygwin_git(str(git_executable))
def get_user_id() -> str:
""":return: String identifying the currently active system user as ``name@node``"""
return "%s@%s" % (getpass.getuser(), platform.node())
def finalize_process(proc: Union["subprocess.Popen[Any]", "Git.AutoInterrupt"], **kwargs: Any) -> None:
"""Wait for the process (clone, fetch, pull or push) and handle its errors
accordingly."""
# TODO: No close proc-streams??
proc.wait(**kwargs)
@overload
def expand_path(p: None) -> None: ...
@overload
def expand_path(p: PathLike) -> Optional[PathLike]:
# TODO: Support for Python 3.5 has been dropped, so these overloads can be improved.
...
def expand_path(p: Union[None, PathLike]) -> Optional[PathLike]:
if p is None:
return None
try:
expanded_path = osp.expanduser(os.fspath(p))
return osp.normpath(osp.abspath(expanded_path))
except Exception:
return None
def _redact_url_credentials(text: str) -> Optional[str]:
"""Return `text` with the userinfo of a URL redacted, or ``None`` if it has none."""
try:
url = urlsplit(text)
except ValueError:
# This is not a valid URL.
return None
if url.password is None and url.username is None:
return None
# Match urllib.parse's userinfo boundary. Keeping the raw hostinfo
# preserves hostname case, IPv6 brackets, and port formatting.
_, _, hostinfo = url.netloc.rpartition("@")
redacted = "*****:*****" if url.password is not None else "*****"
return urlunsplit(url._replace(netloc=f"{redacted}@{hostinfo}"))
def remove_password_if_present(cmdline: Sequence[str]) -> List[str]:
"""Redact credentials in URLs and HTTP Authorization extra headers in a command line.
URLs are also redacted in the key and value of a configuration assignment, such
as ``http.proxy=``. Assignments are recognized in the ``-c =``,
``-c=``, and ``--config==`` forms.
If nothing is found, this just returns the command line as-is.
This should be used for every log line that print a command line, as well as
exception messages.
"""
new_cmdline = []
for to_parse in cmdline:
if to_parse.startswith("--config="):
option = "--config="
elif to_parse.startswith("-c"):
option = "-c"
else:
option = ""
assignment = to_parse[len(option) :]
# Redact the whole URL first, before splitting on a possible '=' in userinfo.
assignment = _redact_url_credentials(assignment) or assignment
config_key, separator, value = assignment.partition("=")
header_name, colon, _ = value.partition(":")
if (
separator
and colon
and config_key.lower().endswith(".extraheader")
and header_name.strip().lower() == "authorization"
):
new_cmdline.append("%s%s%s%s%s *****" % (option, config_key, separator, header_name, colon))
continue
if separator:
value = _redact_url_credentials(value) or value
new_cmdline.append(option + config_key + separator + value)
return new_cmdline
# } END utilities
# { Classes
class RemoteProgress:
"""Handler providing an interface to parse progress information emitted by
:manpage:`git-push(1)` and :manpage:`git-fetch(1)` and to dispatch callbacks
allowing subclasses to react to the progress."""
_num_op_codes: int = 9
(
BEGIN,
END,
COUNTING,
COMPRESSING,
WRITING,
RECEIVING,
RESOLVING,
FINDING_SOURCES,
CHECKING_OUT,
) = [1 None:
self._seen_ops: List[int] = []
self._cur_line: Optional[str] = None
self.error_lines: List[str] = []
self.other_lines: List[str] = []
def _parse_progress_line(self, line: AnyStr) -> Optional[object]:
"""Parse progress information from the given line as retrieved by
:manpage:`git-push(1)` or :manpage:`git-fetch(1)`.
:meta public:
- Lines that do not contain progress info are stored in :attr:`other_lines`.
- Lines that seem to contain an error (i.e. start with ``error:`` or ``fatal:``)
are stored in :attr:`error_lines`.
The base implementation returns ``None``. Subclasses may return another
value for compatibility with existing overrides, but callers should treat
the return value as unspecified.
"""
# handle
# Counting objects: 4, done.
# Compressing objects: 50% (1/2)
# Compressing objects: 100% (2/2)
# Compressing objects: 100% (2/2), done.
if isinstance(line, bytes): # mypy argues about ternary assignment.
line_str = safe_decode(line)
else:
line_str = line
self._cur_line = line_str
if self._cur_line.startswith(("error:", "fatal:")):
self.error_lines.append(self._cur_line)
return None
cur_count, max_count = None, None
match = self.re_op_relative.match(line_str)
if match is None:
match = self.re_op_absolute.match(line_str)
if not match:
self.line_dropped(line_str)
self.other_lines.append(line_str)
return None
# END could not get match
op_code = 0
_remote, op_name, _percent, cur_count, max_count, message = match.groups()
# Get operation ID.
if op_name == "Counting objects":
op_code |= self.COUNTING
elif op_name == "Compressing objects":
op_code |= self.COMPRESSING
elif op_name == "Writing objects":
op_code |= self.WRITING
elif op_name == "Receiving objects":
op_code |= self.RECEIVING
elif op_name == "Resolving deltas":
op_code |= self.RESOLVING
elif op_name == "Finding sources":
op_code |= self.FINDING_SOURCES
elif op_name == "Checking out files":
op_code |= self.CHECKING_OUT
else:
# Note: On Windows it can happen that partial lines are sent.
# Hence we get something like "CompreReceiving objects", which is
# a blend of "Compressing objects" and "Receiving objects".
# This can't really be prevented, so we drop the line verbosely
# to make sure we get informed in case the process spits out new
# commands at some point.
self.line_dropped(line_str)
# Note: Don't add this line to the other lines, as we have to silently
# drop it.
return None
# END handle op code
# Figure out stage.
if op_code not in self._seen_ops:
self._seen_ops.append(op_code)
op_code |= self.BEGIN
# END begin opcode
if message is None:
message = ""
# END message handling
message = message.strip()
if message.endswith(self.DONE_TOKEN):
op_code |= self.END
message = message[: -len(self.DONE_TOKEN)]
# END end message handling
message = message.strip(self.TOKEN_SEPARATOR)
self.update(
op_code,
cur_count and float(cur_count),
max_count and float(max_count),
message,
)
return None
def new_message_handler(self) -> Callable[[str], None]:
"""
:return:
A progress handler suitable for :func:`~git.cmd.handle_process_output`,
passing lines on to this progress handler in a suitable format.
"""
def handler(line: AnyStr) -> None:
self._parse_progress_line(line.rstrip(b"\r\n") if isinstance(line, bytes) else line.rstrip("\r\n"))
# END handler
return handler
def line_dropped(self, line: str) -> None:
"""Called whenever a line could not be understood and was therefore dropped."""
pass
def update(
self,
op_code: int,
cur_count: Union[str, float],
max_count: Union[str, float, None] = None,
message: str = "",
) -> None:
"""Called whenever the progress changes.
:param op_code:
Integer allowing to be compared against Operation IDs and stage IDs.
Stage IDs are :const:`BEGIN` and :const:`END`. :const:`BEGIN` will only be
set once for each Operation ID as well as :const:`END`. It may be that
:const:`BEGIN` and :const:`END` are set at once in case only one progress
message was emitted due to the speed of the operation. Between
:const:`BEGIN` and :const:`END`, none of these flags will be set.
Operation IDs are all held within the :const:`OP_MASK`. Only one Operation
ID will be active per call.
:param cur_count:
Current absolute count of items.
:param max_count:
The maximum count of items we expect. It may be ``None`` in case there is no
maximum number of items or if it is (yet) unknown.
:param message:
In case of the :const:`WRITING` operation, it contains the amount of bytes
transferred. It may possibly be used for other purposes as well.
:note:
You may read the contents of the current line in
:attr:`self._cur_line `.
"""
pass
class CallableRemoteProgress(RemoteProgress):
"""A :class:`RemoteProgress` implementation forwarding updates to any callable.
:note:
Like direct instances of :class:`RemoteProgress`, instances of this
:class:`CallableRemoteProgress` class are not themselves directly callable.
Rather, instances of this class wrap a callable and forward to it. This should
therefore not be confused with :class:`git.types.CallableProgress`.
"""
__slots__ = ("_callable",)
def __init__(self, fn: Callable[..., Any]) -> None:
self._callable = fn
super().__init__()
def update(self, *args: Any, **kwargs: Any) -> None:
self._callable(*args, **kwargs)
class Actor:
"""Actors hold information about a person acting on the repository. They can be
committers and authors or anything with a name and an email as mentioned in the git
log entries."""
# ENVIRONMENT VARIABLES
# These are read when creating new commits.
env_author_name = "GIT_AUTHOR_NAME"
env_author_email = "GIT_AUTHOR_EMAIL"
env_committer_name = "GIT_COMMITTER_NAME"
env_committer_email = "GIT_COMMITTER_EMAIL"
# CONFIGURATION KEYS
conf_name = "name"
conf_email = "email"
__slots__ = ("name", "email")
def __init__(self, name: Optional[str], email: Optional[str]) -> None:
self.name = name
self.email = email
def __eq__(self, other: Any) -> bool:
return self.name == other.name and self.email == other.email
def __ne__(self, other: Any) -> bool:
return not (self == other)
def __hash__(self) -> int:
return hash((self.name, self.email))
def __str__(self) -> str:
return self.name if self.name else ""
def __repr__(self) -> str:
return '' % (self.name, self.email)
@classmethod
def from_string(cls: Type[T_Actor], string: str) -> T_Actor:
"""Create an :class:`Actor` from a string.
:param string:
The string, which is expected to be in regular git format::
John Doe
:return:
:class:`Actor`
"""
line = string.partition("\n")[0]
left_bracket = line.find("", left_bracket + 1)
if left_bracket >= 0 and right_bracket >= 0:
return cls(line[:left_bracket].rstrip(), line[left_bracket + 1 : right_bracket])
# Assume the best and use the whole string as name.
return cls(string, None)
_from_string = from_string
@classmethod
def _main_actor(
cls,
env_name: str,
env_email: str,
config_reader: Union[None, "GitConfigParser", "SectionConstraint[GitConfigParser]"] = None,
) -> "Actor":
actor = Actor("", "")
user_id = None # We use this to avoid multiple calls to getpass.getuser().
def default_email() -> str:
nonlocal user_id
if not user_id:
user_id = get_user_id()
return user_id
def default_name() -> str:
return default_email().split("@")[0]
for attr, evar, cvar, default in (
("name", env_name, cls.conf_name, default_name),
("email", env_email, cls.conf_email, default_email),
):
try:
val = os.environ[evar]
setattr(actor, attr, val)
except KeyError:
if config_reader is not None:
try:
val = config_reader.get("user", cvar)
except Exception:
val = default()
setattr(actor, attr, val)
# END config-reader handling
if not getattr(actor, attr):
setattr(actor, attr, default())
# END handle name
# END for each item to retrieve
return actor
@classmethod
def committer(
cls, config_reader: Union[None, "GitConfigParser", "SectionConstraint[GitConfigParser]"] = None
) -> "Actor":
"""
:return:
:class:`Actor` instance corresponding to the configured committer. It
behaves similar to the git implementation, such that the environment will
override configuration values of `config_reader`. If no value is set at all,
it will be generated.
:param config_reader:
ConfigReader to use to retrieve the values from in case they are not set in
the environment.
"""
return cls._main_actor(cls.env_committer_name, cls.env_committer_email, config_reader)
@classmethod
def author(
cls, config_reader: Union[None, "GitConfigParser", "SectionConstraint[GitConfigParser]"] = None
) -> "Actor":
"""Same as :meth:`committer`, but defines the main author. It may be specified
in the environment, but defaults to the committer."""
return cls._main_actor(cls.env_author_name, cls.env_author_email, config_reader)
class Stats:
"""Represents stat information as presented by git at the end of a merge. It is
created from the output of a diff operation.
Example::
c = Commit( sha1 )
s = c.stats
s.total # full-stat-dict
s.files # dict( filepath : stat-dict )
``stat-dict``
A dictionary with the following keys and values::
deletions = number of deleted lines as int
insertions = number of inserted lines as int
lines = total number of lines changed as int, or deletions + insertions
change_type = type of change as str, A|C|D|M|R|T|U|X|B
``full-stat-dict``
In addition to the items in the stat-dict, it features additional information::
files = number of changed files as int
"""
__slots__ = ("total", "files")
def __init__(self, total: Total_TD, files: Dict[PathLike, Files_TD]) -> None:
self.total = total
self.files = files
@classmethod
def _list_from_string(cls, repo: "Repo", text: str) -> "Stats":
"""Create a :class:`Stats` object from output retrieved by
:manpage:`git-diff(1)`.
:return:
:class:`git.Stats`
"""
hsh: HSH_TD = {
"total": {"insertions": 0, "deletions": 0, "lines": 0, "files": 0},
"files": {},
}
for line in text.split("\n"):
if not line:
continue
(change_type, raw_insertions, raw_deletions, filename) = line.split("\t")
insertions = raw_insertions != "-" and int(raw_insertions) or 0
deletions = raw_deletions != "-" and int(raw_deletions) or 0
hsh["total"]["insertions"] += insertions
hsh["total"]["deletions"] += deletions
hsh["total"]["lines"] += insertions + deletions
hsh["total"]["files"] += 1
files_dict: Files_TD = {
"insertions": insertions,
"deletions": deletions,
"lines": insertions + deletions,
"change_type": change_type,
}
hsh["files"][filename] = files_dict
return Stats(hsh["total"], hsh["files"])
class LockFile:
"""Provides methods to obtain, check for, and release a file based lock which
should be used to handle concurrent access to the same file.
As we are a utility class to be derived from, we only use protected methods.
Locks will automatically be released on destruction.
"""
__slots__ = ("_file_path", "_owns_lock")
def __init__(self, file_path: PathLike) -> None:
self._file_path = file_path
self._owns_lock = False
def __del__(self) -> None:
self._release_lock()
def _lock_file_path(self) -> str:
""":return: Path to lockfile"""
return "%s.lock" % (self._file_path)
def _has_lock(self) -> bool:
"""
:return:
True if we have a lock and if the lockfile still exists
:raise AssertionError:
If our lock-file does not exist.
"""
return self._owns_lock
def _obtain_lock_or_raise(self) -> None:
"""Create a lock file as flag for other instances, mark our instance as
lock-holder.
:raise IOError:
If a lock was already present or a lock file could not be written.
"""
if self._has_lock():
return
lock_file = self._lock_file_path()
# Create the lock in one step. Checking for it first would allow another
# holder to create it between the check and the open.
try:
if sys.platform == "win32":
if "\0" in lock_file:
raise ValueError("embedded null character")
import ctypes
from ctypes import wintypes
# Unlike POSIX, Windows follows dangling symlinks even with O_EXCL.
# Open the reparse point itself so an existing link is rejected.
# Call the Unicode API directly: older _winapi.CreateFile wrappers
# use the ANSI API and can create a lock under the wrong filename.
kernel32 = ctypes.WinDLL("kernel32", use_last_error=True)
create_file = kernel32.CreateFileW
create_file.argtypes = (
wintypes.LPCWSTR,
wintypes.DWORD,
wintypes.DWORD,
wintypes.LPVOID,
wintypes.DWORD,
wintypes.DWORD,
wintypes.HANDLE,
)
create_file.restype = wintypes.HANDLE
close_handle = kernel32.CloseHandle
close_handle.argtypes = (wintypes.HANDLE,)
close_handle.restype = wintypes.BOOL
handle = create_file(
lock_file,
0x40000000, # GENERIC_WRITE
0,
None,
1, # CREATE_NEW
0x00200000, # FILE_FLAG_OPEN_REPARSE_POINT
None,
)
if handle == wintypes.HANDLE(-1).value:
raise ctypes.WinError(ctypes.get_last_error())
if not close_handle(handle):
raise ctypes.WinError(ctypes.get_last_error())
else:
fd = os.open(lock_file, os.O_WRONLY | os.O_CREAT | os.O_EXCL, 0o600)
os.close(fd)
except FileExistsError as e:
raise OSError(
"Lock for file %r did already exist, delete %r in case the lock is illegal"
% (self._file_path, lock_file)
) from e
except OSError as e:
raise OSError(str(e)) from e
self._owns_lock = True
def _obtain_lock(self) -> None:
"""The default implementation will raise if a lock cannot be obtained.
:meta public:
Subclasses may override this method to provide a different implementation.
"""
return self._obtain_lock_or_raise()
def _release_lock(self) -> None:
"""Release our lock if we have one."""
if not self._has_lock():
return
# If someone removed our file beforehand, lets just flag this issue instead of
# failing, to make it more usable.
lfp = self._lock_file_path()
try:
rmfile(lfp)
except OSError:
pass
self._owns_lock = False
class BlockingLockFile(LockFile):
"""The lock file will block until a lock could be obtained, or fail after a
specified timeout.
:note:
Only an existing lock causes retries. Other errors, including a removed or
unwritable parent directory, are raised immediately.
"""
__slots__ = ("_check_interval", "_max_block_time")
def __init__(
self,
file_path: PathLike,
check_interval_s: float = 0.3,
max_block_time_s: int = sys.maxsize,
) -> None:
"""Configure the instance.
:param check_interval_s:
Period of time to sleep until the lock is checked the next time.
By default, it waits a nearly unlimited time.
:param max_block_time_s:
Maximum amount of seconds we may lock.
"""
super().__init__(file_path)
self._check_interval = check_interval_s
self._max_block_time = max_block_time_s
def _obtain_lock(self) -> None:
"""This method blocks until it obtained the lock, or raises :exc:`IOError` if it
ran out of time or lock creation failed for a reason other than an existing lock.
If this method returns, you are guaranteed to own the lock.
"""
starttime = time.time()
maxtime = starttime + float(self._max_block_time)
while True:
try:
super()._obtain_lock()
except OSError as e:
# LockFile retains the creation error as the cause of its OSError.
# Only an existing lock is contention; other errors must not be retried.
if not isinstance(e.__cause__, FileExistsError):
raise
# Abort if the directory containing the existing lock has disappeared.
curtime = time.time()
if not osp.isdir(osp.dirname(self._lock_file_path()) or osp.curdir):
msg = "Directory containing the lockfile %r was not readable anymore after waiting %g seconds" % (
self._lock_file_path(),
curtime - starttime,
)
raise OSError(msg) from e
# END handle missing directory
if curtime >= maxtime:
msg = "Waited %g seconds for lock at %r" % (
maxtime - starttime,
self._lock_file_path(),
)
raise OSError(msg) from e
# END abort if we wait too long
time.sleep(self._check_interval)
else:
break
# END endless loop
class IterableList(List[T_IterableObj]): # type: ignore[type-var]
"""List of iterable objects allowing to query an object by id or by named index::
heads = repo.heads
heads.master
heads['master']
heads[0]
Iterable parent objects:
* :class:`Commit `
* :class:`Submodule `
* :class:`Reference `
* :class:`FetchInfo `
* :class:`PushInfo `
Iterable via inheritance:
* :class:`Head `
* :class:`TagReference `
* :class:`RemoteReference `
This requires an ``id_attribute`` name to be set which will be queried from its
contained items to have a means for comparison.
A prefix can be specified which is to be used in case the id returned by the items
always contains a prefix that does not matter to the user, so it can be left out.
"""
__slots__ = ("_id_attr", "_prefix")
def __new__(cls, id_attr: str, prefix: str = "") -> "IterableList[T_IterableObj]":
return super().__new__(cls)
def __init__(self, id_attr: str, prefix: str = "") -> None:
super().__init__()
self._id_attr = id_attr
self._prefix = prefix
def __contains__(self, attr: object) -> bool:
# First try identity match for performance.
try:
rval = list.__contains__(self, attr)
if rval:
return rval
except (AttributeError, TypeError):
pass
# END handle match
# Otherwise make a full name search.
try:
self.__getattr__(cast(str, attr)) # Use cast to silence mypy.
return True
except (AttributeError, TypeError):
return False
# END handle membership
def __getattr__(self, attr: str) -> T_IterableObj:
attr = self._prefix + attr
for item in self:
if getattr(item, self._id_attr) == attr:
return item
# END for each item
raise AttributeError(f"No item found with id {attr}")
def __getitem__( # type: ignore[override] # pyright: ignore[reportIncompatibleMethodOverride]
self, index: Union[SupportsIndex, int, slice, str]
) -> T_IterableObj:
if isinstance(index, int):
return list.__getitem__(self, index)
elif isinstance(index, slice):
raise ValueError("Index should be an int or str")
else:
try:
return self.__getattr__(cast(str, index))
except AttributeError as e:
raise IndexError(f"No item found with id {self._prefix}{index}") from e
# END handle getattr
def __delitem__(self, index: Union[SupportsIndex, int, slice, str]) -> None:
delindex = cast(int, index)
if isinstance(index, str):
delindex = -1
name = self._prefix + index
for i, item in enumerate(self):
if getattr(item, self._id_attr) == name:
delindex = i
break
# END search index
# END for each item
if delindex == -1:
raise IndexError("Item with name %s not found" % name)
# END handle error
# END get index to delete
list.__delitem__(self, delindex)
@runtime_checkable
class IterableObj(Protocol):
"""Defines an interface for iterable items, so there is a uniform way to retrieve
and iterate items within the git repository.
Subclasses:
* :class:`Submodule `
* :class:`Commit `
* :class:`Reference `
* :class:`PushInfo `
* :class:`FetchInfo `
* :class:`Remote `
"""
__slots__ = ()
_id_attribute_: str
@classmethod
@abstractmethod
def iter_items(cls, repo: "Repo", *args: Any, **kwargs: Any) -> Iterator[T_IterableObj]:
# Return-typed to be compatible with subtypes e.g. Remote.
"""Find (all) items of this type.
Subclasses can specify `args` and `kwargs` differently, and may use them for
filtering. However, when the method is called with no additional positional or
keyword arguments, subclasses are obliged to to yield all items.
:return:
Iterator yielding Items
"""
raise NotImplementedError("To be implemented by Subclass")
@classmethod
def list_items(cls, repo: "Repo", *args: Any, **kwargs: Any) -> IterableList[T_IterableObj]:
"""Find (all) items of this type and collect them into a list.
For more information about the arguments, see :meth:`iter_items`.
:note:
Favor the :meth:`iter_items` method as it will avoid eagerly collecting all
items. When there are many items, that can slow performance and increase
memory usage.
:return:
list(Item,...) list of item instances
"""
out_list: IterableList[T_IterableObj] = IterableList(cls._id_attribute_)
out_list.extend(cls.iter_items(repo, *args, **kwargs))
return out_list
# } END classes