# Copyright (C) 2008, 2009 Michael Trier ([email protected]) 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",
"IndexFileSHA1Writer",
"IterableObj",
"IterableList",
"BlockingLockFile",
"LockFile",
"Actor",
"get_user_id",
"assure_directory_exists",
"RemoteProgress",
"CallableRemoteProgress",
"rmtree",
"unbare_repo",
"HIDE_WINDOWS_KNOWN_ERRORS",
]
if sys.platform == "win32":
__all__.append("to_native_path_windows")
from abc import abstractmethod
import contextlib
from functools import wraps
import getpass
import logging
import os
import os.path as osp
from pathlib import Path
import platform
import re
import shutil
import stat
import subprocess
import time
from urllib.parse import urlsplit, urlunsplit
import warnings
# 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
make_sha,
to_bin_sha, # noqa: F401
to_hex_sha, # noqa: F401
)
# typing ---------------------------------------------------------
from typing import (
Any,
AnyStr,
BinaryIO,
Callable,
Dict,
Generator,
IO,
Iterator,
List,
Optional,
Pattern,
Sequence,
Tuple,
TYPE_CHECKING,
TypeVar,
Union,
cast,
overload,
)
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.types import (
Files_TD,
Has_id_attribute,
HSH_TD,
Literal,
PathLike,
Protocol,
SupportsIndex,
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].
_logger = logging.getLogger(__name__)
def _read_env_flag(name: str, default: bool) -> bool:
"""Read a boolean flag from an environment variable.
:return:
The flag, or the `default` value if absent or ambiguous.
"""
try:
value = os.environ[name]
except KeyError:
return default
_logger.warning(
"The %s environment variable is deprecated. Its effect has never been documented and changes without warning.",
name,
)
adjusted_value = value.strip().lower()
if adjusted_value in {"", "0", "false", "no"}:
return False
if adjusted_value in {"1", "true", "yes"}:
return True
_logger.warning("%s has unrecognized value %r, treating as %r.", name, value, default)
return default
def _read_win_env_flag(name: str, default: bool) -> bool:
"""Read a boolean flag from an environment variable on Windows.
:return:
On Windows, the flag, or the `default` value if absent or ambiguous.
On all other operating systems, ``False``.
:note:
This only accesses the environment on Windows.
"""
return sys.platform == "win32" and _read_env_flag(name, default)
#: We need an easy way to see if Appveyor TCs start failing,
#: so the errors marked with this var are considered "acknowledged" ones, awaiting remedy,
#: till then, we wish to hide them.
HIDE_WINDOWS_KNOWN_ERRORS = _read_win_env_flag("HIDE_WINDOWS_KNOWN_ERRORS", True)
HIDE_WINDOWS_FREEZE_ERRORS = _read_win_env_flag("HIDE_WINDOWS_FREEZE_ERRORS", True)
# { 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, path: PathLike, _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)
try:
function(path)
except PermissionError as ex:
if HIDE_WINDOWS_KNOWN_ERRORS:
from unittest import SkipTest
raise SkipTest(f"FIXME: fails with: PermissionError\n {ex}") from ex
raise
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: BinaryIO, destination: BinaryIO, 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 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.expandvars(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, 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:
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 (since `cache_lru()` unsupported on PY2).
_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:
# TODO: when py3.7 support is dropped, use the new interpolation f"{variable=}"
_logger.debug(f"sys.platform={sys.platform!r}, git_executable={git_executable!r}")
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, "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, expand_vars: bool = ...) -> None: ...
@overload
def expand_path(p: PathLike, expand_vars: bool = ...) -> str:
# TODO: Support for Python 3.5 has been dropped, so these overloads can be improved.
...
def expand_path(p: Union[None, PathLike], expand_vars: bool = True) -> Optional[PathLike]:
if isinstance(p, Path):
return p.resolve()
try:
p = osp.expanduser(p) # type: ignore[arg-type]
if expand_vars:
p = osp.expandvars(p)
return osp.normpath(osp.abspath(p))
except Exception:
return None
def remove_password_if_present(cmdline: Sequence[str]) -> List[str]:
"""Parse any command line argument and if one of the elements is an URL with a
username and/or password, replace them by stars (in-place).
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 index, to_parse in enumerate(cmdline):
new_cmdline.append(to_parse)
try:
url = urlsplit(to_parse)
# Remove password from the URL if present.
if url.password is None and url.username is None:
continue
if url.password is not None:
url = url._replace(netloc=url.netloc.replace(url.password, "*****"))
if url.username is not None:
url = url._replace(netloc=url.netloc.replace(url.username, "*****"))
new_cmdline[index] = urlunsplit(url)
except ValueError:
# This is not a valid URL.
continue
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 << x for x in range(_num_op_codes)]
STAGE_MASK = BEGIN | END
OP_MASK = ~STAGE_MASK
DONE_TOKEN = "done."
TOKEN_SEPARATOR = ", "
__slots__ = (
"_cur_line",
"_seen_ops",
"error_lines", # Lines that started with 'error:' or 'fatal:'.
"other_lines", # Lines not denoting progress (i.e.g. push-infos).
)
re_op_absolute = re.compile(r"(remote: )?([\w\s]+):\s+()(\d+)()(.*)")
re_op_relative = re.compile(r"(remote: )?([\w\s]+):\s+(\d+)% \((\d+)/(\d+)\)(.*)")
def __init__(self) -> 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) -> None:
"""Parse progress information from the given line as retrieved by
:manpage:`git-push(1)` or :manpage:`git-fetch(1)`.
- 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`.
"""
# 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 = line.decode("utf-8")
else:
line_str = line
self._cur_line = line_str
if self._cur_line.startswith(("error:", "fatal:")):
self.error_lines.append(self._cur_line)
return
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
# 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
# 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,
)
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:
return self._parse_progress_line(line.rstrip())
# 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 <_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) -> 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."""
# PRECOMPILED REGEX
name_only_regex = re.compile(r"<(.*)>")
name_email_regex = re.compile(r"(.*) <(.*?)>")
# 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 '