# vfs.py - Mercurial 'vfs' classes

#

#  Copyright Olivia Mackall <olivia@selenic.com>

#

# This software may be used and distributed according to the terms of the

# GNU General Public License version 2 or any later version.

from __future__ import annotations

import abc

import contextlib

import os

import shutil

import stat

import threading

import typing

from typing import (

    Any,

    BinaryIO,

    Callable,

    Dict,

    Iterable,

    Iterator,

    List,

    MutableMapping,

    Optional,

    Set,

    Tuple,

    Type,

    TypeVar,

    Union,

)

from .i18n import _

from . import (

    encoding,

    error,

    pathutil,

    pycompat,

    util,

)

if typing.TYPE_CHECKING:

    from . import (

        ui as uimod,

    )

    _Tbackgroundfilecloser = TypeVar(

        '_Tbackgroundfilecloser', bound='backgroundfilecloser'

    )

    _Tclosewrapbase = TypeVar('_Tclosewrapbase', bound='closewrapbase')

    _OnErrorFn = Callable[[Exception], Optional[object]]

def _avoidambig(path: bytes, oldstat: util.filestat) -> None:

    """Avoid file stat ambiguity forcibly

    This function causes copying ``path`` file, if it is owned by

    another (see issue5418 and issue5584 for detail).

    """

    def checkandavoid():

        newstat = util.filestat.frompath(path)

        # return whether file stat ambiguity is (already) avoided

        return not newstat.isambig(oldstat) or newstat.avoidambig(path, oldstat)

    if not checkandavoid():

        # simply copy to change owner of path to get privilege to

        # advance mtime (see issue5418)

        util.rename(util.mktempcopy(path), path)

        checkandavoid()

class abstractvfs(abc.ABC):

    """Abstract base class; cannot be instantiated"""

    # default directory separator for vfs

    #

    # Other vfs code always use `/` and this works fine because python file API

    # abstract the use of `/` and make it work transparently. For consistency

    # vfs will always use `/` when joining. This avoids some confusion in

    # encoded vfs (see issue6546)

    _dir_sep: bytes = b'/'

    # Used to disable the Rust `InnerRevlog` in case the VFS is not supported

    # by the Rust code

    rust_compatible = True

    # createmode is always available on subclasses

    createmode: int

    _chmod: bool

    # TODO: type return, which is util.posixfile wrapped by a proxy

    @abc.abstractmethod

    def __call__(self, path: bytes, mode: bytes = b'rb', **kwargs) -> Any:

        ...

    @abc.abstractmethod

    def _auditpath(self, path: bytes, mode: bytes) -> None:

        ...

    @abc.abstractmethod

    def join(self, path: Optional[bytes], *insidef: bytes) -> bytes:

        ...

    def tryread(self, path: bytes) -> bytes:

        '''gracefully return an empty string for missing files'''

        try:

            return self.read(path)

        except FileNotFoundError:

            pass

        return b""

    def tryreadlines(self, path: bytes, mode: bytes = b'rb') -> List[bytes]:

        '''gracefully return an empty array for missing files'''

        try:

            return self.readlines(path, mode=mode)

        except FileNotFoundError:

            pass

        return []

    @util.propertycache

    def open(self):

        """Open ``path`` file, which is relative to vfs root.

        Newly created directories are marked as "not to be indexed by

        the content indexing service", if ``notindexed`` is specified

        for "write" mode access.

        """

        return self.__call__

    def read(self, path: bytes) -> bytes:

        with self(path, b'rb') as fp:

            return fp.read()

    def readlines(self, path: bytes, mode: bytes = b'rb') -> List[bytes]:

        with self(path, mode=mode) as fp:

            return fp.readlines()

    def write(

        self, path: bytes, data: bytes, backgroundclose: bool = False, **kwargs

    ) -> int:

        with self(path, b'wb', backgroundclose=backgroundclose, **kwargs) as fp:

            return fp.write(data)

    def writelines(

        self,

        path: bytes,

        data: Iterable[bytes],

        mode: bytes = b'wb',

        notindexed: bool = False,

    ) -> None:

        with self(path, mode=mode, notindexed=notindexed) as fp:

            return fp.writelines(data)

    def append(self, path: bytes, data: bytes) -> int:

        with self(path, b'ab') as fp:

            return fp.write(data)

    def basename(self, path: bytes) -> bytes:

        """return base element of a path (as os.path.basename would do)

        This exists to allow handling of strange encoding if needed."""

        return os.path.basename(path)

    def chmod(self, path: bytes, mode: int) -> None:

        return os.chmod(self.join(path), mode)

    def dirname(self, path: bytes) -> bytes:

        """return dirname element of a path (as os.path.dirname would do)

        This exists to allow handling of strange encoding if needed."""

        return os.path.dirname(path)

    def exists(self, path: Optional[bytes] = None) -> bool:

        return os.path.exists(self.join(path))

    def fstat(self, fp: BinaryIO) -> os.stat_result:

        return util.fstat(fp)

    def isdir(self, path: Optional[bytes] = None) -> bool:

        return os.path.isdir(self.join(path))

    def isfile(self, path: Optional[bytes] = None) -> bool:

        return os.path.isfile(self.join(path))

    def islink(self, path: Optional[bytes] = None) -> bool:

        return os.path.islink(self.join(path))

    def isfileorlink(self, path: Optional[bytes] = None) -> bool:

        """return whether path is a regular file or a symlink

        Unlike isfile, this doesn't follow symlinks."""

        try:

            st = self.lstat(path)

        except OSError:

            return False

        mode = st.st_mode

        return stat.S_ISREG(mode) or stat.S_ISLNK(mode)

    def _join(self, *paths: bytes) -> bytes:

        root_idx = 0

        for idx, p in enumerate(paths):

            if os.path.isabs(p) or p.startswith(self._dir_sep):

                root_idx = idx

        if root_idx != 0:

            paths = paths[root_idx:]

        paths = [p for p in paths if p]

        return self._dir_sep.join(paths)

    def reljoin(self, *paths: bytes) -> bytes:

        """join various elements of a path together (as os.path.join would do)

        The vfs base is not injected so that path stay relative. This exists

        to allow handling of strange encoding if needed."""

        return self._join(*paths)

    def split(self, path: bytes) -> Tuple[bytes, bytes]:

        """split top-most element of a path (as os.path.split would do)

        This exists to allow handling of strange encoding if needed."""

        return os.path.split(path)

    def lexists(self, path: Optional[bytes] = None) -> bool:

        return os.path.lexists(self.join(path))

    def lstat(self, path: Optional[bytes] = None) -> os.stat_result:

        return os.lstat(self.join(path))

    def is_mmap_safe(self, path: Optional[bytes] = None) -> bool:

        """return True if it is safe to read a file content as mmap

        This focus on the file system aspect of such safety, the application

        logic around that file is not taken into account, so caller need to

        make sure the file won't be truncated in a way that will create SIGBUS

        on access.

        The initial motivation for this logic is that if mmap is used on NFS

        and somebody deletes the mapped file (e.g. by renaming on top of it),

        then you get SIGBUS, which can be pretty disruptive: we get core dump

        reports, and the process terminates without writing to the blackbox.

        Instead, in this situation we prefer to read the file normally.

        The risk of ESTALE in the middle of the read remains, but it's

        smaller because we read sooner and the error should be reported

        just as any other error.

        Note that python standard library does not offer the necessary function

        to detect the file stem bits. So this detection rely on compiled bits

        and is not available in pure python.

        """

        # XXX Since we already assume a vfs to address a consistent file system

        # in other location, we could determine the fstype once for the root

        # and cache that value.

        fstype = util.getfstype(self.join(path))

        return fstype is not None and fstype != b'nfs'

    def listdir(self, path: Optional[bytes] = None) -> List[bytes]:

        return os.listdir(self.join(path))

    def makedir(self, path: Optional[bytes] = None, notindexed=True) -> None:

        return util.makedir(self.join(path), notindexed)

    def makedirs(

        self, path: Optional[bytes] = None, mode: Optional[int] = None

    ) -> None:

        return util.makedirs(self.join(path), mode)

    def makelock(self, info: bytes, path: bytes) -> None:

        return util.makelock(info, self.join(path))

    def mkdir(self, path: Optional[bytes] = None) -> None:

        return os.mkdir(self.join(path))

    def mkstemp(

        self,

        suffix: bytes = b'',

        prefix: bytes = b'tmp',

        dir: Optional[bytes] = None,

    ) -> Tuple[int, bytes]:

        fd, name = pycompat.mkstemp(

            suffix=suffix, prefix=prefix, dir=self.join(dir)

        )

        dname, fname = util.split(name)

        if dir:

            return fd, os.path.join(dir, fname)

        else:

            return fd, fname

    # TODO: This doesn't match osutil.listdir().  stat=False in pure;

    #  non-optional bool in cext.  'skip' is bool if we trust cext, or bytes

    #  going by how pure uses it.  Also, cext returns a custom stat structure.

    #  from cext.osutil.pyi:

    #

    #     path: bytes, st: bool, skip: Optional[bool]

    def readdir(

        self, path: Optional[bytes] = None, stat=None, skip=None

    ) -> Any:

        return util.listdir(self.join(path), stat, skip)

    def readlock(self, path: bytes) -> bytes:

        return util.readlock(self.join(path))

    def rename(self, src: bytes, dst: bytes, checkambig: bool = False) -> None:

        """Rename from src to dst

        checkambig argument is used with util.filestat, and is useful

        only if destination file is guarded by any lock

        (e.g. repo.lock or repo.wlock).

        To avoid file stat ambiguity forcibly, checkambig=True involves

        copying ``src`` file, if it is owned by another. Therefore, use

        checkambig=True only in limited cases (see also issue5418 and

        issue5584 for detail).

        """

        self._auditpath(dst, b'w')

        srcpath = self.join(src)

        dstpath = self.join(dst)

        oldstat = util.filestat.frompath(dstpath) if checkambig else None

        util.rename(srcpath, dstpath)

        if oldstat and oldstat.stat:

            _avoidambig(dstpath, oldstat)

    def readlink(self, path: bytes) -> bytes:

        return util.readlink(self.join(path))

    def removedirs(self, path: Optional[bytes] = None) -> None:

        """Remove a leaf directory and all empty intermediate ones"""

        return util.removedirs(self.join(path))

    def rmdir(self, path: Optional[bytes] = None) -> None:

        """Remove an empty directory."""

        return os.rmdir(self.join(path))

    def rmtree(

        self,

        path: Optional[bytes] = None,

        ignore_errors: bool = False,

        forcibly: bool = False,

    ) -> None:

        """Remove a directory tree recursively

        If ``forcibly``, this tries to remove READ-ONLY files, too.

        """

        if forcibly:

            def onexc(function: Callable, path: str, excinfo: Exception):

                # Note: str is passed here even if bytes are passed to rmtree

                # on platforms where `shutil._use_fd_functions == True`.  It is

                # bytes otherwise.  Fortunately, the methods used here accept

                # both.

                if function is not os.remove:

                    raise

                # read-only files cannot be unlinked under Windows

                s = os.stat(path)

                if (s.st_mode & stat.S_IWRITE) != 0:

                    raise

                os.chmod(path, stat.S_IMODE(s.st_mode) | stat.S_IWRITE)

                os.remove(path)

        else:

            def onexc(*args):

                pass

        try:

            # pytype: disable=wrong-keyword-args

            return shutil.rmtree(

                self.join(path), ignore_errors=ignore_errors, onexc=onexc

            )

            # pytype: enable=wrong-keyword-args

        except TypeError:  # onexc was introduced in Python 3.12

            return shutil.rmtree(

                self.join(path), ignore_errors=ignore_errors, onerror=onexc

            )

    def setflags(self, path: bytes, l: bool, x: bool) -> None:

        return util.setflags(self.join(path), l, x)

    def stat(self, path: Optional[bytes] = None) -> os.stat_result:

        return os.stat(self.join(path))

    def unlink(self, path: Optional[bytes] = None) -> None:

        return util.unlink(self.join(path))

    def tryunlink(self, path: Optional[bytes] = None) -> bool:

        """Attempt to remove a file, ignoring missing file errors."""

        return util.tryunlink(self.join(path))

    def unlinkpath(

        self,

        path: Optional[bytes] = None,

        ignoremissing: bool = False,

        rmdir: bool = True,

    ) -> None:

        return util.unlinkpath(

            self.join(path), ignoremissing=ignoremissing, rmdir=rmdir

        )

    # TODO: could be Tuple[float, float] too.

    def utime(

        self, path: Optional[bytes] = None, t: Optional[Tuple[int, int]] = None

    ) -> None:

        return os.utime(self.join(path), t)

    def walk(

        self, path: Optional[bytes] = None, onerror: Optional[_OnErrorFn] = None

    ) -> Iterator[Tuple[bytes, List[bytes], List[bytes]]]:

        """Yield (dirpath, dirs, files) tuple for each directory under path

        ``dirpath`` is relative one from the root of this vfs. This

        uses ``os.sep`` as path separator, even you specify POSIX

        style ``path``.

        "The root of this vfs" is represented as empty ``dirpath``.

        """

        root = os.path.normpath(self.join(None))

        # when dirpath == root, dirpath[prefixlen:] becomes empty

        # because len(dirpath) < prefixlen.

        prefixlen = len(pathutil.normasprefix(root))

        for dirpath, dirs, files in os.walk(self.join(path), onerror=onerror):

            yield (dirpath[prefixlen:], dirs, files)

    @contextlib.contextmanager

    def backgroundclosing(

        self, ui: uimod.ui, expectedcount: int = -1

    ) -> Iterator[Optional[backgroundfilecloser]]:

        """Allow files to be closed asynchronously.

        When this context manager is active, ``backgroundclose`` can be passed

        to ``__call__``/``open`` to result in the file possibly being closed

        asynchronously, on a background thread.

        """

        # Sharing backgroundfilecloser between threads is complex and using

        # multiple instances puts us at risk of running out of file descriptors

        # only allow to use backgroundfilecloser when in main thread.

        if threading.current_thread() is not threading.main_thread():

            yield

            return

        vfs = getattr(self, 'vfs', self)

        if getattr(vfs, '_backgroundfilecloser', None):

            raise error.Abort(

                _(b'can only have 1 active background file closer')

            )

        with backgroundfilecloser(ui, expectedcount=expectedcount) as bfc:

            try:

                vfs._backgroundfilecloser = (

                    bfc  # pytype: disable=attribute-error

                )

                yield bfc

            finally:

                vfs._backgroundfilecloser = (

                    None  # pytype: disable=attribute-error

                )

    def register_file(self, path: bytes) -> None:

        """generic hook point to lets fncache steer its stew"""

    def prepare_streamed_file(

        self, path: bytes, known_directories: Set[bytes]

    ) -> Tuple[bytes, Optional[int]]:

        """make sure we are ready to write a file from a stream clone

        The "known_directories" variable is here to avoid trying to create the

        same directories over and over during a stream clone. It will be

        updated by this function.

        return (path, mode)::

            <path> is the real file system path content should be written to,

            <mode> is the file mode that need to be set if any.

        """

        self.register_file(path)