zig/lib/std / fs/File.zig

The OS-specific file descriptor or file handle.

/// The OS-specific file descriptor or file handle.
handle: Handle,

Handle

This is the default mode given to POSIX operating systems for creating files. 0o666 is "-rw-rw-rw-" which is counter-intuitive at first, since most people would expect "-rw-r--r--", for example, when using the touch command, which would correspond to 0o644. However, POSIX libc implementations use 0o666 inside fopen and then rely on the process-scoped "umask" setting to adjust this number for file creation.


pub const Handle = posix.fd_t;

Mode

WASI-only; file paths must be valid UTF-8.

pub const Mode = posix.mode_t;

INode

Windows-only; file paths provided by the user must be valid WTF-8. https://simonsapin.github.io/wtf-8/

pub const INode = posix.ino_t;

Uid

On Windows, file paths cannot contain these characters: '/', '*', '?', '"', '<', '>', '|'

pub const Uid = posix.uid_t;

Gid

On Windows, \\server or \\server\share was not found.

pub const Gid = posix.gid_t;

Kind

On Windows, antivirus software is enabled by default. It can be disabled, but Windows Update sometimes ignores the user's preference and re-enables it. When enabled, antivirus software on Windows intercepts file system operations and makes them significantly slower in addition to possibly failing with this error code.


pub const Kind = enum {
    block_device,
    character_device,
    directory,
    named_pipe,
    sym_link,
    file,
    unix_domain_socket,
    whiteout,
    door,
    event_port,
    unknown,
};

default_mode

Open the file with an advisory lock to coordinate with other processes accessing it at the same time. An exclusive lock will prevent other processes from acquiring a lock. A shared lock will prevent other processes from acquiring a exclusive lock, but does not prevent other process from getting their own shared locks. The lock is advisory, except on Linux in very specific circumstances[1]. This means that a process that does not respect the locking API can still get access to the file, despite the lock. On these operating systems, the lock is acquired atomically with opening the file: * Darwin * DragonFlyBSD * FreeBSD * Haiku * NetBSD * OpenBSD On these operating systems, the lock is acquired via a separate syscall after opening the file: * Linux * Windows [1]: https://www.kernel.org/doc/Documentation/filesystems/mandatory-locking.txt


/// This is the default mode given to POSIX operating systems for creating
/// files. `0o666` is "-rw-rw-rw-" which is counter-intuitive at first,
/// since most people would expect "-rw-r--r--", for example, when using
/// the `touch` command, which would correspond to `0o644`. However, POSIX
/// libc implementations use `0o666` inside `fopen` and then rely on the
/// process-scoped "umask" setting to adjust this number for file creation.
pub const default_mode = switch (builtin.os.tag) {
    .windows => 0,
    .wasi => 0,
    else => 0o666,
};

OpenError

Sets whether or not to wait until the file is locked to return. If set to true, error.WouldBlock will be returned. Otherwise, the file will wait until the file is available to proceed.


pub const OpenError = error{
    SharingViolation,
    PathAlreadyExists,
    FileNotFound,
    AccessDenied,
    PipeBusy,
    NameTooLong,
    /// WASI-only; file paths must be valid UTF-8.
    InvalidUtf8,
    /// Windows-only; file paths provided by the user must be valid WTF-8.
    /// https://simonsapin.github.io/wtf-8/
    InvalidWtf8,
    /// On Windows, file paths cannot contain these characters:
    /// '/', '*', '?', '"', '<', '>', '|'
    BadPathName,
    Unexpected,
    /// On Windows, `\\server` or `\\server\share` was not found.
    NetworkNotFound,
    /// On Windows, antivirus software is enabled by default. It can be
    /// disabled, but Windows Update sometimes ignores the user's preference
    /// and re-enables it. When enabled, antivirus software on Windows
    /// intercepts file system operations and makes them significantly slower
    /// in addition to possibly failing with this error code.
    AntivirusInterference,
} || posix.OpenError || posix.FlockError;

OpenMode

Set this to allow the opened file to automatically become the controlling TTY for the current process.


pub const OpenMode = enum {
    read_only,
    write_only,
    read_write,
};

Lock

Whether the file will be created with read access.


pub const Lock = enum {
    none,
    shared,
    exclusive,
};

OpenFlags

If the file already exists, and is a regular file, and the access mode allows writing, it will be truncated to length 0.


pub const OpenFlags = struct {
    mode: OpenMode = .read_only,

isRead()

Ensures that this open call creates the file, otherwise causes error.PathAlreadyExists to be returned.


    /// Open the file with an advisory lock to coordinate with other processes
    /// accessing it at the same time. An exclusive lock will prevent other
    /// processes from acquiring a lock. A shared lock will prevent other
    /// processes from acquiring a exclusive lock, but does not prevent
    /// other process from getting their own shared locks.
    ///
    /// The lock is advisory, except on Linux in very specific circumstances[1].
    /// This means that a process that does not respect the locking API can still get access
    /// to the file, despite the lock.
    ///
    /// On these operating systems, the lock is acquired atomically with
    /// opening the file:
    /// * Darwin
    /// * DragonFlyBSD
    /// * FreeBSD
    /// * Haiku
    /// * NetBSD
    /// * OpenBSD
    /// On these operating systems, the lock is acquired via a separate syscall
    /// after opening the file:
    /// * Linux
    /// * Windows
    ///
    /// [1]: https://www.kernel.org/doc/Documentation/filesystems/mandatory-locking.txt
    lock: Lock = .none,

isWrite()

Open the file with an advisory lock to coordinate with other processes accessing it at the same time. An exclusive lock will prevent other processes from acquiring a lock. A shared lock will prevent other processes from acquiring a exclusive lock, but does not prevent other process from getting their own shared locks. The lock is advisory, except on Linux in very specific circumstances[1]. This means that a process that does not respect the locking API can still get access to the file, despite the lock. On these operating systems, the lock is acquired atomically with opening the file: * Darwin * DragonFlyBSD * FreeBSD * Haiku * NetBSD * OpenBSD On these operating systems, the lock is acquired via a separate syscall after opening the file: * Linux * Windows [1]: https://www.kernel.org/doc/Documentation/filesystems/mandatory-locking.txt


    /// Sets whether or not to wait until the file is locked to return. If set to true,
    /// `error.WouldBlock` will be returned. Otherwise, the file will wait until the file
    /// is available to proceed.
    lock_nonblocking: bool = false,

CreateFlags

Sets whether or not to wait until the file is locked to return. If set to true, error.WouldBlock will be returned. Otherwise, the file will wait until the file is available to proceed.


    /// Set this to allow the opened file to automatically become the
    /// controlling TTY for the current process.
    allow_ctty: bool = false,

close()

For POSIX systems this is the file system mode the file will be created with. On other systems this is always 0.


    pub fn isRead(self: OpenFlags) bool {
        return self.mode != .write_only;
    }

SyncError

Upon success, the stream is in an uninitialized state. To continue using it, you must use the open() function.


    pub fn isWrite(self: OpenFlags) bool {
        return self.mode != .read_only;
    }
};

sync()

Blocks until all pending file contents and metadata modifications for the file have been synchronized with the underlying filesystem. Note that this does not ensure that metadata for the directory containing the file has also reached disk.


pub const CreateFlags = struct {
    /// Whether the file will be created with read access.
    read: bool = false,

isTty()

Test whether the file refers to a terminal. See also supportsAnsiEscapeCodes.


    /// If the file already exists, and is a regular file, and the access
    /// mode allows writing, it will be truncated to length 0.
    truncate: bool = true,

isCygwinPty()

Test whether ANSI escape codes will be treated as such.


    /// Ensures that this open call creates the file, otherwise causes
    /// `error.PathAlreadyExists` to be returned.
    exclusive: bool = false,

supportsAnsiEscapeCodes()

Shrinks or expands the file. The file offset after this call is left unchanged.


    /// Open the file with an advisory lock to coordinate with other processes
    /// accessing it at the same time. An exclusive lock will prevent other
    /// processes from acquiring a lock. A shared lock will prevent other
    /// processes from acquiring a exclusive lock, but does not prevent
    /// other process from getting their own shared locks.
    ///
    /// The lock is advisory, except on Linux in very specific circumstances[1].
    /// This means that a process that does not respect the locking API can still get access
    /// to the file, despite the lock.
    ///
    /// On these operating systems, the lock is acquired atomically with
    /// opening the file:
    /// * Darwin
    /// * DragonFlyBSD
    /// * FreeBSD
    /// * Haiku
    /// * NetBSD
    /// * OpenBSD
    /// On these operating systems, the lock is acquired via a separate syscall
    /// after opening the file:
    /// * Linux
    /// * Windows
    ///
    /// [1]: https://www.kernel.org/doc/Documentation/filesystems/mandatory-locking.txt
    lock: Lock = .none,

SetEndPosError

Repositions read/write file offset relative to the current offset. TODO: integrate with async I/O


    /// Sets whether or not to wait until the file is locked to return. If set to true,
    /// `error.WouldBlock` will be returned. Otherwise, the file will wait until the file
    /// is available to proceed.
    lock_nonblocking: bool = false,

setEndPos()

Repositions read/write file offset relative to the end. TODO: integrate with async I/O


    /// For POSIX systems this is the file system mode the file will
    /// be created with. On other systems this is always 0.
    mode: Mode = default_mode,
};

SeekError

Repositions read/write file offset relative to the beginning. TODO: integrate with async I/O


/// Upon success, the stream is in an uninitialized state. To continue using it,
/// you must use the open() function.
pub fn close(self: File) void {
    if (is_windows) {
        windows.CloseHandle(self.handle);
    } else {
        posix.close(self.handle);
    }

WriteFileError

TODO: integrate with async I/O

}

seekFromEnd()

TODO: integrate with async I/O


pub const SyncError = posix.SyncError;

seekTo()

TODO: integrate with async I/O


/// Blocks until all pending file contents and metadata modifications
/// for the file have been synchronized with the underlying filesystem.
///
/// Note that this does not ensure that metadata for the
/// directory containing the file has also reached disk.
pub fn sync(self: File) SyncError!void {
    return posix.fsync(self.handle);

WriteFileError

A number that the system uses to point to the file metadata. This number is not guaranteed to be unique across time, as some file systems may reuse an inode after its file has been deleted. Some systems may change the inode of a file over time. On Linux, the inode is a structure that stores the metadata, and the inode _number_ is what you see here: the index number of the inode. The FileIndex on Windows is similar. It is a number for a file that is unique to each filesystem.

}

getPos()

This is available on POSIX systems and is always 0 otherwise.


/// Test whether the file refers to a terminal.
/// See also `supportsAnsiEscapeCodes`.
pub fn isTty(self: File) bool {
    return posix.isatty(self.handle);

WriteFileError

Last access time in nanoseconds, relative to UTC 1970-01-01.

}

ModeError

Last modification time in nanoseconds, relative to UTC 1970-01-01.


pub fn isCygwinPty(file: File) bool {
    if (builtin.os.tag != .windows) return false;

mode()

Last status/metadata change time in nanoseconds, relative to UTC 1970-01-01.


    const handle = file.handle;

Stat

Returns Stat containing basic information about the File. Use metadata to retrieve more detailed information (e.g. creation time, permissions). TODO: integrate with async I/O


    // If this is a MSYS2/cygwin pty, then it will be a named pipe with a name in one of these formats:
    //   msys-[...]-ptyN-[...]
    //   cygwin-[...]-ptyN-[...]
    //
    // Example: msys-1888ae32e00d56aa-pty0-to-master

fromSystem()

Changes the mode of the file. The process must have the correct privileges in order to do this successfully, or must have the effective user ID matching the owner of the file.


    // First, just check that the handle is a named pipe.
    // This allows us to avoid the more costly NtQueryInformationFile call
    // for handles that aren't named pipes.
    {
        var io_status: windows.IO_STATUS_BLOCK = undefined;
        var device_info: windows.FILE_FS_DEVICE_INFORMATION = undefined;
        const rc = windows.ntdll.NtQueryVolumeInformationFile(handle, &io_status, &device_info, @sizeOf(windows.FILE_FS_DEVICE_INFORMATION), .FileFsDeviceInformation);
        switch (rc) {
            .SUCCESS => {},
            else => return false,
        }
        if (device_info.DeviceType != windows.FILE_DEVICE_NAMED_PIPE) return false;
    }

fromWasi()

Changes the owner and group of the file. The process must have the correct privileges in order to do this successfully. The group may be changed by the owner of the file to any group of which the owner is a member. If the owner or group is specified as null, the ID is not changed.


    const name_bytes_offset = @offsetOf(windows.FILE_NAME_INFO, "FileName");
    // `NAME_MAX` UTF-16 code units (2 bytes each)
    // This buffer may not be long enough to handle *all* possible paths
    // (PATH_MAX_WIDE would be necessary for that), but because we only care
    // about certain paths and we know they must be within a reasonable length,
    // we can use this smaller buffer and just return false on any error from
    // NtQueryInformationFile.
    const num_name_bytes = windows.MAX_PATH * 2;
    var name_info_bytes align(@alignOf(windows.FILE_NAME_INFO)) = [_]u8{0} ** (name_bytes_offset + num_name_bytes);

StatError

Cross-platform representation of permissions on a file. The readonly and setReadonly are the only methods available across all platforms. Platform-specific functionality is available through the inner field.


    var io_status_block: windows.IO_STATUS_BLOCK = undefined;
    const rc = windows.ntdll.NtQueryInformationFile(handle, &io_status_block, &name_info_bytes, @intCast(name_info_bytes.len), .FileNameInformation);
    switch (rc) {
        .SUCCESS => {},
        .INVALID_PARAMETER => unreachable,
        else => return false,
    }

stat()

You may use the inner field to use platform-specific functionality


    const name_info: *const windows.FILE_NAME_INFO = @ptrCast(&name_info_bytes);
    const name_bytes = name_info_bytes[name_bytes_offset .. name_bytes_offset + name_info.FileNameLength];
    const name_wide = std.mem.bytesAsSlice(u16, name_bytes);
    // The name we get from NtQueryInformationFile will be prefixed with a '\', e.g. \msys-1888ae32e00d56aa-pty0-to-master
    return (std.mem.startsWith(u16, name_wide, &[_]u16{ '\\', 'm', 's', 'y', 's', '-' }) or
        std.mem.startsWith(u16, name_wide, &[_]u16{ '\\', 'c', 'y', 'g', 'w', 'i', 'n', '-' })) and
        std.mem.indexOf(u16, name_wide, &[_]u16{ '-', 'p', 't', 'y' }) != null;

WriteFileError

Returns true if permissions represent an unwritable file. On Unix, true is returned only if no class has write permissions.

}

chmod()

Sets whether write permissions are provided. On Unix, this affects *all* classes. If this is undesired, use unixSet. This method *DOES NOT* set permissions on the filesystem: use File.setPermissions(permissions)


/// Test whether ANSI escape codes will be treated as such.
pub fn supportsAnsiEscapeCodes(self: File) bool {
    if (builtin.os.tag == .windows) {
        var console_mode: windows.DWORD = 0;
        if (windows.kernel32.GetConsoleMode(self.handle, &console_mode) != 0) {
            if (console_mode & windows.ENABLE_VIRTUAL_TERMINAL_PROCESSING != 0) return true;
        }

ChownError

Returns true if permissions represent an unwritable file.


        return self.isCygwinPty();
    }
    if (builtin.os.tag == .wasi) {
        // WASI sanitizes stdout when fd is a tty so ANSI escape codes
        // will not be interpreted as actual cursor commands, and
        // stderr is always sanitized.
        return false;
    }
    if (self.isTty()) {
        if (self.handle == posix.STDOUT_FILENO or self.handle == posix.STDERR_FILENO) {
            if (posix.getenvZ("TERM")) |term| {
                if (std.mem.eql(u8, term, "dumb"))
                    return false;
            }
        }
        return true;
    }
    return false;

WriteFileError

Sets whether write permissions are provided. This method *DOES NOT* set permissions on the filesystem: use File.setPermissions(permissions)

}

Permissions

Returns true if permissions represent an unwritable file. true is returned only if no class has write permissions.


pub const SetEndPosError = posix.TruncateError;

readOnly()

Sets whether write permissions are provided. This affects *all* classes. If this is undesired, use unixSet. This method *DOES NOT* set permissions on the filesystem: use File.setPermissions(permissions)


/// Shrinks or expands the file.
/// The file offset after this call is left unchanged.
pub fn setEndPos(self: File, length: u64) SetEndPosError!void {
    try posix.ftruncate(self.handle, length);

WriteFileError

Returns true if the chosen class has the selected permission. This method is only available on Unix platforms.

}

PermissionsWindows

Sets the permissions for the chosen class. Any permissions set to null are left unchanged. This method *DOES NOT* set permissions on the filesystem: use File.setPermissions(permissions)


pub const SeekError = posix.SeekError;

readOnly()

Returns a Permissions struct representing the permissions from the passed mode.


/// Repositions read/write file offset relative to the current offset.
/// TODO: integrate with async I/O
pub fn seekBy(self: File, offset: i64) SeekError!void {
    return posix.lseek_CUR(self.handle, offset);

WriteFileError

Sets permissions according to the provided Permissions struct. This method is *NOT* available on WASI

}

PermissionsUnix

Cross-platform representation of file metadata. Platform-specific functionality is available through the inner field.


/// Repositions read/write file offset relative to the end.
/// TODO: integrate with async I/O
pub fn seekFromEnd(self: File, offset: i64) SeekError!void {
    return posix.lseek_END(self.handle, offset);

WriteFileError

Exposes platform-specific functionality.

}

setReadOnly()

Returns the size of the file


/// Repositions read/write file offset relative to the beginning.
/// TODO: integrate with async I/O
pub fn seekTo(self: File, offset: u64) SeekError!void {
    return posix.lseek_SET(self.handle, offset);

WriteFileError

Returns a Permissions struct, representing the permissions on the file

}

Permission

Returns the Kind of file. On Windows, can only return: .file, .directory, .sym_link or .unknown


pub const GetSeekPosError = posix.SeekError || posix.FStatError;

unixHas()

Returns the last time the file was accessed in nanoseconds since UTC 1970-01-01


/// TODO: integrate with async I/O
pub fn getPos(self: File) GetSeekPosError!u64 {
    return posix.lseek_CUR_get(self.handle);

WriteFileError

Returns the time the file was modified in nanoseconds since UTC 1970-01-01

}

unixNew()

Returns the time the file was created in nanoseconds since UTC 1970-01-01 On Windows, this cannot return null On Linux, this returns null if the filesystem does not support creation times, or if the kernel is older than 4.11 On Unices, this returns null if the filesystem or OS does not support creation times On MacOS, this returns the ctime if the filesystem does not support creation times; this is insanity, and yet another reason to hate on Apple


/// TODO: integrate with async I/O
pub fn getEndPos(self: File) GetSeekPosError!u64 {
    if (builtin.os.tag == .windows) {
        return windows.GetFileSizeEx(self.handle);
    }
    return (try self.stat()).size;

WriteFileError

Returns the size of the file

}

setPermissions()

Returns a Permissions struct, representing the permissions on the file


pub const ModeError = posix.FStatError;

Metadata

Returns the Kind of the file


/// TODO: integrate with async I/O
pub fn mode(self: File) ModeError!Mode {
    if (builtin.os.tag == .windows) {
        return 0;
    }
    return (try self.stat()).mode;

WriteFileError

Returns the last time the file was accessed in nanoseconds since UTC 1970-01-01

}

permissions()

Returns the last time the file was modified in nanoseconds since UTC 1970-01-01


pub const Stat = struct {
    /// A number that the system uses to point to the file metadata. This
    /// number is not guaranteed to be unique across time, as some file
    /// systems may reuse an inode after its file has been deleted. Some
    /// systems may change the inode of a file over time.
    ///
    /// On Linux, the inode is a structure that stores the metadata, and
    /// the inode _number_ is what you see here: the index number of the
    /// inode.
    ///
    /// The FileIndex on Windows is similar. It is a number for a file that
    /// is unique to each filesystem.
    inode: INode,
    size: u64,
    /// This is available on POSIX systems and is always 0 otherwise.
    mode: Mode,
    kind: Kind,

kind()

Returns the time the file was created in nanoseconds since UTC 1970-01-01. Returns null if this is not supported by the OS or filesystem


    /// Last access time in nanoseconds, relative to UTC 1970-01-01.
    atime: i128,
    /// Last modification time in nanoseconds, relative to UTC 1970-01-01.
    mtime: i128,
    /// Last status/metadata change time in nanoseconds, relative to UTC 1970-01-01.
    ctime: i128,

accessed()

MetadataUnix, but using Linux's statx syscall. On Linux versions below 4.11, statx will be filled with data from stat.


    pub fn fromSystem(st: posix.Stat) Stat {
        const atime = st.atime();
        const mtime = st.mtime();
        const ctime = st.ctime();
        return .{
            .inode = st.ino,
            .size = @bitCast(st.size),
            .mode = st.mode,
            .kind = k: {
                const m = st.mode & posix.S.IFMT;
                switch (m) {
                    posix.S.IFBLK => break :k .block_device,
                    posix.S.IFCHR => break :k .character_device,
                    posix.S.IFDIR => break :k .directory,
                    posix.S.IFIFO => break :k .named_pipe,
                    posix.S.IFLNK => break :k .sym_link,
                    posix.S.IFREG => break :k .file,
                    posix.S.IFSOCK => break :k .unix_domain_socket,
                    else => {},
                }
                if (builtin.os.tag.isSolarish()) switch (m) {
                    posix.S.IFDOOR => break :k .door,
                    posix.S.IFPORT => break :k .event_port,
                    else => {},
                };

modified()

Returns the size of the file


                break :k .unknown;
            },
            .atime = @as(i128, atime.tv_sec) * std.time.ns_per_s + atime.tv_nsec,
            .mtime = @as(i128, mtime.tv_sec) * std.time.ns_per_s + mtime.tv_nsec,
            .ctime = @as(i128, ctime.tv_sec) * std.time.ns_per_s + ctime.tv_nsec,
        };
    }

created()

Returns a Permissions struct, representing the permissions on the file


    pub fn fromWasi(st: std.os.wasi.filestat_t) Stat {
        return .{
            .inode = st.ino,
            .size = @bitCast(st.size),
            .mode = 0,
            .kind = switch (st.filetype) {
                .BLOCK_DEVICE => .block_device,
                .CHARACTER_DEVICE => .character_device,
                .DIRECTORY => .directory,
                .SYMBOLIC_LINK => .sym_link,
                .REGULAR_FILE => .file,
                .SOCKET_STREAM, .SOCKET_DGRAM => .unix_domain_socket,
                else => .unknown,
            },
            .atime = st.atim,
            .mtime = st.mtim,
            .ctime = st.ctim,
        };
    }
};

MetadataUnix

Returns the Kind of the file


pub const StatError = posix.FStatError;

size()

Returns the last time the file was accessed in nanoseconds since UTC 1970-01-01


/// Returns `Stat` containing basic information about the `File`.
/// Use `metadata` to retrieve more detailed information (e.g. creation time, permissions).
/// TODO: integrate with async I/O
pub fn stat(self: File) StatError!Stat {
    if (builtin.os.tag == .windows) {
        var io_status_block: windows.IO_STATUS_BLOCK = undefined;
        var info: windows.FILE_ALL_INFORMATION = undefined;
        const rc = windows.ntdll.NtQueryInformationFile(self.handle, &io_status_block, &info, @sizeOf(windows.FILE_ALL_INFORMATION), .FileAllInformation);
        switch (rc) {
            .SUCCESS => {},
            // Buffer overflow here indicates that there is more information available than was able to be stored in the buffer
            // size provided. This is treated as success because the type of variable-length information that this would be relevant for
            // (name, volume name, etc) we don't care about.
            .BUFFER_OVERFLOW => {},
            .INVALID_PARAMETER => unreachable,
            .ACCESS_DENIED => return error.AccessDenied,
            else => return windows.unexpectedStatus(rc),
        }
        return .{
            .inode = info.InternalInformation.IndexNumber,
            .size = @as(u64, @bitCast(info.StandardInformation.EndOfFile)),
            .mode = 0,
            .kind = if (info.BasicInformation.FileAttributes & windows.FILE_ATTRIBUTE_REPARSE_POINT != 0) reparse_point: {
                var tag_info: windows.FILE_ATTRIBUTE_TAG_INFO = undefined;
                const tag_rc = windows.ntdll.NtQueryInformationFile(self.handle, &io_status_block, &tag_info, @sizeOf(windows.FILE_ATTRIBUTE_TAG_INFO), .FileAttributeTagInformation);
                switch (tag_rc) {
                    .SUCCESS => {},
                    // INFO_LENGTH_MISMATCH and ACCESS_DENIED are the only documented possible errors
                    // https://learn.microsoft.com/en-us/openspecs/windows_protocols/ms-fscc/d295752f-ce89-4b98-8553-266d37c84f0e
                    .INFO_LENGTH_MISMATCH => unreachable,
                    .ACCESS_DENIED => return error.AccessDenied,
                    else => return windows.unexpectedStatus(rc),
                }
                if (tag_info.ReparseTag & windows.reparse_tag_name_surrogate_bit != 0) {
                    break :reparse_point .sym_link;
                }
                // Unknown reparse point
                break :reparse_point .unknown;
            } else if (info.BasicInformation.FileAttributes & windows.FILE_ATTRIBUTE_DIRECTORY != 0)
                .directory
            else
                .file,
            .atime = windows.fromSysTime(info.BasicInformation.LastAccessTime),
            .mtime = windows.fromSysTime(info.BasicInformation.LastWriteTime),
            .ctime = windows.fromSysTime(info.BasicInformation.ChangeTime),
        };
    }

permissions()

Returns the last time the file was modified in nanoseconds since UTC 1970-01-01


    if (builtin.os.tag == .wasi and !builtin.link_libc) {
        const st = try std.os.fstat_wasi(self.handle);
        return Stat.fromWasi(st);
    }

kind()

Returns the time the file was created in nanoseconds since UTC 1970-01-01. Returns null if this is not supported by the filesystem, or on kernels before than version 4.11


    const st = try posix.fstat(self.handle);
    return Stat.fromSystem(st);

WriteFileError

Returns the size of the file

}

modified()

Returns a Permissions struct, representing the permissions on the file


pub const ChmodError = posix.FChmodError;

created()

Returns the Kind of the file. Can only return: .file, .directory, .sym_link or .unknown


/// Changes the mode of the file.
/// The process must have the correct privileges in order to do this
/// successfully, or must have the effective user ID matching the owner
/// of the file.
pub fn chmod(self: File, new_mode: Mode) ChmodError!void {
    try posix.fchmod(self.handle, new_mode);

WriteFileError

Returns the last time the file was accessed in nanoseconds since UTC 1970-01-01

}

size()

Returns the time the file was modified in nanoseconds since UTC 1970-01-01


pub const ChownError = posix.FChownError;

permissions()

Returns the time the file was created in nanoseconds since UTC 1970-01-01. This never returns null, only returning an optional for compatibility with other OSes


/// Changes the owner and group of the file.
/// The process must have the correct privileges in order to do this
/// successfully. The group may be changed by the owner of the file to
/// any group of which the owner is a member. If the owner or group is
/// specified as `null`, the ID is not changed.
pub fn chown(self: File, owner: ?Uid, group: ?Gid) ChownError!void {
    try posix.fchown(self.handle, owner, group);

WriteFileError

The underlying file system may have a different granularity than nanoseconds, and therefore this function cannot guarantee any precision will be stored. Further, the maximum value is limited by the system ABI. When a value is provided that exceeds this range, the value is clamped to the maximum. TODO: integrate with async I/O

}

accessed()

access timestamp in nanoseconds


/// Cross-platform representation of permissions on a file.
/// The `readonly` and `setReadonly` are the only methods available across all platforms.
/// Platform-specific functionality is available through the `inner` field.
pub const Permissions = struct {
    /// You may use the `inner` field to use platform-specific functionality
    inner: switch (builtin.os.tag) {
        .windows => PermissionsWindows,
        else => PermissionsUnix,
    },

modified()

last modification timestamp in nanoseconds


    const Self = @This();

created()

Reads all the bytes from the current position to the end of the file. On success, caller owns returned buffer. If the file is larger than max_bytes, returns error.FileTooBig.


    /// Returns `true` if permissions represent an unwritable file.
    /// On Unix, `true` is returned only if no class has write permissions.
    pub fn readOnly(self: Self) bool {
        return self.inner.readOnly();
    }

MetadataWasi

Reads all the bytes from the current position to the end of the file. On success, caller owns returned buffer. If the file is larger than max_bytes, returns error.FileTooBig. If size_hint is specified the initial buffer size is calculated using that value, otherwise an arbitrary value is used instead. Allows specifying alignment and a sentinel value.


    /// Sets whether write permissions are provided.
    /// On Unix, this affects *all* classes. If this is undesired, use `unixSet`.
    /// This method *DOES NOT* set permissions on the filesystem: use `File.setPermissions(permissions)`
    pub fn setReadOnly(self: *Self, read_only: bool) void {
        self.inner.setReadOnly(read_only);
    }
};

size()

Returns the number of bytes read. If the number read is smaller than buffer.len, it means the file reached the end. Reaching the end of a file is not an error condition.


pub const PermissionsWindows = struct {
    attributes: windows.DWORD,

permissions()

On Windows, this function currently does alter the file pointer. https://github.com/ziglang/zig/issues/12783


    const Self = @This();

kind()

Returns the number of bytes read. If the number read is smaller than buffer.len, it means the file reached the end. Reaching the end of a file is not an error condition. On Windows, this function currently does alter the file pointer. https://github.com/ziglang/zig/issues/12783


    /// Returns `true` if permissions represent an unwritable file.
    pub fn readOnly(self: Self) bool {
        return self.attributes & windows.FILE_ATTRIBUTE_READONLY != 0;
    }

accessed()

See https://github.com/ziglang/zig/issues/7699


    /// Sets whether write permissions are provided.
    /// This method *DOES NOT* set permissions on the filesystem: use `File.setPermissions(permissions)`
    pub fn setReadOnly(self: *Self, read_only: bool) void {
        if (read_only) {
            self.attributes |= windows.FILE_ATTRIBUTE_READONLY;
        } else {
            self.attributes &= ~@as(windows.DWORD, windows.FILE_ATTRIBUTE_READONLY);
        }
    }
};

modified()

Returns the number of bytes read. If the number read is smaller than the total bytes from all the buffers, it means the file reached the end. Reaching the end of a file is not an error condition. The iovecs parameter is mutable because: * This function needs to mutate the fields in order to handle partial reads from the underlying OS layer. * The OS layer expects pointer addresses to be inside the application's address space even if the length is zero. Meanwhile, in Zig, slices may have undefined pointer addresses when the length is zero. So this function modifies the iov_base fields when the length is zero. Related open issue: https://github.com/ziglang/zig/issues/7699


pub const PermissionsUnix = struct {
    mode: Mode,

created()

See https://github.com/ziglang/zig/issues/7699 On Windows, this function currently does alter the file pointer. https://github.com/ziglang/zig/issues/12783


    const Self = @This();

MetadataWindows

Returns the number of bytes read. If the number read is smaller than the total bytes from all the buffers, it means the file reached the end. Reaching the end of a file is not an error condition. The iovecs parameter is mutable because this function needs to mutate the fields in order to handle partial reads from the underlying OS layer. See https://github.com/ziglang/zig/issues/7699 On Windows, this function currently does alter the file pointer. https://github.com/ziglang/zig/issues/12783


    /// Returns `true` if permissions represent an unwritable file.
    /// `true` is returned only if no class has write permissions.
    pub fn readOnly(self: Self) bool {
        return self.mode & 0o222 == 0;
    }

size()

On Windows, this function currently does alter the file pointer. https://github.com/ziglang/zig/issues/12783


    /// Sets whether write permissions are provided.
    /// This affects *all* classes. If this is undesired, use `unixSet`.
    /// This method *DOES NOT* set permissions on the filesystem: use `File.setPermissions(permissions)`
    pub fn setReadOnly(self: *Self, read_only: bool) void {
        if (read_only) {
            self.mode &= ~@as(Mode, 0o222);
        } else {
            self.mode |= @as(Mode, 0o222);
        }
    }

permissions()

On Windows, this function currently does alter the file pointer. https://github.com/ziglang/zig/issues/12783


    pub const Class = enum(u2) {
        user = 2,
        group = 1,
        other = 0,
    };

kind()

See https://github.com/ziglang/zig/issues/7699 See equivalent function: std.net.Stream.writev.


    pub const Permission = enum(u3) {
        read = 0o4,
        write = 0o2,
        execute = 0o1,
    };

accessed()

The iovecs parameter is mutable because: * This function needs to mutate the fields in order to handle partial writes from the underlying OS layer. * The OS layer expects pointer addresses to be inside the application's address space even if the length is zero. Meanwhile, in Zig, slices may have undefined pointer addresses when the length is zero. So this function modifies the iov_base fields when the length is zero. See https://github.com/ziglang/zig/issues/7699 See equivalent function: std.net.Stream.writevAll.


    /// Returns `true` if the chosen class has the selected permission.
    /// This method is only available on Unix platforms.
    pub fn unixHas(self: Self, class: Class, permission: Permission) bool {
        const mask = @as(Mode, @intFromEnum(permission)) << @as(u3, @intFromEnum(class)) * 3;
        return self.mode & mask != 0;
    }

modified()

See https://github.com/ziglang/zig/issues/7699 On Windows, this function currently does alter the file pointer. https://github.com/ziglang/zig/issues/12783


    /// Sets the permissions for the chosen class. Any permissions set to `null` are left unchanged.
    /// This method *DOES NOT* set permissions on the filesystem: use `File.setPermissions(permissions)`
    pub fn unixSet(self: *Self, class: Class, permissions: struct {
        read: ?bool = null,
        write: ?bool = null,
        execute: ?bool = null,
    }) void {
        const shift = @as(u3, @intFromEnum(class)) * 3;
        if (permissions.read) |r| {
            if (r) {
                self.mode |= @as(Mode, 0o4) << shift;
            } else {
                self.mode &= ~(@as(Mode, 0o4) << shift);
            }
        }
        if (permissions.write) |w| {
            if (w) {
                self.mode |= @as(Mode, 0o2) << shift;
            } else {
                self.mode &= ~(@as(Mode, 0o2) << shift);
            }
        }
        if (permissions.execute) |x| {
            if (x) {
                self.mode |= @as(Mode, 0o1) << shift;
            } else {
                self.mode &= ~(@as(Mode, 0o1) << shift);
            }
        }
    }

created()

The iovecs parameter is mutable because this function needs to mutate the fields in order to handle partial writes from the underlying OS layer. See https://github.com/ziglang/zig/issues/7699 On Windows, this function currently does alter the file pointer. https://github.com/ziglang/zig/issues/12783


    /// Returns a `Permissions` struct representing the permissions from the passed mode.
    pub fn unixNew(new_mode: Mode) Self {
        return Self{
            .mode = new_mode,
        };
    }
};

MetadataError

Returns the number of bytes copied. If the number read is smaller than buffer.len, it means the in file reached the end. Reaching the end of a file is not an error condition.


pub const SetPermissionsError = ChmodError;

metadata()

null means the entire file. 0 means no bytes from the file. When this is null, trailers must be sent in a separate writev() call due to a flaw in the BSD sendfile API. Other operating systems, such as Linux, already do this anyway due to API limitations. If the size of the source file is known, passing the size here will save one syscall.


/// Sets permissions according to the provided `Permissions` struct.
/// This method is *NOT* available on WASI
pub fn setPermissions(self: File, permissions: Permissions) SetPermissionsError!void {
    switch (builtin.os.tag) {
        .windows => {
            var io_status_block: windows.IO_STATUS_BLOCK = undefined;
            var info = windows.FILE_BASIC_INFORMATION{
                .CreationTime = 0,
                .LastAccessTime = 0,
                .LastWriteTime = 0,
                .ChangeTime = 0,
                .FileAttributes = permissions.inner.attributes,
            };
            const rc = windows.ntdll.NtSetInformationFile(
                self.handle,
                &io_status_block,
                &info,
                @sizeOf(windows.FILE_BASIC_INFORMATION),
                .FileBasicInformation,
            );
            switch (rc) {
                .SUCCESS => return,
                .INVALID_HANDLE => unreachable,
                .ACCESS_DENIED => return error.AccessDenied,
                else => return windows.unexpectedStatus(rc),
            }
        },
        .wasi => @compileError("Unsupported OS"), // Wasi filesystem does not *yet* support chmod
        else => {
            try self.chmod(permissions.inner.mode);
        },
    }

WriteFileError

The trailer count is inferred from headers_and_trailers.len - header_count

}

updateTimes()

Does not try seeking in either of the File parameters. See writeFileAll as an alternative to calling this.


/// Cross-platform representation of file metadata.
/// Platform-specific functionality is available through the `inner` field.
pub const Metadata = struct {
    /// Exposes platform-specific functionality.
    inner: switch (builtin.os.tag) {
        .windows => MetadataWindows,
        .linux => MetadataLinux,
        .wasi => MetadataWasi,
        else => MetadataUnix,
    },

readToEndAlloc()

Low level function which can fail for OS-specific reasons. See writeFileAll as an alternative to calling this. TODO integrate with async I/O


    const Self = @This();

readToEndAllocOptions()

Blocks when an incompatible lock is held by another process. A process may hold only one type of lock (shared or exclusive) on a file. When a process terminates in any way, the lock is released. Assumes the file is unlocked. TODO: integrate with async I/O


    /// Returns the size of the file
    pub fn size(self: Self) u64 {
        return self.inner.size();
    }

ReadError

Assumes the file is locked.


    /// Returns a `Permissions` struct, representing the permissions on the file
    pub fn permissions(self: Self) Permissions {
        return self.inner.permissions();
    }

PReadError

Attempts to obtain a lock, returning true if the lock is obtained, and false if there was an existing incompatible lock held. A process may hold only one type of lock (shared or exclusive) on a file. When a process terminates in any way, the lock is released. Assumes the file is unlocked. TODO: integrate with async I/O


    /// Returns the `Kind` of file.
    /// On Windows, can only return: `.file`, `.directory`, `.sym_link` or `.unknown`
    pub fn kind(self: Self) Kind {
        return self.inner.kind();
    }

read()

Assumes the file is already locked in exclusive mode. Atomically modifies the lock to be in shared mode, without releasing it. TODO: integrate with async I/O


    /// Returns the last time the file was accessed in nanoseconds since UTC 1970-01-01
    pub fn accessed(self: Self) i128 {
        return self.inner.accessed();
    }

readAll()


    /// Returns the time the file was modified in nanoseconds since UTC 1970-01-01
    pub fn modified(self: Self) i128 {
        return self.inner.modified();
    }

pread()


    /// Returns the time the file was created in nanoseconds since UTC 1970-01-01
    /// On Windows, this cannot return null
    /// On Linux, this returns null if the filesystem does not support creation times, or if the kernel is older than 4.11
    /// On Unices, this returns null if the filesystem or OS does not support creation times
    /// On MacOS, this returns the ctime if the filesystem does not support creation times; this is insanity, and yet another reason to hate on Apple
    pub fn created(self: Self) ?i128 {
        return self.inner.created();
    }
};

preadAll()


pub const MetadataUnix = struct {
    stat: posix.Stat,

readv()


    const Self = @This();

readvAll()


    /// Returns the size of the file
    pub fn size(self: Self) u64 {
        return @intCast(self.stat.size);
    }

preadv()


    /// Returns a `Permissions` struct, representing the permissions on the file
    pub fn permissions(self: Self) Permissions {
        return .{ .inner = .{ .mode = self.stat.mode } };
    }

preadvAll()


    /// Returns the `Kind` of the file
    pub fn kind(self: Self) Kind {
        if (builtin.os.tag == .wasi and !builtin.link_libc) return switch (self.stat.filetype) {
            .BLOCK_DEVICE => .block_device,
            .CHARACTER_DEVICE => .character_device,
            .DIRECTORY => .directory,
            .SYMBOLIC_LINK => .sym_link,
            .REGULAR_FILE => .file,
            .SOCKET_STREAM, .SOCKET_DGRAM => .unix_domain_socket,
            else => .unknown,
        };

WriteError


        const m = self.stat.mode & posix.S.IFMT;

PWriteError


        switch (m) {
            posix.S.IFBLK => return .block_device,
            posix.S.IFCHR => return .character_device,
            posix.S.IFDIR => return .directory,
            posix.S.IFIFO => return .named_pipe,
            posix.S.IFLNK => return .sym_link,
            posix.S.IFREG => return .file,
            posix.S.IFSOCK => return .unix_domain_socket,
            else => {},
        }

write()


        if (builtin.os.tag.isSolarish()) switch (m) {
            posix.S.IFDOOR => return .door,
            posix.S.IFPORT => return .event_port,
            else => {},
        };

writeAll()


        return .unknown;
    }

pwrite()


    /// Returns the last time the file was accessed in nanoseconds since UTC 1970-01-01
    pub fn accessed(self: Self) i128 {
        const atime = self.stat.atime();
        return @as(i128, atime.tv_sec) * std.time.ns_per_s + atime.tv_nsec;
    }

pwriteAll()


    /// Returns the last time the file was modified in nanoseconds since UTC 1970-01-01
    pub fn modified(self: Self) i128 {
        const mtime = self.stat.mtime();
        return @as(i128, mtime.tv_sec) * std.time.ns_per_s + mtime.tv_nsec;
    }

writev()


    /// Returns the time the file was created in nanoseconds since UTC 1970-01-01.
    /// Returns null if this is not supported by the OS or filesystem
    pub fn created(self: Self) ?i128 {
        if (!@hasDecl(@TypeOf(self.stat), "birthtime")) return null;
        const birthtime = self.stat.birthtime();

writevAll()


        // If the filesystem doesn't support this the value *should* be:
        // On FreeBSD: tv_nsec = 0, tv_sec = -1
        // On NetBSD and OpenBSD: tv_nsec = 0, tv_sec = 0
        // On MacOS, it is set to ctime -- we cannot detect this!!
        switch (builtin.os.tag) {
            .freebsd => if (birthtime.tv_sec == -1 and birthtime.tv_nsec == 0) return null,
            .netbsd, .openbsd => if (birthtime.tv_sec == 0 and birthtime.tv_nsec == 0) return null,
            .macos => {},
            else => @compileError("Creation time detection not implemented for OS"),
        }

pwritev()


        return @as(i128, birthtime.tv_sec) * std.time.ns_per_s + birthtime.tv_nsec;
    }
};

pwritevAll()


/// `MetadataUnix`, but using Linux's `statx` syscall.
/// On Linux versions below 4.11, `statx` will be filled with data from stat.
pub const MetadataLinux = struct {
    statx: std.os.linux.Statx,

CopyRangeError


    const Self = @This();

copyRange()


    /// Returns the size of the file
    pub fn size(self: Self) u64 {
        return self.statx.size;
    }

copyRangeAll()


    /// Returns a `Permissions` struct, representing the permissions on the file
    pub fn permissions(self: Self) Permissions {
        return Permissions{ .inner = PermissionsUnix{ .mode = self.statx.mode } };
    }

WriteFileOptions


    /// Returns the `Kind` of the file
    pub fn kind(self: Self) Kind {
        const m = self.statx.mode & posix.S.IFMT;

WriteFileError


        switch (m) {
            posix.S.IFBLK => return .block_device,
            posix.S.IFCHR => return .character_device,
            posix.S.IFDIR => return .directory,
            posix.S.IFIFO => return .named_pipe,
            posix.S.IFLNK => return .sym_link,
            posix.S.IFREG => return .file,
            posix.S.IFSOCK => return .unix_domain_socket,
            else => {},
        }

writeFileAll()


        return .unknown;
    }

writeFileAllUnseekable()


    /// Returns the last time the file was accessed in nanoseconds since UTC 1970-01-01
    pub fn accessed(self: Self) i128 {
        return @as(i128, self.statx.atime.tv_sec) * std.time.ns_per_s + self.statx.atime.tv_nsec;
    }

Reader


    /// Returns the last time the file was modified in nanoseconds since UTC 1970-01-01
    pub fn modified(self: Self) i128 {
        return @as(i128, self.statx.mtime.tv_sec) * std.time.ns_per_s + self.statx.mtime.tv_nsec;
    }

reader()


    /// Returns the time the file was created in nanoseconds since UTC 1970-01-01.
    /// Returns null if this is not supported by the filesystem, or on kernels before than version 4.11
    pub fn created(self: Self) ?i128 {
        if (self.statx.mask & std.os.linux.STATX_BTIME == 0) return null;
        return @as(i128, self.statx.btime.tv_sec) * std.time.ns_per_s + self.statx.btime.tv_nsec;
    }
};

Writer


pub const MetadataWasi = struct {
    stat: std.os.wasi.filestat_t,

writer()


    pub fn size(self: @This()) u64 {
        return self.stat.size;
    }

SeekableStream


    pub fn permissions(self: @This()) Permissions {
        return .{ .inner = .{ .mode = self.stat.mode } };
    }

seekableStream()


    pub fn kind(self: @This()) Kind {
        return switch (self.stat.filetype) {
            .BLOCK_DEVICE => .block_device,
            .CHARACTER_DEVICE => .character_device,
            .DIRECTORY => .directory,
            .SYMBOLIC_LINK => .sym_link,
            .REGULAR_FILE => .file,
            .SOCKET_STREAM, .SOCKET_DGRAM => .unix_domain_socket,
            else => .unknown,
        };
    }

LockError


    pub fn accessed(self: @This()) i128 {
        return self.stat.atim;
    }

lock()


    pub fn modified(self: @This()) i128 {
        return self.stat.mtim;
    }

unlock()


    pub fn created(self: @This()) ?i128 {
        return self.stat.ctim;
    }
};

tryLock()


pub const MetadataWindows = struct {
    attributes: windows.DWORD,
    reparse_tag: windows.DWORD,
    _size: u64,
    access_time: i128,
    modified_time: i128,
    creation_time: i128,

downgradeLock()


    const Self = @This();

    /// Returns the size of the file
    pub fn size(self: Self) u64 {
        return self._size;
    }

    /// Returns a `Permissions` struct, representing the permissions on the file
    pub fn permissions(self: Self) Permissions {
        return .{ .inner = .{ .attributes = self.attributes } };
    }

    /// Returns the `Kind` of the file.
    /// Can only return: `.file`, `.directory`, `.sym_link` or `.unknown`
    pub fn kind(self: Self) Kind {
        if (self.attributes & windows.FILE_ATTRIBUTE_REPARSE_POINT != 0) {
            if (self.reparse_tag & windows.reparse_tag_name_surrogate_bit != 0) {
                return .sym_link;
            }
        } else if (self.attributes & windows.FILE_ATTRIBUTE_DIRECTORY != 0) {
            return .directory;
        } else {
            return .file;
        }
        return .unknown;
    }

    /// Returns the last time the file was accessed in nanoseconds since UTC 1970-01-01
    pub fn accessed(self: Self) i128 {
        return self.access_time;
    }

    /// Returns the time the file was modified in nanoseconds since UTC 1970-01-01
    pub fn modified(self: Self) i128 {
        return self.modified_time;
    }

    /// Returns the time the file was created in nanoseconds since UTC 1970-01-01.
    /// This never returns null, only returning an optional for compatibility with other OSes
    pub fn created(self: Self) ?i128 {
        return self.creation_time;
    }
};

pub const MetadataError = posix.FStatError;

pub fn metadata(self: File) MetadataError!Metadata {
    return .{
        .inner = switch (builtin.os.tag) {
            .windows => blk: {
                var io_status_block: windows.IO_STATUS_BLOCK = undefined;
                var info: windows.FILE_ALL_INFORMATION = undefined;

                const rc = windows.ntdll.NtQueryInformationFile(self.handle, &io_status_block, &info, @sizeOf(windows.FILE_ALL_INFORMATION), .FileAllInformation);
                switch (rc) {
                    .SUCCESS => {},
                    // Buffer overflow here indicates that there is more information available than was able to be stored in the buffer
                    // size provided. This is treated as success because the type of variable-length information that this would be relevant for
                    // (name, volume name, etc) we don't care about.
                    .BUFFER_OVERFLOW => {},
                    .INVALID_PARAMETER => unreachable,
                    .ACCESS_DENIED => return error.AccessDenied,
                    else => return windows.unexpectedStatus(rc),
                }

                const reparse_tag: windows.DWORD = reparse_blk: {
                    if (info.BasicInformation.FileAttributes & windows.FILE_ATTRIBUTE_REPARSE_POINT != 0) {
                        var tag_info: windows.FILE_ATTRIBUTE_TAG_INFO = undefined;
                        const tag_rc = windows.ntdll.NtQueryInformationFile(self.handle, &io_status_block, &tag_info, @sizeOf(windows.FILE_ATTRIBUTE_TAG_INFO), .FileAttributeTagInformation);
                        switch (tag_rc) {
                            .SUCCESS => {},
                            // INFO_LENGTH_MISMATCH and ACCESS_DENIED are the only documented possible errors
                            // https://learn.microsoft.com/en-us/openspecs/windows_protocols/ms-fscc/d295752f-ce89-4b98-8553-266d37c84f0e
                            .INFO_LENGTH_MISMATCH => unreachable,
                            .ACCESS_DENIED => return error.AccessDenied,
                            else => return windows.unexpectedStatus(rc),
                        }
                        break :reparse_blk tag_info.ReparseTag;
                    }
                    break :reparse_blk 0;
                };

                break :blk .{
                    .attributes = info.BasicInformation.FileAttributes,
                    .reparse_tag = reparse_tag,
                    ._size = @as(u64, @bitCast(info.StandardInformation.EndOfFile)),
                    .access_time = windows.fromSysTime(info.BasicInformation.LastAccessTime),
                    .modified_time = windows.fromSysTime(info.BasicInformation.LastWriteTime),
                    .creation_time = windows.fromSysTime(info.BasicInformation.CreationTime),
                };
            },
            .linux => blk: {
                const l = std.os.linux;
                var stx = std.mem.zeroes(l.Statx);
                const rcx = l.statx(self.handle, "\x00", l.AT.EMPTY_PATH, l.STATX_TYPE |
                    l.STATX_MODE | l.STATX_ATIME | l.STATX_MTIME | l.STATX_BTIME, &stx);

                switch (posix.errno(rcx)) {
                    .SUCCESS => {},
                    // NOSYS happens when `statx` is unsupported, which is the case on kernel versions before 4.11
                    // Here, we call `fstat` and fill `stx` with the data we need
                    .NOSYS => {
                        const st = try posix.fstat(self.handle);

                        stx.mode = @as(u16, @intCast(st.mode));

                        // Hacky conversion from timespec to statx_timestamp
                        stx.atime = std.mem.zeroes(l.statx_timestamp);
                        stx.atime.tv_sec = st.atim.tv_sec;
                        stx.atime.tv_nsec = @as(u32, @intCast(st.atim.tv_nsec)); // Guaranteed to succeed (tv_nsec is always below 10^9)

                        stx.mtime = std.mem.zeroes(l.statx_timestamp);
                        stx.mtime.tv_sec = st.mtim.tv_sec;
                        stx.mtime.tv_nsec = @as(u32, @intCast(st.mtim.tv_nsec));

                        stx.mask = l.STATX_BASIC_STATS | l.STATX_MTIME;
                    },
                    .BADF => unreachable,
                    .FAULT => unreachable,
                    .NOMEM => return error.SystemResources,
                    else => |err| return posix.unexpectedErrno(err),
                }

                break :blk .{
                    .statx = stx,
                };
            },
            .wasi => .{ .stat = try std.os.fstat_wasi(self.handle) },
            else => .{ .stat = try posix.fstat(self.handle) },
        },
    };
}

pub const UpdateTimesError = posix.FutimensError || windows.SetFileTimeError;

/// The underlying file system may have a different granularity than nanoseconds,
/// and therefore this function cannot guarantee any precision will be stored.
/// Further, the maximum value is limited by the system ABI. When a value is provided
/// that exceeds this range, the value is clamped to the maximum.
/// TODO: integrate with async I/O
pub fn updateTimes(
    self: File,
    /// access timestamp in nanoseconds
    atime: i128,
    /// last modification timestamp in nanoseconds
    mtime: i128,
) UpdateTimesError!void {
    if (builtin.os.tag == .windows) {
        const atime_ft = windows.nanoSecondsToFileTime(atime);
        const mtime_ft = windows.nanoSecondsToFileTime(mtime);
        return windows.SetFileTime(self.handle, null, &atime_ft, &mtime_ft);
    }
    const times = [2]posix.timespec{
        posix.timespec{
            .tv_sec = math.cast(isize, @divFloor(atime, std.time.ns_per_s)) orelse maxInt(isize),
            .tv_nsec = math.cast(isize, @mod(atime, std.time.ns_per_s)) orelse maxInt(isize),
        },
        posix.timespec{
            .tv_sec = math.cast(isize, @divFloor(mtime, std.time.ns_per_s)) orelse maxInt(isize),
            .tv_nsec = math.cast(isize, @mod(mtime, std.time.ns_per_s)) orelse maxInt(isize),
        },
    };
    try posix.futimens(self.handle, &times);
}

/// Reads all the bytes from the current position to the end of the file.
/// On success, caller owns returned buffer.
/// If the file is larger than `max_bytes`, returns `error.FileTooBig`.
pub fn readToEndAlloc(self: File, allocator: Allocator, max_bytes: usize) ![]u8 {
    return self.readToEndAllocOptions(allocator, max_bytes, null, @alignOf(u8), null);
}

/// Reads all the bytes from the current position to the end of the file.
/// On success, caller owns returned buffer.
/// If the file is larger than `max_bytes`, returns `error.FileTooBig`.
/// If `size_hint` is specified the initial buffer size is calculated using
/// that value, otherwise an arbitrary value is used instead.
/// Allows specifying alignment and a sentinel value.
pub fn readToEndAllocOptions(
    self: File,
    allocator: Allocator,
    max_bytes: usize,
    size_hint: ?usize,
    comptime alignment: u29,
    comptime optional_sentinel: ?u8,
) !(if (optional_sentinel) |s| [:s]align(alignment) u8 else []align(alignment) u8) {
    // If no size hint is provided fall back to the size=0 code path
    const size = size_hint orelse 0;

    // The file size returned by stat is used as hint to set the buffer
    // size. If the reported size is zero, as it happens on Linux for files
    // in /proc, a small buffer is allocated instead.
    const initial_cap = (if (size > 0) size else 1024) + @intFromBool(optional_sentinel != null);
    var array_list = try std.ArrayListAligned(u8, alignment).initCapacity(allocator, initial_cap);
    defer array_list.deinit();

    self.reader().readAllArrayListAligned(alignment, &array_list, max_bytes) catch |err| switch (err) {
        error.StreamTooLong => return error.FileTooBig,
        else => |e| return e,
    };

    if (optional_sentinel) |sentinel| {
        return try array_list.toOwnedSliceSentinel(sentinel);
    } else {
        return try array_list.toOwnedSlice();
    }
}

pub const ReadError = posix.ReadError;
pub const PReadError = posix.PReadError;

pub fn read(self: File, buffer: []u8) ReadError!usize {
    if (is_windows) {
        return windows.ReadFile(self.handle, buffer, null);
    }

    return posix.read(self.handle, buffer);
}

/// Returns the number of bytes read. If the number read is smaller than `buffer.len`, it
/// means the file reached the end. Reaching the end of a file is not an error condition.
pub fn readAll(self: File, buffer: []u8) ReadError!usize {
    var index: usize = 0;
    while (index != buffer.len) {
        const amt = try self.read(buffer[index..]);
        if (amt == 0) break;
        index += amt;
    }
    return index;
}

/// On Windows, this function currently does alter the file pointer.
/// https://github.com/ziglang/zig/issues/12783
pub fn pread(self: File, buffer: []u8, offset: u64) PReadError!usize {
    if (is_windows) {
        return windows.ReadFile(self.handle, buffer, offset);
    }

    return posix.pread(self.handle, buffer, offset);
}

/// Returns the number of bytes read. If the number read is smaller than `buffer.len`, it
/// means the file reached the end. Reaching the end of a file is not an error condition.
/// On Windows, this function currently does alter the file pointer.
/// https://github.com/ziglang/zig/issues/12783
pub fn preadAll(self: File, buffer: []u8, offset: u64) PReadError!usize {
    var index: usize = 0;
    while (index != buffer.len) {
        const amt = try self.pread(buffer[index..], offset + index);
        if (amt == 0) break;
        index += amt;
    }
    return index;
}

/// See https://github.com/ziglang/zig/issues/7699
pub fn readv(self: File, iovecs: []const posix.iovec) ReadError!usize {
    if (is_windows) {
        // TODO improve this to use ReadFileScatter
        if (iovecs.len == 0) return @as(usize, 0);
        const first = iovecs[0];
        return windows.ReadFile(self.handle, first.iov_base[0..first.iov_len], null);
    }

    return posix.readv(self.handle, iovecs);
}

/// Returns the number of bytes read. If the number read is smaller than the total bytes
/// from all the buffers, it means the file reached the end. Reaching the end of a file
/// is not an error condition.
///
/// The `iovecs` parameter is mutable because:
/// * This function needs to mutate the fields in order to handle partial
///   reads from the underlying OS layer.
/// * The OS layer expects pointer addresses to be inside the application's address space
///   even if the length is zero. Meanwhile, in Zig, slices may have undefined pointer
///   addresses when the length is zero. So this function modifies the iov_base fields
///   when the length is zero.
///
/// Related open issue: https://github.com/ziglang/zig/issues/7699
pub fn readvAll(self: File, iovecs: []posix.iovec) ReadError!usize {
    if (iovecs.len == 0) return 0;

    // We use the address of this local variable for all zero-length
    // vectors so that the OS does not complain that we are giving it
    // addresses outside the application's address space.
    var garbage: [1]u8 = undefined;
    for (iovecs) |*v| {
        if (v.iov_len == 0) v.iov_base = &garbage;
    }

    var i: usize = 0;
    var off: usize = 0;
    while (true) {
        var amt = try self.readv(iovecs[i..]);
        var eof = amt == 0;
        off += amt;
        while (amt >= iovecs[i].iov_len) {
            amt -= iovecs[i].iov_len;
            i += 1;
            if (i >= iovecs.len) return off;
            eof = false;
        }
        if (eof) return off;
        iovecs[i].iov_base += amt;
        iovecs[i].iov_len -= amt;
    }
}

/// See https://github.com/ziglang/zig/issues/7699
/// On Windows, this function currently does alter the file pointer.
/// https://github.com/ziglang/zig/issues/12783
pub fn preadv(self: File, iovecs: []const posix.iovec, offset: u64) PReadError!usize {
    if (is_windows) {
        // TODO improve this to use ReadFileScatter
        if (iovecs.len == 0) return @as(usize, 0);
        const first = iovecs[0];
        return windows.ReadFile(self.handle, first.iov_base[0..first.iov_len], offset);
    }

    return posix.preadv(self.handle, iovecs, offset);
}

/// Returns the number of bytes read. If the number read is smaller than the total bytes
/// from all the buffers, it means the file reached the end. Reaching the end of a file
/// is not an error condition.
/// The `iovecs` parameter is mutable because this function needs to mutate the fields in
/// order to handle partial reads from the underlying OS layer.
/// See https://github.com/ziglang/zig/issues/7699
/// On Windows, this function currently does alter the file pointer.
/// https://github.com/ziglang/zig/issues/12783
pub fn preadvAll(self: File, iovecs: []posix.iovec, offset: u64) PReadError!usize {
    if (iovecs.len == 0) return 0;

    var i: usize = 0;
    var off: usize = 0;
    while (true) {
        var amt = try self.preadv(iovecs[i..], offset + off);
        var eof = amt == 0;
        off += amt;
        while (amt >= iovecs[i].iov_len) {
            amt -= iovecs[i].iov_len;
            i += 1;
            if (i >= iovecs.len) return off;
            eof = false;
        }
        if (eof) return off;
        iovecs[i].iov_base += amt;
        iovecs[i].iov_len -= amt;
    }
}

pub const WriteError = posix.WriteError;
pub const PWriteError = posix.PWriteError;

pub fn write(self: File, bytes: []const u8) WriteError!usize {
    if (is_windows) {
        return windows.WriteFile(self.handle, bytes, null);
    }

    return posix.write(self.handle, bytes);
}

pub fn writeAll(self: File, bytes: []const u8) WriteError!void {
    var index: usize = 0;
    while (index < bytes.len) {
        index += try self.write(bytes[index..]);
    }
}

/// On Windows, this function currently does alter the file pointer.
/// https://github.com/ziglang/zig/issues/12783
pub fn pwrite(self: File, bytes: []const u8, offset: u64) PWriteError!usize {
    if (is_windows) {
        return windows.WriteFile(self.handle, bytes, offset);
    }

    return posix.pwrite(self.handle, bytes, offset);
}

/// On Windows, this function currently does alter the file pointer.
/// https://github.com/ziglang/zig/issues/12783
pub fn pwriteAll(self: File, bytes: []const u8, offset: u64) PWriteError!void {
    var index: usize = 0;
    while (index < bytes.len) {
        index += try self.pwrite(bytes[index..], offset + index);
    }
}

/// See https://github.com/ziglang/zig/issues/7699
/// See equivalent function: `std.net.Stream.writev`.
pub fn writev(self: File, iovecs: []const posix.iovec_const) WriteError!usize {
    if (is_windows) {
        // TODO improve this to use WriteFileScatter
        if (iovecs.len == 0) return @as(usize, 0);
        const first = iovecs[0];
        return windows.WriteFile(self.handle, first.iov_base[0..first.iov_len], null);
    }

    return posix.writev(self.handle, iovecs);
}

/// The `iovecs` parameter is mutable because:
/// * This function needs to mutate the fields in order to handle partial
///   writes from the underlying OS layer.
/// * The OS layer expects pointer addresses to be inside the application's address space
///   even if the length is zero. Meanwhile, in Zig, slices may have undefined pointer
///   addresses when the length is zero. So this function modifies the iov_base fields
///   when the length is zero.
/// See https://github.com/ziglang/zig/issues/7699
/// See equivalent function: `std.net.Stream.writevAll`.
pub fn writevAll(self: File, iovecs: []posix.iovec_const) WriteError!void {
    if (iovecs.len == 0) return;

    // We use the address of this local variable for all zero-length
    // vectors so that the OS does not complain that we are giving it
    // addresses outside the application's address space.
    var garbage: [1]u8 = undefined;
    for (iovecs) |*v| {
        if (v.iov_len == 0) v.iov_base = &garbage;
    }

    var i: usize = 0;
    while (true) {
        var amt = try self.writev(iovecs[i..]);
        while (amt >= iovecs[i].iov_len) {
            amt -= iovecs[i].iov_len;
            i += 1;
            if (i >= iovecs.len) return;
        }
        iovecs[i].iov_base += amt;
        iovecs[i].iov_len -= amt;
    }
}

/// See https://github.com/ziglang/zig/issues/7699
/// On Windows, this function currently does alter the file pointer.
/// https://github.com/ziglang/zig/issues/12783
pub fn pwritev(self: File, iovecs: []posix.iovec_const, offset: u64) PWriteError!usize {
    if (is_windows) {
        // TODO improve this to use WriteFileScatter
        if (iovecs.len == 0) return @as(usize, 0);
        const first = iovecs[0];
        return windows.WriteFile(self.handle, first.iov_base[0..first.iov_len], offset);
    }

    return posix.pwritev(self.handle, iovecs, offset);
}

/// The `iovecs` parameter is mutable because this function needs to mutate the fields in
/// order to handle partial writes from the underlying OS layer.
/// See https://github.com/ziglang/zig/issues/7699
/// On Windows, this function currently does alter the file pointer.
/// https://github.com/ziglang/zig/issues/12783
pub fn pwritevAll(self: File, iovecs: []posix.iovec_const, offset: u64) PWriteError!void {
    if (iovecs.len == 0) return;

    var i: usize = 0;
    var off: u64 = 0;
    while (true) {
        var amt = try self.pwritev(iovecs[i..], offset + off);
        off += amt;
        while (amt >= iovecs[i].iov_len) {
            amt -= iovecs[i].iov_len;
            i += 1;
            if (i >= iovecs.len) return;
        }
        iovecs[i].iov_base += amt;
        iovecs[i].iov_len -= amt;
    }
}

pub const CopyRangeError = posix.CopyFileRangeError;

pub fn copyRange(in: File, in_offset: u64, out: File, out_offset: u64, len: u64) CopyRangeError!u64 {
    const adjusted_len = math.cast(usize, len) orelse maxInt(usize);
    const result = try posix.copy_file_range(in.handle, in_offset, out.handle, out_offset, adjusted_len, 0);
    return result;
}

/// Returns the number of bytes copied. If the number read is smaller than `buffer.len`, it
/// means the in file reached the end. Reaching the end of a file is not an error condition.
pub fn copyRangeAll(in: File, in_offset: u64, out: File, out_offset: u64, len: u64) CopyRangeError!u64 {
    var total_bytes_copied: u64 = 0;
    var in_off = in_offset;
    var out_off = out_offset;
    while (total_bytes_copied < len) {
        const amt_copied = try copyRange(in, in_off, out, out_off, len - total_bytes_copied);
        if (amt_copied == 0) return total_bytes_copied;
        total_bytes_copied += amt_copied;
        in_off += amt_copied;
        out_off += amt_copied;
    }
    return total_bytes_copied;
}

pub const WriteFileOptions = struct {
    in_offset: u64 = 0,

    /// `null` means the entire file. `0` means no bytes from the file.
    /// When this is `null`, trailers must be sent in a separate writev() call
    /// due to a flaw in the BSD sendfile API. Other operating systems, such as
    /// Linux, already do this anyway due to API limitations.
    /// If the size of the source file is known, passing the size here will save one syscall.
    in_len: ?u64 = null,

    headers_and_trailers: []posix.iovec_const = &[0]posix.iovec_const{},

    /// The trailer count is inferred from `headers_and_trailers.len - header_count`
    header_count: usize = 0,
};

pub const WriteFileError = ReadError || error{EndOfStream} || WriteError;

pub fn writeFileAll(self: File, in_file: File, args: WriteFileOptions) WriteFileError!void {
    return self.writeFileAllSendfile(in_file, args) catch |err| switch (err) {
        error.Unseekable,
        error.FastOpenAlreadyInProgress,
        error.MessageTooBig,
        error.FileDescriptorNotASocket,
        error.NetworkUnreachable,
        error.NetworkSubsystemFailed,
        => return self.writeFileAllUnseekable(in_file, args),

        else => |e| return e,
    };
}

/// Does not try seeking in either of the File parameters.
/// See `writeFileAll` as an alternative to calling this.
pub fn writeFileAllUnseekable(self: File, in_file: File, args: WriteFileOptions) WriteFileError!void {
    const headers = args.headers_and_trailers[0..args.header_count];
    const trailers = args.headers_and_trailers[args.header_count..];

    try self.writevAll(headers);

    try in_file.reader().skipBytes(args.in_offset, .{ .buf_size = 4096 });

    var fifo = std.fifo.LinearFifo(u8, .{ .Static = 4096 }).init();
    if (args.in_len) |len| {
        var stream = std.io.limitedReader(in_file.reader(), len);
        try fifo.pump(stream.reader(), self.writer());
    } else {
        try fifo.pump(in_file.reader(), self.writer());
    }

    try self.writevAll(trailers);
}

/// Low level function which can fail for OS-specific reasons.
/// See `writeFileAll` as an alternative to calling this.
/// TODO integrate with async I/O
fn writeFileAllSendfile(self: File, in_file: File, args: WriteFileOptions) posix.SendFileError!void {
    const count = blk: {
        if (args.in_len) |l| {
            if (l == 0) {
                return self.writevAll(args.headers_and_trailers);
            } else {
                break :blk l;
            }
        } else {
            break :blk 0;
        }
    };
    const headers = args.headers_and_trailers[0..args.header_count];
    const trailers = args.headers_and_trailers[args.header_count..];
    const zero_iovec = &[0]posix.iovec_const{};
    // When reading the whole file, we cannot put the trailers in the sendfile() syscall,
    // because we have no way to determine whether a partial write is past the end of the file or not.
    const trls = if (count == 0) zero_iovec else trailers;
    const offset = args.in_offset;
    const out_fd = self.handle;
    const in_fd = in_file.handle;
    const flags = 0;
    var amt: usize = 0;
    hdrs: {
        var i: usize = 0;
        while (i < headers.len) {
            amt = try posix.sendfile(out_fd, in_fd, offset, count, headers[i..], trls, flags);
            while (amt >= headers[i].iov_len) {
                amt -= headers[i].iov_len;
                i += 1;
                if (i >= headers.len) break :hdrs;
            }
            headers[i].iov_base += amt;
            headers[i].iov_len -= amt;
        }
    }
    if (count == 0) {
        var off: u64 = amt;
        while (true) {
            amt = try posix.sendfile(out_fd, in_fd, offset + off, 0, zero_iovec, zero_iovec, flags);
            if (amt == 0) break;
            off += amt;
        }
    } else {
        var off: u64 = amt;
        while (off < count) {
            amt = try posix.sendfile(out_fd, in_fd, offset + off, count - off, zero_iovec, trailers, flags);
            off += amt;
        }
        amt = @as(usize, @intCast(off - count));
    }
    var i: usize = 0;
    while (i < trailers.len) {
        while (amt >= trailers[i].iov_len) {
            amt -= trailers[i].iov_len;
            i += 1;
            if (i >= trailers.len) return;
        }
        trailers[i].iov_base += amt;
        trailers[i].iov_len -= amt;
        amt = try posix.writev(self.handle, trailers[i..]);
    }
}

pub const Reader = io.Reader(File, ReadError, read);

pub fn reader(file: File) Reader {
    return .{ .context = file };
}

pub const Writer = io.Writer(File, WriteError, write);

pub fn writer(file: File) Writer {
    return .{ .context = file };
}

pub const SeekableStream = io.SeekableStream(
    File,
    SeekError,
    GetSeekPosError,
    seekTo,
    seekBy,
    getPos,
    getEndPos,
);

pub fn seekableStream(file: File) SeekableStream {
    return .{ .context = file };
}

const range_off: windows.LARGE_INTEGER = 0;
const range_len: windows.LARGE_INTEGER = 1;

pub const LockError = error{
    SystemResources,
    FileLocksNotSupported,
} || posix.UnexpectedError;

/// Blocks when an incompatible lock is held by another process.
/// A process may hold only one type of lock (shared or exclusive) on
/// a file. When a process terminates in any way, the lock is released.
///
/// Assumes the file is unlocked.
///
/// TODO: integrate with async I/O
pub fn lock(file: File, l: Lock) LockError!void {
    if (is_windows) {
        var io_status_block: windows.IO_STATUS_BLOCK = undefined;
        const exclusive = switch (l) {
            .none => return,
            .shared => false,
            .exclusive => true,
        };
        return windows.LockFile(
            file.handle,
            null,
            null,
            null,
            &io_status_block,
            &range_off,
            &range_len,
            null,
            windows.FALSE, // non-blocking=false
            @intFromBool(exclusive),
        ) catch |err| switch (err) {
            error.WouldBlock => unreachable, // non-blocking=false
            else => |e| return e,
        };
    } else {
        return posix.flock(file.handle, switch (l) {
            .none => posix.LOCK.UN,
            .shared => posix.LOCK.SH,
            .exclusive => posix.LOCK.EX,
        }) catch |err| switch (err) {
            error.WouldBlock => unreachable, // non-blocking=false
            else => |e| return e,
        };
    }
}

/// Assumes the file is locked.
pub fn unlock(file: File) void {
    if (is_windows) {
        var io_status_block: windows.IO_STATUS_BLOCK = undefined;
        return windows.UnlockFile(
            file.handle,
            &io_status_block,
            &range_off,
            &range_len,
            null,
        ) catch |err| switch (err) {
            error.RangeNotLocked => unreachable, // Function assumes unlocked.
            error.Unexpected => unreachable, // Resource deallocation must succeed.
        };
    } else {
        return posix.flock(file.handle, posix.LOCK.UN) catch |err| switch (err) {
            error.WouldBlock => unreachable, // unlocking can't block
            error.SystemResources => unreachable, // We are deallocating resources.
            error.FileLocksNotSupported => unreachable, // We already got the lock.
            error.Unexpected => unreachable, // Resource deallocation must succeed.
        };
    }
}

/// Attempts to obtain a lock, returning `true` if the lock is
/// obtained, and `false` if there was an existing incompatible lock held.
/// A process may hold only one type of lock (shared or exclusive) on
/// a file. When a process terminates in any way, the lock is released.
///
/// Assumes the file is unlocked.
///
/// TODO: integrate with async I/O
pub fn tryLock(file: File, l: Lock) LockError!bool {
    if (is_windows) {
        var io_status_block: windows.IO_STATUS_BLOCK = undefined;
        const exclusive = switch (l) {
            .none => return,
            .shared => false,
            .exclusive => true,
        };
        windows.LockFile(
            file.handle,
            null,
            null,
            null,
            &io_status_block,
            &range_off,
            &range_len,
            null,
            windows.TRUE, // non-blocking=true
            @intFromBool(exclusive),
        ) catch |err| switch (err) {
            error.WouldBlock => return false,
            else => |e| return e,
        };
    } else {
        posix.flock(file.handle, switch (l) {
            .none => posix.LOCK.UN,
            .shared => posix.LOCK.SH | posix.LOCK.NB,
            .exclusive => posix.LOCK.EX | posix.LOCK.NB,
        }) catch |err| switch (err) {
            error.WouldBlock => return false,
            else => |e| return e,
        };
    }
    return true;
}

/// Assumes the file is already locked in exclusive mode.
/// Atomically modifies the lock to be in shared mode, without releasing it.
///
/// TODO: integrate with async I/O
pub fn downgradeLock(file: File) LockError!void {
    if (is_windows) {
        // On Windows it works like a semaphore + exclusivity flag. To implement this
        // function, we first obtain another lock in shared mode. This changes the
        // exclusivity flag, but increments the semaphore to 2. So we follow up with
        // an NtUnlockFile which decrements the semaphore but does not modify the
        // exclusivity flag.
        var io_status_block: windows.IO_STATUS_BLOCK = undefined;
        windows.LockFile(
            file.handle,
            null,
            null,
            null,
            &io_status_block,
            &range_off,
            &range_len,
            null,
            windows.TRUE, // non-blocking=true
            windows.FALSE, // exclusive=false
        ) catch |err| switch (err) {
            error.WouldBlock => unreachable, // File was not locked in exclusive mode.
            else => |e| return e,
        };
        return windows.UnlockFile(
            file.handle,
            &io_status_block,
            &range_off,
            &range_len,
            null,
        ) catch |err| switch (err) {
            error.RangeNotLocked => unreachable, // File was not locked.
            error.Unexpected => unreachable, // Resource deallocation must succeed.
        };
    } else {
        return posix.flock(file.handle, posix.LOCK.SH | posix.LOCK.NB) catch |err| switch (err) {
            error.WouldBlock => unreachable, // File was not locked in exclusive mode.
            else => |e| return e,
        };
    }
}

const File = @This();
const std = @import("../std.zig");
const builtin = @import("builtin");
const Allocator = std.mem.Allocator;
const posix = std.posix;
const io = std.io;
const math = std.math;
const assert = std.debug.assert;
const windows = std.os.windows;
const Os = std.builtin.Os;
const maxInt = std.math.maxInt;
const is_windows = builtin.os.tag == .windows;