# -*- coding: utf-8 -*-
# :-
# :Copyright: (c) 2020-2025 Franz Glasner
# :License:   BSD-3-Clause
# :-
r"""Utility sub-module to implement a file and stream digest computations.

"""

__all__ = ["compute_digest_file", "compute_digest_stream"]


import errno
import io
import os
try:
    import mmap
except ImportError:
    mmap = None
import stat

from . import PY2
from . import constants


def compute_digest_file(hashobj, path, use_mmap=None):
    """Compute the digest for a file with a filename of an open fd.

    :param hashobj: a :mod:`hashlib` compatible hash algorithm type or factory
    :param path: filename within the filesystem or a file descriptor opened in
                 binary mode (also a socket or pipe)
    :param use_mmap: Use the :mod:`mmap` module if available.
                     If `None` determine automatically.
    :type use_mmap: bool or None
    :return: the digest in binary form
    :rtype: bytes

    If a file descriptor is given is must support :func:`os.read`.

    """
    h = hashobj()
    if isinstance(path, constants.PATH_TYPES):
        flags = os.O_RDONLY | getattr(os, "O_BINARY", 0) \
            | getattr(os, "O_SEQUENTIAL", 0) | getattr(os, "O_NOCTTY", 0)
        fd = os.open(path, flags)
        own_fd = True
    else:
        fd = path
        own_fd = False
    try:
        try:
            st = os.fstat(fd)
        except TypeError:
            #
            # "fd" is most probably a Python socket object.
            # (a pipe typically supports fstat)
            #
            use_mmap = False
        else:
            if stat.S_ISREG(st[stat.ST_MODE]):
                filesize = st[stat.ST_SIZE]
                if (use_mmap is None) \
                        and (filesize > constants.MAX_AUTO_MAP_SIZE):
                    #
                    # This is borrowed from FreeBSD's cp(1) implementation:
                    # Mmap and process if less than 8M (the limit is
                    # so we don't totally trash memory on big files.
                    # This is really a minor hack, but it wins some
                    # CPU back.  Some filesystems, such as smbnetfs,
                    # don't support mmap, so this is a best-effort
                    # attempt.
                    #
                    use_mmap = False
            else:
                use_mmap = False
        if use_mmap is None:
            use_mmap = True
        if mmap is None or not use_mmap:
            # No mmap available or wanted -> use traditional low-level file IO
            fadvise = getattr(os, "posix_fadvise", None)
            if fadvise:
                fadvise(fd, 0, 0, os.POSIX_FADV_SEQUENTIAL)
            if not PY2:
                fileobj = io.FileIO(fd, mode="r", closefd=False)
                buf = bytearray(constants.READ_CHUNK_SIZE)
                with memoryview(buf) as full_view:
                    while True:
                        try:
                            n = fileobj.readinto(buf)
                        except OSError as e:
                            if e.errno not in (errno.EAGAIN,
                                               errno.EWOULDBLOCK,
                                               errno.EINTR):
                                raise
                        else:
                            if n == 0:
                                break
                            if n == constants.READ_CHUNK_SIZE:
                                h.update(buf)
                            else:
                                with full_view[:n] as partial_view:
                                    h.update(partial_view)
            else:
                while True:
                    try:
                        buf = os.read(fd, constants.READ_CHUNK_SIZE)
                    except OSError as e:
                        if e.errno not in (errno.EAGAIN,
                                           errno.EWOULDBLOCK,
                                           errno.EINTR):
                            raise
                    else:
                        if len(buf) == 0:
                            break
                        h.update(buf)
        else:
            #
            # Use mmap
            #
            # NOTE: On Windows mmapped files with length 0 are not supported.
            #       So ensure to not call mmap.mmap() if the file size is 0.
            #
            madvise = getattr(mmap.mmap, "madvise", None)
            if filesize <= constants.MAP_WINDOW_SIZE:
                mapsize = filesize
            else:
                mapsize = constants.MAP_WINDOW_SIZE
            mapoffset = 0
            rest = filesize
            while rest > 0:
                m = mmap.mmap(fd,
                              mapsize,
                              access=mmap.ACCESS_READ,
                              offset=mapoffset)
                if madvise:
                    madvise(m, mmap.MADV_SEQUENTIAL)
                try:
                    h.update(m)
                finally:
                    m.close()
                rest -= mapsize
                mapoffset += mapsize
                if rest < mapsize:
                    mapsize = rest
    finally:
        if own_fd:
            os.close(fd)
    return h.digest()


def compute_digest_stream(hashobj, instream):
    """Compute the digest for a given byte string `instream`.

    :param hashobj: a :mod:`hashlib` compatible hash algorithm type or factory
    :param instream: a bytes input stream to read the data to be hashed from
    :return: the digest in binary form
    :rtype: bytes

    """
    h = hashobj()
    while True:
        try:
            buf = instream.read(constants.READ_CHUNK_SIZE)
        except OSError as e:
            if e.errno not in (errno.EAGAIN, errno.EWOULDBLOCK, errno.EINTR):
                raise
        else:
            if buf is not None:
                if len(buf) == 0:
                    break
                h.update(buf)
    return h.digest()