r"""
:Author:    Franz Glasner
:Copyright: (c) 2020 Franz Glasner.
            All rights reserved.
:License:   BSD 3-Clause "New" or "Revised" License.
            See :ref:`LICENSE <license>` for details.
            If you cannot find LICENSE see
            <https://opensource.org/licenses/BSD-3-Clause>
:ID:        @(#) $HGid$

"""

from __future__ import print_function


__version__ = "0.1"

__revision__ = "|VCSRevision|"
__date__ = "|VCSJustDate|"


import argparse
import hashlib
import io
try:
    import mmap
except ImportError:
    mmap = None
import os
import re
import stat
import sys


PY2 = sys.version_info[0] < 3

CHUNK_SIZE = 1024*1024
MAP_CHUNK_SIZE = 64*1024*1024


def main(argv=None):
    aparser = argparse.ArgumentParser(
        description="Python implementation of shasum",
        fromfile_prefix_chars='@')
    aparser.add_argument(
        "--algorithm", "-a", action="store", type=argv2algo,
        help="1 (default), 224, 256, 384, 512, 3-224, 3-256, 3-384, 3-512, blake2b, blake2s, md5")
    aparser.add_argument(
        "--binary", "-b", action="store_false", dest="text_mode", default=False,
        help="read in binary mode (default)")
    aparser.add_argument(
        "--bsd", "-B", action="store_true", dest="bsd", default=False,
        help="Write BSD style output. This is also the default output format of :command:`openssl dgst`.")
    aparser.add_argument(
        "--check", "-c", action="store_true",
        help="""Read digests from FILEs and check them.
If this option is specified, the FILE options become checklists. Each
checklist should contain hash results in a supported format, which will
be verified against the specified paths. Output consists of the digest
used, the file name, and an OK, FAILED, or MISSING for the result of
the comparison. This will validate any of the supported checksums.
If no file is given, stdin is used.""")
    aparser.add_argument(
        "--checklist", "-C", metavar="CHECKLIST",
        help="""Compare the checksum of each FILE against the checksums in
the CHECKLIST. Any specified FILE that is not listed in the CHECKLIST will
generate an error.""")

    aparser.add_argument(
        "--reverse", "-r", action="store_false", dest="bsd", default=False,
        help="explicitely select normal coreutils style output (to be option compatible with BSD style commands and :command:`openssl dgst -r`)")
    aparser.add_argument(
        "--tag", action="store_true", dest="bsd", default=False,
        help="alias for the `--bsd' option (to be compatible with :command:`b2sum`)")
    aparser.add_argument(
        "--text", "-t", action="store_true", dest="text_mode", default=False,
        help="read in text mode (not supported)")
    aparser.add_argument(
        "--version", "-v", action="version", version="%s (rv:%s)" % (__version__, __revision__))
    aparser.add_argument(
        "files", nargs="*", metavar="FILE")

    opts = aparser.parse_args(args=argv)

    if opts.text_mode:
        print("ERROR: text mode not supported", file=sys.stderr)
        sys.exit(78)   # :manpage:`sysexits(3)`  EX_CONFIG

    if opts.check and opts.checklist:
        print("ERROR: only one of --check or --checklist allowed",
              file=sys.stderr)
        sys.exit(64)   # :manpage:`sysexits(3)`  EX_USAGE

    if not opts.algorithm:
        opts.algorithm = argv2algo("1")

    if opts.check:
        return verify_digests_from_files(opts)
    elif opts.checklist:
        return verify_digests_with_checklist(opts)
    else:
        return generate_digests(opts)


def generate_digests(opts):
    if opts.bsd:
        out = out_bsd
    else:
        out = out_std
    if not opts.files or (len(opts.files) == 1 and opts.files[0] == '-'):
        if PY2:
            if sys.platform == "win32":
                import os, msvcrt   # noqa: E401
                msvcrt.setmode(sys.stdin.fileno(), os.O_BINARY)
            source = sys.stdin
        else:
            source = sys.stdin.buffer
        out(sys.stdout,
            compute_digest_stream(opts.algorithm[0], source),
            None,
            opts.algorithm[1],
            True)
    else:
        for fn in opts.files:
            out(sys.stdout,
                compute_digest_file(opts.algorithm[0], fn),
                fn,
                opts.algorithm[1],
                True)
    return 0


def verify_digests_with_checklist(opts):
    exit_code = 0
    if not opts.files or (len(opts.files) == 1 and opts.files[0] == '-'):
        if PY2:
            if sys.platform == "win32":
                import os, msvcrt   # noqa: E401
                msvcrt.setmode(sys.stdin.fileno(), os.O_BINARY)
            source = sys.stdin
        else:
            source = sys.stdin.buffer
        pl = get_parsed_digest_line_from_checklist(opts.checklist, opts, None)
        if pl is None:
            exit_code = 1
            print("-: MISSING")
        else:
            tag, algo, cl_filename, cl_digest = pl
            computed_digest = compute_digest_stream(algo, source)
            if cl_digest.lower() == computed_digest.lower():
                res = "OK"
            else:
                res = "FAILED"
                exit_code = 1
            print("{}: {}: {}".format(tag, "-", res))
    else:
        for fn in opts.files:
            pl = get_parsed_digest_line_from_checklist(opts.checklist, opts, fn)
            if pl is None:
                print("{}: MISSING".format(fn))
                exit_code = 1
            else:
                tag, algo, cl_filename, cl_digest = pl
                computed_digest = compute_digest_file(algo, fn)
                if cl_digest.lower() == computed_digest.lower():
                    res = "OK"
                else:
                    exit_code = 1
                    res = "FAILED"
                print("{}: {}: {}".format(tag, fn, res))
    return exit_code


def verify_digests_from_files(opts):
    exit_code = 0
    if not opts.files or (len(opts.files) == 1 and opts.files[0] == '-'):
        for checkline in sys.stdin:
            if not checkline:
                continue
            r, fn, tag = handle_checkline(opts, checkline)
            print("{}: {}: {}".format(tag, fn, r.upper()))
            if r != "ok" and exit_code == 0:
                exit_code = 1
    else:
        for fn in opts.files:
            with io.open(fn, "rt", encoding="utf-8") as checkfile:
                for checkline in checkfile:
                    if not checkline:
                        continue
                    r, fn, tag = handle_checkline(opts, checkline)
                    print("{}: {}: {}".format(tag, fn, r.upper()))
                    if r != "ok" and exit_code == 0:
                        exit_code = 1
    return exit_code


def handle_checkline(opts, line):
    """
    :return: a tuple with static "ok", "missing", or "failed", the filename and
             the digest used
    :rtype: tuple(str, str, str)

    """
    parts = parse_digest_line(opts, line)
    if not parts:
        raise ValueError(
            "improperly formatted digest line: {}".format(line))
    tag, algo, fn, digest = parts
    try:
        d = compute_digest_file(algo, fn)
        if d.lower() == digest.lower():
            return ("ok", fn, tag)
        else:
            return ("failed", fn, tag)
    except EnvironmentError:
        return ("missing", fn, tag)


def get_parsed_digest_line_from_checklist(checklist, opts, filename):
    if filename is None:
        filenames = ("-", "stdin", "", )
    else:
        filenames = (
            normalize_filename(filename, strip_leading_dot_slash=True),)
    with io.open(checklist, "rt", encoding="utf-8") as clf:
        for checkline in clf:
            if not checkline:
                continue
            parts = parse_digest_line(opts, checkline)
            if not parts:
                raise ValueError(
                    "improperly formatted digest line: {}".format(checkline))
            fn = normalize_filename(parts[2], strip_leading_dot_slash=True)
            if fn in filenames:
                return parts
        else:
            return None


def parse_digest_line(opts, line):
    """Parse a `line` of a digest file and return its parts.

    :return: a tuple of the normalized algorithm tag, the algorithm
             constructor, the filename and the hex digest;
             if `line` cannot be parsed successfully `None` is returned
    :rtype: tuple(str, obj, str, str) or None

    Handles coreutils and BSD-style file formats.

    """
    # determine checkfile format (BSD or coreutils)
    # BSD?
    mo = re.search(r"\A(\S+)\s*\((.*)\)\s*=\s*(.+)\n?\Z", line)
    if mo:
        # (tag, algorithm, filename, digest)
        return (mo.group(1),
                algotag2algotype(mo.group(1)),
                mo.group(2),
                mo.group(3))
    else:
        # coreutils?
        mo = re.search(r"([^\ ]+) [\*\ ]?(.+)\n?\Z", line)
        if mo:
            # (tag, algorithm, filename, digest)
            return (opts.algorithm[1],
                    opts.algorithm[0],
                    mo.group(2),
                    mo.group(1))
        else:
            return None


def argv2algo(s):
    """Convert a commane line algorithm specifier into a tuple with the
    type/factory of the digest and the algorithms tag for output purposes.

    :param str s: the specifier from the commane line
    :return: the internal digest specification
    :rtype: a tuple (digest_type_or_factory, name_in_output)

    String comparisons are done case-insensitively.

    """
    s = s.lower()
    if s in ("1", "sha1"):
        return (hashlib.sha1, "SHA1")
    elif s in ("224", "sha224"):
        return (hashlib.sha224, "SHA224")
    elif s in ("256", "sha256"):
        return (hashlib.sha256, "SHA256")
    elif s in ("384", "sha384"):
        return (hashlib.sha384, "SHA384")
    elif s in ("512", "sha512"):
        return (hashlib.sha512, "SHA512")
    elif s in ("3-224", "sha3-224"):
        return (hashlib.sha3_224, "SHA3-224")
    elif s in ("3-256", "sha3-256"):
        return (hashlib.sha3_256, "SHA3-256")
    elif s in ("3-384", "sha3-384"):
        return (hashlib.sha3_384, "SHA3-384")
    elif s in ("3-512", "sha3-512"):
        return (hashlib.sha3_512, "SHA3-512")
    elif s in ("blake2b", "blake2b-512"):
        return (hashlib.blake2b, "BLAKE2b")
    elif s in ("blake2s", "blake2s-256"):
        return (hashlib.blake2s, "BLAKE2s")
    elif s == "md5":
        return (hashlib.md5, "MD5")
    else:
        raise argparse.ArgumentTypeError(
            "`{}' is not a recognized algorithm".format(s))


def algotag2algotype(s):
    """Convert the algorithm specifier in a BSD-style digest file to the
    type/factory of the corresponding algorithm.

    :param str s: the tag (i.e. normalized name) or the algorithm
    :return: the digest type or factory for `s`

    All string comparisons are case-sensitive.

    """
    if s == "SHA1":
        return hashlib.sha1
    elif s == "SHA224":
        return hashlib.sha224
    elif s == "SHA256":
        return hashlib.sha256
    elif s == "SHA384":
        return hashlib.sha384
    elif s == "SHA512":
        return hashlib.sha512
    elif s == "SHA3-224":
        return hashlib.sha3_224
    elif s == "SHA3-256":
        return hashlib.sha3_256
    elif s == "SHA3-384":
        return hashlib.sha3_384
    elif s == "SHA3-512":
        return hashlib.sha3_512
    elif s == "BLAKE2b":
        return hashlib.blake2b
    elif s == "BLAKE2s":
        return hashlib.blake2s
    elif s == "MD5":
        return hashlib.md5
    else:
        raise ValueError("unknown algorithm: {}".format(s))


def out_bsd(dest, digest, filename, digestname, binary):
    """BSD format output, also :command:`openssl dgst` and
    :command:`b2sum --tag" format output

    """
    if filename is None:
        print(digest, file=dest)
    else:
        print("{} ({}) = {}".format(digestname,
                                    normalize_filename(filename),
                                    digest),
              file=dest)


def out_std(dest, digest, filename, digestname, binary):
    """Coreutils format (:command:`shasum` et al.)

    """
    print("{} {}{}".format(
                digest,
                '*' if binary else ' ',
                '-' if filename is None else normalize_filename(filename)),
          file=dest)


def compute_digest_file(hashobj, filename):
    """
    :param hashobj: a :mod:`hashlib` compatible hash algorithm type or factory
    :param str filename: filename within the filesystem
    :return: the digest in hex form
    :rtype: str

    """
    h = hashobj()
    flags = os.O_RDONLY
    try:
        flags |= os.O_BINARY
    except AttributeError:
        pass
    fd = os.open(filename, flags)
    try:
        st = os.fstat(fd)
        filesize = st[stat.ST_SIZE]
        #
        # On Windows mmapped file with length 0 are not supported
        # -> use low-level IO
        #
        if mmap is None:
            while True:
                buf = os.read(fd, CHUNK_SIZE)
                if len(buf) == 0:
                    break
                h.update(buf)
        else:
            # mmap
            if filesize < MAP_CHUNK_SIZE:
                mapsize = filesize
                mapoffset = 0
            else:
                mapsize = MAP_CHUNK_SIZE
                mapoffset = 0
            rest = filesize
            while rest > 0:
                m = mmap.mmap(fd,
                              mapsize,
                              access=mmap.ACCESS_READ,
                              offset=mapoffset)
                try:
                    h.update(m)
                finally:
                    m.close()
                rest -= mapsize
                mapoffset += mapsize
                if rest < mapsize:
                    mapsize = rest
    finally:
        os.close(fd)
    return h.hexdigest()


def compute_digest_stream(hashobj, 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 hex form
    :rtype: str

    """
    h = hashobj()
    while True:
        buf = instream.read(CHUNK_SIZE)
        if buf is not None:
            if len(buf) == 0:
                break
            h.update(buf)
    return h.hexdigest()


def normalize_filename(filename, strip_leading_dot_slash=False):
    filename = filename.replace("\\", "/")
    if strip_leading_dot_slash:
        while filename.startswith("./"):
            filename = filename[2:]
    return filename


if __name__ == "__main__":
    sys.exit(main())