Ticket #2864: uc1541

#! /usr/bin/env python
"""
UC1541 Virtual filesystem

This extfs provides an access to disk image files for the Commodore
VIC20/C64/C128. It requires the utility c1541 that comes bundled with Vice,
the emulator for the VIC20, C64, C128 and other computers made by Commodore.

Remarks
-------

Due to different way of representing file entries on regular D64 disk images,
there could be issues with filenames that are transfered from/to the image.
Following rules was applied to represent a single file entry:

1. An extension is attached to the end of a filename depending on a file type.
   Possible extensions are: prg, del, seq, usr and rel.
2. Every non-ASCII character (which could be some of characters specific to
   PET-ASCII, or be a control character) will be replaced by dot (.), since
   c1541 program will list them that way.
3. Every slash character (/) will be replaced by pipe character (|).
4. Leading space will be replaced by tilda (~).

While copying from D64 image to filesystem, filenames will be stored as they
are seen on a listing.

While copying from filesystem to D64 image, filename conversion will be done:
1. Every $ and * characters will be replaced by question mark (?)
2. Every pipe (|) and backslash (\) characters will be replaced by slash (/)
3. Every tilda (~) will be replaced by a space
4. 'prg' extension will be truncated

Representation of a directory can be sometimes confusing - in case when one
copied file without extension it stays there in such form, till next access
(after flushing VFS). Also file sizes are not accurate, since D64 directory
entries have sizes stored as 256 bytes blocks.

Configuration
-------------

Here are specific for this script variable, which while set, can influence
script behaviour:

UC1541_DEBUG - if set, uc1541 will produce log in /tmp/uc1541.log file

UC1541_VERBOSE - of set, script will be more verbose, i.e. error messages form
c1541 program will be passed to Midnight Commander, so that user will be aware
of error cause if any.

UC1541_HIDE_DEL - if set, no DEL entries will be shown

Changelog:
    2.3 Re added and missing method _correct_fname used for writing files
        into d64 image.
    2.2 Fixed bug(?) with unusual sector end (marked as sector 0, not 255),
        causing endless directory reading on random locations.
    2.1 Fixed bug with filenames containing slash.
    2.0 Added reading raw D64 image, and mapping for jokers. Now it is
        possible to read files with PET-ASCII/control sequences in filenames.
        Working with d64 images only. Added workaround for space at the
        beggining of the filename.
    1.2 Added configuration env variables: UC1541_VERBOSE and UC1541_HIDE_DEL.
        First one, if set to any value, will cause that error messages from
        c1541 program will be redirected as a failure messages visible in MC.
        The other variable, when set to any value, cause "del" entries to be
        not shown in the lister.
    1.1 Added protect bits, added failsafe for argparse module
    1.0 Initial release

Author: Roman 'gryf' Dobosz <gryf73@gmail.com>
Date: 2012-09-24
Version: 2.3
Licence: BSD
"""

import sys
import re
import os
from subprocess import Popen, PIPE

if os.getenv('UC1541_DEBUG'):
    import logging
    LOG = logging.getLogger('UC1541')
    LOG.setLevel(logging.DEBUG)
    FILE_HANDLER = logging.FileHandler("/tmp/uc1541.log")
    FILE_FORMATTER = logging.Formatter("%(asctime)s %(levelname)-8s "
                                       "%(lineno)s %(funcName)s - %(message)s")
    FILE_HANDLER.setFormatter(FILE_FORMATTER)
    FILE_HANDLER.setLevel(logging.DEBUG)
    LOG.addHandler(FILE_HANDLER)
else:
    class LOG(object):
        """
        Dummy logger object. does nothing.
        """
        @classmethod
        def debug(*args, **kwargs):
            pass

        @classmethod
        def info(*args, **kwargs):
            pass

        @classmethod
        def warning(*args, **kwargs):
            pass

        @classmethod
        def error(*args, **kwargs):
            pass

        @classmethod
        def critical(*args, **kwargs):
            pass


class D64(object):
    """
    Implement d64 directory reader
    """
    CHAR_MAP = {32: ' ', 33: '!', 34: '"', 35: '#', 37: '%', 38: '&', 39: "'",
                40: '(', 41: ')', 42: '*', 43: '+', 44: ',', 45: '-', 46: '.',
                47: '/', 48: '0', 49: '1', 50: '2', 51: '3', 52: '4', 53: '5',
                54: '6', 55: '7', 56: '8', 57: '9', 59: ';', 60: '<', 61: '=',
                62: '>', 63: '?', 64: '@', 65: 'a', 66: 'b', 67: 'c', 68: 'd',
                69: 'e', 70: 'f', 71: 'g', 72: 'h', 73: 'i', 74: 'j', 75: 'k',
                76: 'l', 77: 'm', 78: 'n', 79: 'o', 80: 'p', 81: 'q', 82: 'r',
                83: 's', 84: 't', 85: 'u', 86: 'v', 87: 'w', 88: 'x', 89: 'y',
                90: 'z', 91: '[', 93: ']', 97: 'A', 98: 'B', 99: 'C',
                100: 'D', 101: 'E', 102: 'F', 103: 'G', 104: 'H', 105: 'I',
                106: 'J', 107: 'K', 108: 'L', 109: 'M', 110: 'N', 111: 'O',
                112: 'P', 113: 'Q', 114: 'R', 115: 'S', 116: 'T', 117: 'U',
                118: 'V', 119: 'W', 120: 'X', 121: 'Y', 122: 'Z', 193: 'A',
                194: 'B', 195: 'C', 196: 'D', 197: 'E', 198: 'F', 199: 'G',
                200: 'H', 201: 'I', 202: 'J', 203: 'K', 204: 'L', 205: 'M',
                206: 'N', 207: 'O', 208: 'P', 209: 'Q', 210: 'R', 211: 'S',
                212: 'T', 213: 'U', 214: 'V', 215: 'W', 216: 'X', 217: 'Y',
                218: 'Z'}

    FILE_TYPES = {0b000: 'del',
                  0b001: 'seq',
                  0b010: 'prg',
                  0b011: 'usr',
                  0b100: 'rel'}

    def __init__(self, dimage):
        """
        Init
        """
        LOG.debug('image: %s', dimage)
        dimage = open(dimage, 'rb')
        self.raw = dimage.read()
        dimage.close()

        self.current_sector_data = None
        self.next_sector = 0
        self.next_track = None
        self._dir_contents = []

    def _map_filename(self, string):
        """
        Transcode filename to ASCII compatible. Replace not supported
        characters with jokers.
        """

        filename = list()

        for chr_ in string:
            if ord(chr_) == 160:  # shift+space character; $a0
                break

            character = D64.CHAR_MAP.get(ord(chr_), '?')
            filename.append(character)

        LOG.debug("string: ``%s'' mapped to: ``%s''", string,
                  "".join(filename))
        return "".join(filename)

    def _go_to_next_sector(self):
        """
        Fetch (if exist) next sector from a directory chain
        Return False if the chain ends, True otherwise
        """

        if self.next_track == 0 and self.next_sector in (0, 255):
            LOG.debug("End of directory")
            return False

        if self.next_track is None:
            LOG.debug("Going to the track: 18,1")
            offset = self._get_d64_offset(18, 1)
        else:
            offset = self._get_d64_offset(self.next_track, self.next_sector)
            LOG.debug("Going to the track: %s,%s", self.next_track,
                      self.next_sector)

        self.current_sector_data = self.raw[offset:offset + 256]

        self.next_track = ord(self.current_sector_data[0])
        self.next_sector = ord(self.current_sector_data[1])
        LOG.debug("Next track: %s,%s", self.next_track, self.next_sector)
        return True

    def _get_ftype(self, num):
        """
        Get filetype as a string
        """
        return D64.FILE_TYPES.get(int("%d%d%d" % (num & 4 and 1,
                                                  num & 2 and 1,
                                                  num & 1), 2), '???')

    def _get_d64_offset(self, track, sector):
        """
        Return offset (in bytes) for specified track and sector.
        """

        offset = 0
        truncate_track = 0

        if track > 17:
            offset = 17 * 21 * 256
            truncate_track = 17

        if track > 24:
            offset += 6 * 19 * 256
            truncate_track = 24

        if track > 30:
            offset += 5 * 18 * 256
            truncate_track = 30

        track = track - truncate_track
        offset += track * sector * 256

        return offset

    def _harvest_entries(self):
        """
        Traverse through sectors and store entries in _dir_contents
        """
        sector = self.current_sector_data
        for x in range(8):
            entry = sector[:32]
            ftype = ord(entry[2])

            if ftype == 0:  # deleted
                sector = sector[32:]
                continue

            type_verbose = self._get_ftype(ftype)

            protect = ord(entry[2]) & 64 and "<" or " "
            fname = entry[5:21]
            if ftype == 'rel':
                size = ord(entry[23])
            else:
                size = ord(entry[30]) + ord(entry[31]) * 226

            self._dir_contents.append({'fname': self._map_filename(fname),
                                       'ftype': type_verbose,
                                       'size': size,
                                       'protect': protect})
            sector = sector[32:]

    def list_dir(self):
        """
        Return directory list as list of dict with keys:
            fname, ftype, protect and size
        """
        while self._go_to_next_sector():
            self._harvest_entries()

        return self._dir_contents


class Uc1541(object):
    """
    Class for interact with c1541 program and MC
    """
    PRG = re.compile(r'(\d+)\s+"([^"]*)".+?\s(del|prg|rel|seq|usr)([\s<])')

    def __init__(self, archname):
        self.arch = archname
        self.out = ''
        self.err = ''
        self._verbose = os.getenv("UC1541_VERBOSE", False)
        self._hide_del = os.getenv("UC1541_HIDE_DEL", False)

        self.pyd64 = D64(archname).list_dir()
        self.file_map = {}
        self.directory = []

    def list(self):
        """
        Output list contents of D64 image.
        Convert filenames to be Unix filesystem friendly
        Add suffix to show user what kind of file do he dealing with.
        """
        LOG.info("List contents of %s", self.arch)
        directory = self._get_dir()

        for entry in directory:
            sys.stdout.write("%(perms)s   1 %(uid)-8d %(gid)-8d %(size)8d "
                             "Jan 01 1980 %(display_name)s\n" % entry)
        return 0

    def rm(self, dst):
        """
        Remove file from D64 image
        """
        LOG.info("Removing file %s", dst)
        dst = self._get_masked_fname(dst)

        if not self._call_command('delete', dst=dst):
            return self._show_error()

        # During removing, a message containing ERRORCODE is sent to stdout
        # instead of stderr. Everything other than 'ERRORCODE 1' (which means:
        # 'everything fine') is actually a failure. In case of verbose error
        # output it is needed to copy self.out to self.err.
        if '\nERRORCODE 1\n' not in self.out:
            self.err = self.out
            return self._show_error()

        return 0

    def copyin(self, dst, src):
        """
        Copy file to the D64 image. Destination filename has to be corrected.
        """
        LOG.info("Copy into D64 %s as %s", src, dst)
        dst = self._correct_fname(dst)

        if not self._call_command('write', src=src, dst=dst):
            return self._show_error()

        return 0

    def copyout(self, src, dst):
        """
        Copy file form the D64 image. Source filename has to be corrected,
        since it's representation differ from the real one inside D64 image.
        """
        LOG.info("Copy form D64 %s as %s", src, dst)
        if not src.endswith(".prg"):
            return "cannot read"

        src = self._get_masked_fname(src)

        if not self._call_command('read', src=src, dst=dst):
            return self._show_error()

        return 0

    def _correct_fname(self, fname):
        """
        Return filename with mapped characters, without .prg extension.
        Characters like $, *, + in filenames are perfectly legal, but c1541
        program seem to have issues with it while writing, so it will also be
        replaced.
        """
        char_map = {'|': "/",
                    "\\": "/",
                    "~": " ",
                    "$": "?",
                    "*": "?"}

        if fname.lower().endswith(".prg"):
            fname = fname[:-4]

        new_fname = []
        for char in fname:
            trans = char_map.get(char)
            new_fname.append(trans if trans else char)

        return "".join(new_fname)

    def _get_masked_fname(self, fname):
        """
        Return masked filename with '?' jokers instead of non ASCII
        characters, useful for copying or deleting files with c1541. In case
        of several files with same name exists in directory, only first one
        will be operative (first as appeared in directory).

        Warning! If there are two different names but the only difference is in
        non-ASCII characters (some PET ASCII or control characters) there is
        a risk that one can remove both files.
        """
        directory = self._get_dir()

        for entry in directory:
            if entry['display_name'] == fname:
                return entry['pattern_name']

    def _get_dir(self):
        """
        Retrieve directory via c1541 program
        """
        directory = []

        uid = os.getuid()
        gid = os.getgid()

        if not self._call_command('list'):
            return self._show_error()

        idx = 0
        for line in self.out.split("\n"):
            if Uc1541.PRG.match(line):
                blocks, fname, ext, rw = Uc1541.PRG.match(line).groups()

                if ext == 'del' and self._hide_del:
                    continue

                display_name = ".".join([fname, ext])
                pattern_name = self.pyd64[idx]['fname']

                if '/' in display_name:
                    display_name = display_name.replace('/', '|')

                # workaround for space at the beggining of the filename
                if display_name[0] == ' ':
                    display_name = '~' + display_name[1:]

                if ext == 'del':
                    perms = "----------"
                else:
                    perms = "-r%s-r--r--" % (rw.strip() and "-" or "w")

                directory.append({'pattern_name': pattern_name,
                                  'display_name': display_name,
                                  'uid': uid,
                                  'gid': gid,
                                  'size': int(blocks) * 256,
                                  'perms': perms})
                idx += 1
        return directory

    def _show_error(self):
        """
        Pass out error output from c1541 execution
        """
        if self._verbose:
            sys.exit(self.err)
        else:
            sys.exit(1)

    def _call_command(self, cmd, src=None, dst=None):
        """
        Return status of the provided command, which can be one of:
            write
            read
            delete
            dir/list
        """
        command = ['c1541', '-attach', self.arch, '-%s' % cmd]
        if src and dst:
            command.append(src)
            command.append(dst)
        elif src or dst:
            command.append(src and src or dst)

        self.out, self.err = Popen(command, stdout=PIPE,
                                   stderr=PIPE).communicate()
        return not self.err


CALL_MAP = {'list': lambda a: Uc1541(a.ARCH).list(),
            'copyin': lambda a: Uc1541(a.ARCH).copyin(a.SRC, a.DST),
            'copyout': lambda a: Uc1541(a.ARCH).copyout(a.SRC, a.DST),
            'rm': lambda a: Uc1541(a.ARCH).rm(a.DST)}


def parse_args():
    """
    Use ArgumentParser to check for script arguments and execute.
    """
    parser = ArgumentParser()
    subparsers = parser.add_subparsers(help='supported commands')
    parser_list = subparsers.add_parser('list', help="List contents of D64 "
                                        "image")
    parser_copyin = subparsers.add_parser('copyin', help="Copy file into D64 "
                                          "image")
    parser_copyout = subparsers.add_parser('copyout', help="Copy file out of "
                                           "D64 image")
    parser_rm = subparsers.add_parser('rm', help="Delete file from D64 image")

    parser_list.add_argument('ARCH', help="D64 Image filename")
    parser_list.set_defaults(func=CALL_MAP['list'])

    parser_copyin.add_argument('ARCH', help="D64 Image filename")
    parser_copyin.add_argument('SRC', help="Source filename")
    parser_copyin.add_argument('DST', help="Destination filename (to be "
                               "written into D64 image)")
    parser_copyin.set_defaults(func=CALL_MAP['copyin'])

    parser_copyout.add_argument('ARCH', help="D64 Image filename")
    parser_copyout.add_argument('SRC', help="Source filename (to be read from"
                                " D64 image")
    parser_copyout.add_argument('DST', help="Destination filename")
    parser_copyout.set_defaults(func=CALL_MAP['copyout'])

    parser_rm.add_argument('ARCH', help="D64 Image filename")
    parser_rm.add_argument('DST', help="File inside D64 image to be deleted")
    parser_rm.set_defaults(func=CALL_MAP['rm'])

    args = parser.parse_args()
    return args.func(args)


def no_parse():
    """
    Failsafe argument "parsing". Note, that it blindly takes positional
    arguments without checking them. In case of wrong arguments it will
    silently exit
    """
    try:
        if sys.argv[1] not in ('list', 'copyin', 'copyout', 'rm'):
            sys.exit(2)
    except IndexError:
        sys.exit(2)

    class Arg(object):
        DST = None
        SRC = None
        ARCH = None

    arg = Arg()

    try:
        arg.ARCH = sys.argv[2]
        if sys.argv[1] in ('copyin', 'copyout'):
            arg.SRC = sys.argv[3]
            arg.DST = sys.argv[4]
        elif sys.argv[1] == 'rm':
            arg.DST = sys.argv[3]
    except IndexError:
        sys.exit(2)

    CALL_MAP[sys.argv[1]](arg)

if __name__ == "__main__":
    LOG.debug("Script params: %s", str(sys.argv))
    try:
        from argparse import ArgumentParser
        parse_func = parse_args
    except ImportError:
        parse_func = no_parse

    parse_func()