macresources/bin/rfx

#!/usr/bin/env python3

# Copyright (c) 2020 Elliot Nunn

# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:

# The above copyright notice and this permission notice shall be included in all
# copies or substantial portions of the Software.

# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.


import macresources
import sys
import tempfile
import os
from os import path
import re
import subprocess


HELP = '''Usage: rfx [-c] command [arg | arg//type/id | arg//type | arg// ...]

Expose MacOS resource forks to command

Resources specified as filename//type/id are converted to tempfiles
before command is run, then back after command returns. Truncated //
arguments are wildcards.

Supports .rdump Rez files and .hqx BinHex files. Otherwise .rdump will
be appended implicitly.

Examples:
    rfx mv Doc.rdump//STR/0 Doc.rdump//STR/1
    rfx cp App.hqx//PICT allpictures/
    rfx rm System/..namedfork/rsrc//vers/2'''


if len(sys.argv) < 2 or sys.argv[1].startswith('-'):
    sys.exit(HELP)


def is_rez(the_path):
    return path.splitext(the_path)[1].lower() == '.rdump'


def is_hqx(the_path):
    return path.splitext(the_path)[1].lower() == '.hqx'


def is_fork(the_path):
    return the_path.lower().endswith('/..namedfork/rsrc') or path.splitext(the_path)[1].lower() == '.rsrc'


resourcefork_cache = {} # the_path, mutable list of resurces
inodes = {} # deduplicates file paths so we don't screw it up
hqx_saved_data = {} # stores data fork and Finder info so we don't strip it
def get_cached_file(the_path):
    path_user_entered = the_path # only for error messages

    if not (is_rez(the_path) or is_fork(the_path) or is_hqx(the_path)):
        the_path += '.rdump' # will cause is_rez to return true

    # The path is already in the cache! Hooray!
    try: return resourcefork_cache[the_path]
    except KeyError: pass

    # Hack to stop us being fooled by the same file with multiple names
    # (Doesn't help if the file doesn't exist yet -- oh well)
    try:
        stat = os.stat(the_path)
        stat = (stat.st_dev, stat.st_ino)
        the_path = inodes.setdefault(stat, the_path)

        # Have one more crack at the main cache
        try: return resourcefork_cache[the_path]
        except KeyError: pass

    except FileNotFoundError:
        pass

    try:
        with open(the_path, 'rb') as f:
            raw = f.read()

        try:
            if is_rez(the_path):
                resources = list(macresources.parse_rez_code(raw))
            elif is_fork(the_path):
                resources = list(macresources.parse_file(raw))
            elif is_hqx(the_path):
                from macresources import binhex
                hb = binhex.HexBin(raw)
                hqx_saved_data[the_path] = (hb.FName, hb.FInfo, hb.read())
                rsrc = hb.read_rsrc()
                resources = macresources.parse_file(rsrc)
        except:
            sys.exit('Corrupt: ' + repr(path_user_entered))

    except FileNotFoundError: # Treat as empty resource fork
        if is_rez(the_path):
            resources = []
        elif is_fork(the_path):
            resources = []
        elif is_hqx(the_path):
            try:
                valid_filename = path.basename(the_path)[:-4].replace(':', path.sep)
                valid_filename.encode('mac_roman')
                if len(valid_filename) > 31: raise ValueError
            except:
                sys.exit('Name not suitable for a new BinHex: ' + repr(path_user_entered))

            hqx_saved_data[the_path] = (valid_filename, None, b'')
            resources = []

    resourcefork_cache[the_path] = resources
    return resources


def flush_cache():
    for the_path, resources in resourcefork_cache.items():
        # No change, do not write the file
        if not any(getattr(res, '__rfx_dirty', False) for res in resources): continue

        # Weed out the ghost resources
        resources = [res for res in resources if not getattr(res, '__rfx_ghost', False)]

        # Support commands that pack/unpack GreggyBits etc (mistake here very rare!)
        for res in resources:
            if getattr(res, '__rfx_dirty', False):
                is_compressed = (res.startswith(b'\xA8\x9F\x65\x72') and
                    len(res) >= 6 and
                    len(res) >= int.from_bytes(res[4:6], 'big')) # hdrlen thing
                res.attribs = (res.attribs & ~1) | int(is_compressed)

        if is_rez(the_path):
            # For BASE.rdump to be valid, BASE must exist (my rule)
            try:
                with open(path.splitext(the_path)[0], 'x'): pass
            except FileExistsError:
                pass

            with open(the_path, 'wb') as f:
                f.write(macresources.make_rez_code(resources, ascii_clean=True))

        elif is_fork(the_path):
            # For BASE/..namedfork/rsrc to be openable by macOS, BASE must exist
            if the_path.lower().endswith('/..namedfork/rsrc'):
                try:
                    with open(the_path[:-17], 'x'): pass
                except FileExistsError:
                    pass

            with open(the_path, 'wb') as f:
                f.write(macresources.make_file(resources))

        elif is_hqx(the_path):
            # Get back the non-resource-fork stuff for the BinHex file
            from macresources import binhex
            fname, finfo, data = hqx_saved_data[the_path]
            rsrc = macresources.make_file(resources)
            bh = binhex.BinHex((fname, finfo, len(data), len(rsrc)), the_path)
            bh.write(b'')
            bh.write_rsrc(rsrc)
            bh.close()


def escape_ostype(ostype):
    escaped = ''
    for char in ostype:
        if ord('A') <= char <= ord('Z') or ord('a') <= char <= ord('z'):
            escaped += chr(char)
        else:
            escaped += '_%02X' % char
    return escaped


with tempfile.TemporaryDirectory() as backup_tmp_dir:
    new_argv = [sys.argv[1]]
    to_retrieve = []

    for i, arg in enumerate(sys.argv[2:], 1):
        m = re.match(r'(.*[^/])//(?:([^/]{1,4})(?:/(-?\d+)?)?)?$'.replace('/', re.escape(path.sep)), arg)

        if not m:
            # Do not expand this argument
            new_argv.append(arg)
        else:
            # Expand arg into 1+ fake-resource tempfiles, each backed by a Resource object
            res_file = m.group(1)
            res_type = m.group(2).encode('mac_roman').ljust(4)[:4] if m.group(2) else None
            res_id = int(m.group(3)) if m.group(3) else None

            if res_type is None:
                # File// = every resource
                arg_resources = get_cached_file(res_file)
            elif res_id is None:
                # File//Type/ = resources of type (can omit trailing slash)
                arg_resources = [foundres for foundres in get_cached_file(res_file) if foundres.type == res_type]
            else:
                # File//Type/ID = 1 resource
                for foundres in get_cached_file(res_file):
                    if foundres.type == res_type and foundres.id == res_id:
                        arg_resources = [foundres]
                        break
                else:
                    arg_resources = [macresources.Resource(res_type, res_id)]
                    arg_resources[0].__rfx_ghost = True
                    arg_resources[0].__rfx_dirty = False
                    get_cached_file(res_file).append(arg_resources[0])

            if not arg_resources:
                # Failed to expand so leave unchanged
                new_argv.append(arg)
            else:
                # Expand!
                for j, res in enumerate(arg_resources, 1):
                    enclosing_dir = path.join(backup_tmp_dir, '%d.%d' % (i,j))
                    os.mkdir(enclosing_dir)

                    tmp_file = path.join(enclosing_dir, '%s.%d' % (escape_ostype(res.type), res.id))
                    if not getattr(res, '__rfx_ghost', False):
                        with open(tmp_file, 'wb') as f:
                            f.write(res)

                    to_retrieve.append((tmp_file, res))
                    new_argv.append(tmp_file)

    result = subprocess.run(new_argv)

    for tmp_file, res in to_retrieve:
        try:
            with open(tmp_file, 'rb') as f:
                d = f.read()

            if getattr(res, '__rfx_ghost', False) or d != res:
                res[:] = d
                res.__rfx_dirty = True

            res.__rfx_ghost = False

        except FileNotFoundError:
            if not getattr(res, '__rfx_ghost', False):
                res.__rfx_dirty = True

            res.__rfx_ghost = True

flush_cache()

sys.exit(result.returncode)