medashare/ui/medashare/cli.py

#!/usr/bin/env python

import collections
import stat
import sys
import logging

# useful for debugging when stderr is redirected/captured
_real_stderr = sys.stderr
_sql_verbose = False

if False:
	lvl = logging.DEBUG
	lvl = logging.INFO
	_handler = logging.StreamHandler(_real_stderr)
	_handler.setLevel(lvl)
	_handler.setFormatter(logging.Formatter('%(asctime)s: %(message)s'))
	import sqlalchemy
	logging.getLogger('sqlalchemy').addHandler(_handler)
	logging.getLogger('sqlalchemy.engine').setLevel(lvl)

from .utils import _debprint, enable_debug, disable_debug

def _getquery(q, objstr):
	return repr(str(q.compile(objstr._engine,
	    compile_kwargs={"literal_binds": True})).replace('\n', ' '))

#import pdb, sys; mypdb = pdb.Pdb(stdout=sys.stderr); mypdb.set_trace()

from edgold.ed448 import EDDSA448
from unittest import mock

from .hostid import hostuuid
from .tags import TagCache
from . import orm
from .magic import detect_from_filename

from .btv import _TestCases as bttestcase, validate_file

import base64
import base58
from .btv import bencode
import copy
import datetime
import functools
import hashlib
import importlib
import io
import itertools
import json
import libarchive
import magic
import operator
import os.path
import pathlib
import pasn1
import re
import shutil
import socket
import sqlalchemy
from sqlalchemy import create_engine, select, insert, func, delete, text
from sqlalchemy.orm import sessionmaker, aliased, load_only
import string
import subprocess
import sys
import tempfile
import unittest
import uuid

# The UUID for the namespace representing the path to a file
_NAMESPACE_MEDASHARE_PATH = uuid.UUID('f6f36b62-3770-4a68-bc3d-dc3e31e429e6')
_NAMESPACE_MEDASHARE_CONTAINER = uuid.UUID('890a9d5c-0626-4de1-ab05-9e14947391eb')

_defaulthash = 'sha512'
_validhashes = set([ 'sha256', 'sha512' ])
_hashlengths = { len(getattr(hashlib, x)().hexdigest()): x for x in
	_validhashes }

def _makehashuri(hashstr):
	hash, value = ObjectStore.makehash(hashstr).split(':')

	return f'hash://{hash}/{value}'

def _keyordering(x):
	k, v = x
	try:
		return (MDBase._common_names_list.index(k), k, v)
	except ValueError:
		return (2**32, k, v)

def _iterdictlist(obj, **kwargs):
	l = list(sorted(obj.items(**kwargs), key=_keyordering))
	for k, v in l:
		if isinstance(v, list):
			for i in sorted(v):
				yield k, i
		else:
			yield k, v

_metaopre = re.compile('meta:(?P<key>[a-zA-Z]+)(?P<op>=|<|>|!=|<=|>=)(?P<val>.*)')
_metaoplookup = {
	'=': operator.eq,	'!=': operator.ne,
	'<': operator.lt,	'>': operator.gt,
	'<=': operator.le,	'>=': operator.ge
}
_metavalueconv = {
	'modified':	datetime.datetime.fromisoformat,
}

from .utils import _makeuuid, _makedatetime, _asn1coder

from .mdb import MDBase

class MetaData(MDBase):
	_type = 'metadata'

	_uniq_properties = set([ 'ms:tag' ])

class Identity(MDBase):
	_type = 'identity'

	# Identites don't need a created by
	_common_properties = [ x for x in MDBase._common_properties if x !=
	    'created_by_ref' ]
	_common_optional = set([ x for x in MDBase._common_optional if x !=
	    'parent_refs' ] + [ 'name', 'pubkey' ])
	_common_names = set(_common_properties + list(
	    MDBase._generated_properties.keys()))

class Persona(object):
	'''The object that represents a persona, or identity.  It will
	create the proper identity object, serialize for saving keys,
	create objects for that persona and other management.'''

	def __init__(self, identity=None, key=None):
		if identity is None:
			self._identity = Identity()
		else:
			self._identity = identity

		self._key = key
		self._pubkey = None

		if 'pubkey' in self._identity:
			pubkeybytes = self._identity.pubkey
			self._pubkey = EDDSA448(pub=pubkeybytes)

		self._created_by_ref = self._identity.uuid

	def Host(self, *args, **kwargs):
		kwargs['created_by_ref'] = self.uuid

		return self.sign(Host(*args, **kwargs))

	def Mapping(self, *args, **kwargs):
		kwargs['created_by_ref'] = self.uuid

		return self.sign(Mapping(*args, **kwargs))

	def Container(self, *args, **kwargs):
		kwargs['created_by_ref'] = self.uuid

		return self.sign(Container(*args, **kwargs))

	def MetaData(self, *args, **kwargs):
		kwargs['created_by_ref'] = self.uuid

		return self.sign(MetaData(*args, **kwargs))

	@property
	def uuid(self):
		'''Return the UUID of the identity represented.'''

		return self._identity.uuid

	def __repr__(self):	# pragma: no cover
		r = '<Persona: has key: %s, has pubkey: %s, identity: %s>' % \
		    (self._key is not None, self._pubkey is not None,
		    repr(self._identity))

		return r

	@classmethod
	def from_pubkey(cls, pubkeystr):
		pubstr = base58.b58decode_check(pubkeystr)

		uuid, pubkey = _asn1coder.loads(pubstr)

		ident = Identity(uuid=uuid, pubkey=pubkey)

		return cls(ident)

	def get_identity(self):
		'''Return the Identity object for this Persona.'''

		return self._identity

	def get_pubkey(self):
		'''Get a printable version of the public key.  This is used
		for importing into different programs, or for shared.'''

		idobj = self._identity
		pubstr = _asn1coder.dumps([ idobj.uuid, idobj.pubkey ])

		return base58.b58encode_check(pubstr).decode('ascii')

	def new_version(self, *args):
		'''Update the Persona's Identity object.'''

		self._identity = self.sign(self._identity.new_version(*args))

		return self._identity

	def store(self, fname):
		'''Store the Persona to a file.  If there is a private
		key associated w/ the Persona, it will be saved as well.'''

		with open(fname, 'wb') as fp:
			obj = {
				'identity': self._identity,
			}
			if self._key is not None:
				obj['key'] = \
				    self._key.export_key('raw')

			fp.write(_asn1coder.dumps(obj))

	@classmethod
	def load(cls, fname):
		'''Load the Persona from the provided file.'''

		with open(fname, 'rb') as fp:
			objs = _asn1coder.loads(fp.read())

		kwargs = {}
		if 'key' in objs:
			kwargs['key'] = EDDSA448(objs['key'])

		return cls(Identity(objs['identity']), **kwargs)

	def generate_key(self):
		'''Generate a key for this Identity.

		Raises a RuntimeError if a key is already present.'''

		if self._key:
			raise RuntimeError('a key already exists')

		self._key = EDDSA448.generate()
		self._pubkey = self._key.public_key()
		pubkey = self._pubkey.export_key('raw')
		self._identity = self.sign(self._identity.new_version(('pubkey',
		    pubkey)))

	def _makesigbytes(self, obj):
		obj = dict(obj.items(False))
		try:
			del obj['sig']
		except KeyError:
			pass

		return _asn1coder.dumps(obj)

	def sign(self, obj):
		'''Takes the object, adds a signature, and returns the new
		object.'''

		sigbytes = self._makesigbytes(obj)
		sig = self._key.sign(sigbytes)
		newobj = MDBase.create_obj(obj)
		newobj.sig = sig

		return newobj

	def verify(self, obj):
		sigbytes = self._makesigbytes(obj)

		pubkey = self._pubkey.export_key('raw')
		self._pubkey.verify(obj['sig'], sigbytes)

		return True

	def by_file(self, fname):
		'''Return a file object for the file named fname.'''

		fobj = FileObject.from_file(fname, self._created_by_ref)

		return self.sign(fobj)

class ObjectStore(object):
	'''A container to store the various MetaData objects.'''

	# The _uuids property contains both the UUIDv4 for objects, and
	# looking up the UUIDv5 for FileObjects.

	def __init__(self, engine, version='head'):
		# Uncomment when working on the db schema
		#orm.Base.metadata.create_all(engine)

		self._engine = engine
		self._ses = sessionmaker(engine)

		self._handle_migration(version)

	def _handle_migration(self, version):
		'''Handle migrating the database to a newer version.'''

		# running commands directly:
		# pydoc3 alembic.config.Config
		# pydoc3 alembic.commands

		# inspecting the scripts directly:
		# alembic/script/base.py:61

		from alembic import command
		from alembic.config import Config

		config = Config()
		config.set_main_option("script_location", "medashare:alembic")

		with self._engine.begin() as connection:
			config.attributes['engine'] = self._engine
			command.upgrade(config, version)

	def get_host(self, hostuuid):
		hostuuid = _makeuuid(hostuuid)

		with self._ses() as session:
			a = session.get(orm.HostTable, hostuuid)

			if a is None:
				raise KeyError(hostuuid)

			return self._by_id(a.objid, session)

	def get_by_type(self, _type):
		try:
			if issubclass(_type, MDBase):
				_type = _type._type
		except TypeError:
			pass

		with self._ses() as session:
			for i in session.query(orm.MetaDataObject.data).where(
			    orm.MetaDataObject.type == _type):
				yield i.data

	def get_hosts(self):
		return self.get_by_type(Host)

	@staticmethod
	def makehash(hashstr, strict=True):
		'''Take a hash or hash string, and return a valid hash
		string from it.

		This makes sure that it is of the correct type and length.

		If strict is False, the function will detect the length and
		return a valid hash string if one can be found.

		By default, the string must be prepended by the type,
		followed by a colon, followed by the value in hex in all
		lower case characters.'''

		try:
			hash, value = hashstr.split(':')
		except ValueError:
			if strict:
				raise

			hash = _hashlengths[len(hashstr)]
			value = hashstr

		bvalue = value.encode('ascii')
		if strict and len(bvalue.translate(None,
		    string.hexdigits.lower().encode('ascii'))) != 0:
			raise ValueError('value has invalid hex digits (must be lower case)', value)

		if hash in _validhashes:
			return ':'.join((hash, value))

		raise ValueError

	def __len__(self):
		with self._ses() as session:
			return list(session.query(func.count(
			    orm.MetaDataObject.uuid)))[0][0]

	def __iter__(self):
		with self._ses() as session:
			for i in session.query(orm.MetaDataObject.data).all():
				yield i.data

	@classmethod
	def load(cls, fname):
		engine = create_engine("sqlite+pysqlite:///%s" % fname,
		    echo=_sql_verbose, future=True)

		return cls(engine)

	def store(self, fname):
		'''Write out the objects in the store to the file named
		fname.'''

		pass

	def _add_uuidv5(self, id, obj, session):
		session.execute(delete(orm.UUIDv5Table).where(
		    orm.UUIDv5Table.uuid == id))

		o = orm.UUIDv5Table(uuid=id, objid=obj.uuid)
		session.add(o)

	def _lock(self, session):
		'''Function to issue a write to "lock" the database transaction.'''

		res = list(session.scalars(select(orm.Dummy).where(
		    orm.Dummy.id == 1)))

		if res:
			session.delete(res[0])
		else:
			d = orm.Dummy(id=1)
			session.add(d)

	@staticmethod
	def _update_metadata_indexes(session, obj, strcache):
		# sqlalchemy doesn't cache inserts, so don't insert dups
		# ourselves
		propmapcache = set()

		# clear out old data
		stmt = delete(orm.PropertyMapping).where(
		    orm.PropertyMapping.obj == obj.uuid)
		session.execute(stmt)

		props = [ x for x in obj.items() if x[0] not in {
		    'hashes',
		    'sig',
		    'parent_refs',
		    } ]
		for k, vids in props:
			kid = strcache[k]

			if not isinstance(vids, list):
				vids = [ vids ]

			vids = [ strcache[sv] for sv in vids ]

			for v in vids:
				if (obj.uuid, kid, v) in propmapcache:
					continue

				session.add(orm.PropertyMapping(obj=obj.uuid,
				    keyid=kid, valueid=v))
				propmapcache.add((obj.uuid, kid, v))

	def loadobj(self, obj):
		'''Load obj into the data store.'''

		obj = MDBase.create_obj(obj)

		with self._ses() as session:
			self._lock(session)

			oldobj = session.get(orm.MetaDataObject, obj.uuid)
			#if oldobj.modified > obj.modified:
			#	return
			if oldobj is not None:
				# XXX - missing cleanup of indexes
				session.delete(oldobj)

			sobj = orm.MetaDataObject(uuid=obj.uuid, type=obj.type,
			    modified=obj.modified, data=obj)
			session.add(sobj)

			if obj.type == 'file':
				objid = _makeuuid(obj.id)
				oldobj = self._by_id(objid, session)
				if oldobj is not None:
					# pick which obj
					if oldobj.modified > obj.modified:
						session.delete(session.get(
						    orm.MetaDataObject,
						    obj.uuid))
						obj = oldobj
					else:
						# get ride of old obj
						session.delete(session.get(
						    orm.MetaDataObject,
						    oldobj.uuid))

				self._add_uuidv5(obj.id, obj, session)
			elif obj.type == 'container':
				self._add_uuidv5(obj.make_id(obj.uri), obj,
				    session)
			elif obj.type == 'host':
				o = orm.HostTable(hostid=_makeuuid(
				    obj.hostuuid), objid=obj.uuid)
				session.add(o)
			elif obj.type == 'mapping':
				hostid = _makeuuid(hostuuid())

				maps = [ (lambda a, b: orm.HostMapping(
				    hostid=uuid.UUID(a), objid=obj.uuid))(
				    *x.split(':', 1)) for x in obj.mapping ]
				session.add_all(maps)
			elif obj.type == 'metadata':
				self._update_metadata_indexes(session, obj,
				    StringCache(session))

			try:
				hashes = obj.hashes
			except AttributeError:
				pass
			else:
				for j in hashes:
					h = self.makehash(j)
					r = session.get(orm.HashTable,
					    dict(hash=h, uuid=obj.uuid))
					if r is None:
						session.add(orm.HashTable(
						    hash=h, uuid=obj.uuid))

			session.commit()

	def drop_uuid(self, uuid):
		uuid = _makeuuid(uuid)

		with self._ses() as session:
			obj = session.get(orm.MetaDataObject, uuid)
			session.delete(obj)

			obj = obj.data

			if obj.type == 'file':
				session.execute(delete(orm.UUIDv5Table).where(
				    orm.UUIDv5Table.uuid == obj.id))

			for j in obj.hashes:
				h = self.makehash(j)
				session.execute(delete(orm.HashTable).where(
				    orm.HashTable.hash == h and
				    orm.HashTable.uuid == obj.uuid))

			session.commit()

	def by_id(self, id):
		'''Look up an object by it's UUID.'''

		id = _makeuuid(id)

		with self._ses() as session:
			res = self._by_id(id, session)
			if res is None:
				raise KeyError(id)

		return res

	def _by_id(self, id, session):
		if id.version == 5:
			res = session.get(orm.UUIDv5Table, id)
			if res is None:
				return

			id = res.objid

		res = session.get(orm.MetaDataObject, id)
		if res is None:
			return

		return res.data

	def by_hash(self, hash, types=None):
		'''Look up an object by it's hash value.

		types is either a list of types, or None meaning all.'''

		h = self.makehash(hash, strict=False)

		types = True if types is None else \
		    orm.MetaDataObject.type.in_(types)

		sel = select(orm.MetaDataObject.data).where(
		    orm.MetaDataObject.uuid == orm.HashTable.uuid,
		     orm.HashTable.hash == h, types)

		with self._ses() as session:
			r = list(session.scalars(sel))

		return r

	def get_metadata(self, fname, persona, create_metadata=True):
		'''Get all MetaData objects for fname, or create one if
		not found.

		If a FileObject is not present, one will be created.

		A Persona must be passed in to create the FileObject and
		MetaData objects as needed.

		A MetaData object will be created if create_metadata is
		True, which is the default.

		Note: if a new MetaData object is created, it is not
		stored in the database automatically.  It is expected that
		it will be modified and then saved, so call ObjectStore.loadobj
		with it to save it.
		'''

		try:
			fobj = self.by_file(fname, ('file',))[0]
		except KeyError:
			fobj = persona.by_file(fname)

			self.loadobj(fobj)

		# we now have the fobj, get the metadata for it.
		try:
			objs = self.by_file(fname)
		except KeyError:
			if create_metadata:
				objs = [ persona.MetaData(hashes=fobj.hashes) ]
			else:
				objs = [ ]

		return objs

	def get_hostmappings(self):
		'''Returns the tuple (lclpath, hostid, rempath) for all
		the mappings for this hostid.'''

		hostid = _makeuuid(hostuuid())

		sel = select(orm.MetaDataObject.data).where(
		    orm.HostMapping.hostid == hostid,
		    orm.HostMapping.objid == orm.MetaDataObject.uuid)

		res = []
		with self._ses() as session:
			# XXX - view
			for obj in session.scalars(sel):
				maps = [ (lambda a, b: (uuid.UUID(a),
				    pathlib.Path(b).resolve()))(*x.split(':',
				    1)) for x in obj.mapping ]
				for idx, (id, path) in enumerate(maps):
					if hostid == id:
						# add other to mapping
						other = tuple(maps[(idx + 1) %
						    2])
						res.append((path, ) + other)
		return res

	def by_file(self, fname, types=('metadata', )):
		'''Return a metadata object for the file named fname.

		Will check the mapping database to get hashes, and possibly
		return that FileObject if requested.

		Will raise a KeyError if this file does not exist in
		the database.

		Will raise a ValueError if fname currently does not
		match what is in the database.
		'''

		fid = FileObject.make_id(fname)

		#print('bf:', repr(fid), file=_real_stderr)
		try:
			fobj = self.by_id(fid)
			lclfile = None
		except KeyError:
			# check mappings
			fname = pathlib.Path(fname).resolve()
			for lclpath, hostid, rempath in self.get_hostmappings():
				if fname.parts[:len(lclpath.parts)] == lclpath.parts:
					try:
						rempath = pathlib.Path(
						    *rempath.parts +
						    fname.parts[len(
						    lclpath.parts):])
						fid = FileObject.make_id(
						    rempath, hostid)
						fobj = self.by_id(fid)
						lclfile = fname
						break
					except KeyError:
						continue
			else:
				raise

		fobj.verify(lclfile)

		for i in fobj.hashes:
			j = self.by_hash(i)

			# Filter out non-metadata objects
			j = [ x for x in j if x.type in types ]
			if j:
				return j
		else:
			raise KeyError('unable to find metadata for file: %s' %
			    repr(fname))

def _readfp(fp):
	while True:
		r = fp.read(64*1024)
		# libarchive returns None on EOF
		if r == b'' or r is None:
			return

		yield r

def _hashfile(fname):
	with open(fname, 'rb') as fp:
		return _hashfp(fp)

def _hashfp(fp):
	hash = getattr(hashlib, _defaulthash)()
	for r in _readfp(fp):
		hash.update(r)

	return '%s:%s' % (_defaulthash, hash.hexdigest())

class Host(MDBase):
	_type = 'host'

	_class_instance_properties = {
		'hostuuid': _makeuuid,
	}

class Mapping(MDBase):
	_type = 'mapping'

class FileObject(MDBase):
	_type = 'file'

	_class_instance_properties = {
		'hostid': _makeuuid,
		'id': _makeuuid,
		'mtime': _makedatetime,
	}

	@property
	def fullname(self):
		return os.path.join(self.dir, self.filename)

	@staticmethod
	def prep_mapping(mapping):
		r = collections.defaultdict(lambda: [])
		for lclp, hid, remp in mapping:
			r[hid].append((pathlib.PosixPath(lclp),
			    pathlib.PosixPath(remp)))

		return r

	def get_lcl_name(self, mapping):
		'''Using mapping (obtained from the object store's
		get_hostmappings method and passed through local prep_mapping
		method), try to map the file to a local path.

		if not possible, returned will be hostid:remotepath.
		'''

		lclhostid = hostuuid()

		#_debprint('gln:', repr(self), repr(mapping), repr(lclhostid))

		if self.hostid == lclhostid:
			return self.fullname

		for lclp, remp in mapping[self.hostid]:
			fname = pathlib.PosixPath(self.fullname)
			if fname.parts[:len(remp.parts)] == remp.parts:
				return pathlib.Path(
				    *lclp.parts +
				    fname.parts[len(
				    remp.parts):])
		else:
			return str(self.hostid) + ':' + self.fullname

	@staticmethod
	def make_id(fname, hostid=None):
		'''Take a local file name, and make the id for it.  Note that
		converts from the local path separator to a forward slash so
		that it will be the same between Windows and Unix systems.'''

		if hostid is None:
			hostid = hostuuid()

		fname = os.path.realpath(fname)
		return uuid.uuid5(_NAMESPACE_MEDASHARE_PATH,
		    str(hostid) + '/'.join(os.path.split(fname)))

	_statsymbtoname = { getattr(stat, x): 'stat.' + x for x in dir(stat) if x.startswith('S_') }

	@classmethod
	def _modetosymbolic(cls, mode): # pragma: no cover
		r = []
		while mode:
			nbit = -mode & mode

			r.append(cls._statsymbtoname[nbit])

			mode = mode & ~nbit

		return '|'.join(r)

	@classmethod
	def _real_stat_repr(cls, st): # pragma: no cover
		return 'os.stat_result' \
		    '((%s, %d, %d, %d, %d, %d, %d, %d, %.6f, %d))' % \
		    (cls._modetosymbolic(st.st_mode), 10, 100, 1, 100, 100,
		    st.st_size, st.st_atime, st.st_mtime, st.st_ctime)

	@classmethod
	def from_file(cls, filename, created_by_ref):
		filename = os.path.abspath(filename)
		s = os.stat(filename)
		# keep so that when new files are added, it's easy to get stat
		#_debprint(repr(filename), cls._real_stat_repr(s))
		# XXX - race here, fix w/ checking mtime before/after?
		obj = {
			'created_by_ref': created_by_ref,
			'hostid': hostuuid(),
			'dir': os.path.dirname(filename),
			'filename': os.path.basename(filename),
			'id': cls.make_id(filename),
			'mtime': datetime.datetime.fromtimestamp(s.st_mtime,
			    tz=datetime.timezone.utc),
			'size': s.st_size,
			'hashes': [ _hashfile(filename), ],
		}

		return cls(obj)

	def verify(self, lclfile=None):
		'''Verify that this FileObject is still valid.  It will
		by default, only do a mtime verification.

		It will raise a ValueError if the file does not match.'''

		if lclfile is None:
			s = os.stat(os.path.join(self.dir, self.filename))
		else:
			s = os.stat(lclfile)

		mtimets = datetime.datetime.fromtimestamp(s.st_mtime,
		    tz=datetime.timezone.utc).timestamp()

		#print(repr(self), repr(s), s.st_mtime, file=_real_stderr)
		if self.mtime.timestamp() != mtimets or \
		    self.size != s.st_size:
			_debprint('times:', self.mtime.timestamp(), mtimets)
			raise ValueError('file %s has changed' %
			    repr(self.filename))

class Container(MDBase):
	_type = 'container'

	_common_optional = MDBase._common_optional | set([ 'uri' ])

	@staticmethod
	def make_id(uri):
		return uuid.uuid5(_NAMESPACE_MEDASHARE_CONTAINER, uri)

def enumeratedir(_dir, created_by_ref):
	'''Enumerate all the files and directories (not recursive) in _dir.

	Returned is a list of FileObjects.'''

	return [FileObject.from_file(os.path.join(_dir, x),
	    created_by_ref) for x in sorted(os.listdir(_dir)) if not
	    os.path.isdir(os.path.join(_dir, x)) ]

def _get_paths(options):
	fnames = (
	    '.medashare_identity.pasn1',
	    '.medashare_store.sqlite3',
	    '.medashare_cache.pasn1' )

	if 'MEDASHARE_PATH' in os.environ:
		return ( os.path.expanduser(
		    os.path.join(os.environ['MEDASHARE_PATH'], x)) for x in
		    fnames )

	return ( os.path.expanduser('~/' + x) for x in fnames )

class StringCache:
	def __init__(self, session):
		self._ses = session
		self._cache = {}

	def __getitem__(self, k):
		try:
			return self._cache[k]
		except KeyError:
			pass

		v = self._ses.execute(select(orm.StringTable.id).where(
		    orm.StringTable.str == k)).first()
		if v is None:
			# not present, insert it
			st = self._ses.add(orm.StringTable(str=k))

			v = self._ses.execute(select(orm.StringTable.id)
			    .where(orm.StringTable.str == k)).first()

		v = v[0]

		self._cache[k] = v

		return v

def init_datastructs(f):
	@functools.wraps(f)
	def wrapper(options):

		identfname, storefname, cachefname = _get_paths(options)

		# create the persona
		try:
			persona = Persona.load(identfname)
		except FileNotFoundError:
			print('ERROR: Identity not created, create w/ genident.',
			    file=sys.stderr)
			sys.exit(1)

		# create the object store
		engine = create_engine("sqlite+pysqlite:///%s" % storefname,
		    echo=_sql_verbose, future=True)

		objstr = ObjectStore(engine)

		# create the cache
		cache = TagCache.load(cachefname)

		try:
			r = f(options, persona, objstr, cache)

			# per: https://www.sqlite.org/lang_analyze.html
			with engine.connect() as con:
				con.execute(text('PRAGMA analysis_limit = 400;'))
				con.execute(text('PRAGMA optimize;'))
				con.commit()
			return r
		finally:
			if cache.modified:
				cache.store(cachefname)
	return wrapper

def cmd_genident(options):
	identfname, _, _ = _get_paths(options)

	if os.path.exists(identfname):
		print('Error: Identity already created.', file=sys.stderr)
		sys.exit(1)

	persona = Persona()
	persona.generate_key()

	persona.new_version(*(x.split('=', 1) for x in options.tagvalue))

	persona.store(identfname)

def cmd_ident(options):
	identfname, _, _ = _get_paths(options)

	persona = Persona.load(identfname)

	if options.tagvalue:
		persona.new_version(*(x.split('=', 1) for x in
		    options.tagvalue))

		persona.store(identfname)
	else:
		ident = persona.get_identity()
		for k, v in _iterdictlist(ident, skipcommon=False):
			print('%s:\t%s' % (k, v))

def cmd_pubkey(options):
	identfname, _, _ = _get_paths(options)

	persona = Persona.load(identfname)

	print(persona.get_pubkey())

@init_datastructs
def cmd_modify(options, persona, objstr, cache):
	# because of how argparse works, only one file will be collected
	# multiple files will end up in modtagvalues, so we need to
	# find and move them.

	for idx, i in enumerate(options.modtagvalues):
		if i[0] not in { '+', '-' }:
			# move remaining files
			options.files[0:0] = options.modtagvalues[idx:]
			del options.modtagvalues[idx:]
			break

	props = [[ x[0] ] + x[1:].split('=', 1) for x in options.modtagvalues]
	if any(x[0] not in ('+', '-') for x in props):
		print('ERROR: tag needs to start with a "+" (add) or a "-" (remove).', file=sys.stderr)
		sys.exit(1)

	badtags = list(x[1] for x in props if x[1] in (MDBase._common_names |
	    MDBase._common_optional))
	if any(badtags):
		print('ERROR: invalid tag%s: %s.' % ( 's' if
		    len(badtags) > 1 else '', repr(badtags)), file=sys.stderr)
		sys.exit(1)

	adds = [ x[1:] for x in props if x[0] == '+' ]

	if any((len(x) != 2 for x in adds)):
		print('ERROR: invalid tag, needs an "=".', file=sys.stderr)
		sys.exit(1)

	dels = [ x[1:] for x in props if x[0] == '-' ]

	for i in options.files:
		#print('a:', repr(i), file=_real_stderr)

		try:
			objs = objstr.get_metadata(i, persona)
			#print('d:', repr(i), repr(objs), file=_real_stderr)
		except FileNotFoundError:
			print('ERROR: file not found: %s, or invalid tag specification.' % repr(i), file=sys.stderr)
			sys.exit(1)


		for j in objs:
			#print('c:', repr(j), file=_real_stderr)
			# make into key/values
			# copy as we modify it later, which is bad
			obj = j.__to_dict__().copy()

			# delete tags
			for k in dels:
				try:
					key, v = k
				except ValueError:
					try:
						del obj[k[0]]
					except KeyError:
						pass
				else:
					obj[key].remove(v)

			# add tags
			uniqify = set()
			for k, v in adds:
				obj.setdefault(k, []).append(v)
				if k in j._uniq_properties:
					uniqify.add(k)

			for k in uniqify:
				obj[k] = list(set(obj[k]))

			#print('a:', repr(obj), file=_real_stderr)
			del obj['modified']
			nobj = MDBase.create_obj(obj)

			objstr.loadobj(nobj)

def printhost(host):
	print('%s\t%s' % (host.name, host.hostuuid))

@init_datastructs
def cmd_mapping(options, persona, objstr, cache):
	if options.mapping is None:
		maps = objstr.get_by_type('mapping')

		maps = (' <-> '.join(i['mapping']) for i in maps)

		for i in maps:
			print(i)
		return

	parts = [ x.split(':', 1) for x in options.mapping ]

	if len(parts[0]) == 1:
		parts[0] = [ str(hostuuid()), parts[0][0] ]

	if parts[0][0] == str(hostuuid()):
		parts[0][1] = str(pathlib.Path(parts[0][1]).resolve())

	if parts[1][1][0] != '/':
		print('ERROR: host path must be absolute, is %s.' %
		    repr(parts[1][1][0]), file=sys.stderr)
		sys.exit(1)

	try:
		[ objstr.get_host(x[0]) for x in parts ]
	except KeyError as e:
		print('ERROR: Unable to find host %s' %
		    str(e.args[0]), file=sys.stderr)
		sys.exit(1)

	m = persona.Mapping(mapping=[ ':'.join(x) for x in parts ])

	objstr.loadobj(m)

@init_datastructs
def cmd_hosts(options, persona, objstr, cache):
	selfuuid = hostuuid()

	try:
		host = objstr.get_host(selfuuid)
	except KeyError:
		host = persona.Host(name=socket.gethostname(), hostuuid=selfuuid)
		objstr.loadobj(host)

	printhost(host)

	hosts = objstr.get_hosts()

	for i in hosts:
		if i == host:
			continue

		printhost(i)

def genstartstop(cnt, idx):
	idx = min(idx, cnt - 10)
	idx = max(0, idx)

	maxstart = max(0, cnt - 20)

	startidx = min(max(0, idx - 10), maxstart)
	endidx = min(cnt, startidx + 20)

	return startidx, endidx

def getnextfile(files, idx):
	# original data incase of abort
	origfiles = files
	origidx = idx

	# current selection (last file or dir)
	curselidx = origidx

	currentcnt = None

	while True:
		if len(files) != currentcnt:
			currentcnt = len(files)
			maxidx = max(0, currentcnt - 10)
			idx = min(maxidx, idx)

		startidx, endidx = genstartstop(currentcnt, idx)
		subset = files[startidx:endidx]
		selfile = -1 if curselidx < startidx or curselidx >= startidx + \
		    20 else curselidx - startidx
		print('%2d) Parent directory' % 0)
		for i, f in enumerate(subset):
			print('%2d)%1s%s%s' % (i + 1, '*' if i == selfile else '', repr(str(f)), '/' if f.is_dir() else ''))

		print('P) Previous page')
		print('N) Next page')
		print('A) Abort')
		print('Selection:')
		inp = sys.stdin.readline().strip()

		if inp.lower() == 'p':
			idx = max(0, idx - 19)
			continue
		if inp.lower() == 'n':
			idx = min(currentcnt - 1, idx + 19)
			continue
		if inp.lower() == 'a':
			return origfiles, origidx

		try:
			inp = int(inp)
		except ValueError:
			print('Invalid selection.')
			continue

		if inp == 0:
			curdir = files[idx].parent
			files = sorted(files[idx].parent.parent.iterdir())
			idx = curselidx = files.index(curdir)
			continue
		if inp < 1 or inp > len(subset):
			print('Invalid selection.')
			continue

		newidx = startidx - 1 + inp

		if files[newidx].is_dir():
			files = sorted(files[newidx].iterdir())
			curselidx = idx = 0
			continue

		return files, newidx

def checkforfile(objstr, curfile, ask=False):
	try:
		fobj = objstr.by_file(curfile, ('file',))
	except (ValueError, KeyError):
		if not ask:
			return

		while True:
			print('file unknown, hash(y/n)?')
			inp = sys.stdin.readline().strip().lower()
			if inp == 'n':
				return
			if inp == 'y':
				break

		try:
			fobj = persona.by_file(curfile)
		except (FileNotFoundError, KeyError) as e:
			print('ERROR: file not found: %s' % repr(curfile), file=sys.stderr)
			return
		else:
			objstr.loadobj(fobj)

	return fobj

from contextlib import contextmanager
import time

@contextmanager
def timeblock(tag):
	s = time.time()
	try:
		yield
	finally:
		e = time.time()
		_debprint(tag, e - s)

@init_datastructs
def cmd_interactive(options, persona, objstr, cache):
	files = [ pathlib.Path(x) for x in options.files ]

	cache.count = 15

	autoskip = True

	idx = 0
	if not files:
		files = sorted(pathlib.Path('.').iterdir())

	while True:
		with timeblock('checkfile'):
			curfile = files[idx]

			fobj = checkforfile(objstr, curfile, not autoskip)
			#_debprint(repr(fobj))

			if fobj is None and autoskip and idx > 0 and idx < len(files) - 1:
				# if we are auto skipping, and within range, continue
				if inp == '1':
					idx = max(0, idx - 1)
					continue
				if inp == '2':
					idx = min(len(files) - 1, idx + 1)
					continue

		print('Current: %s' % repr(str(curfile)))

		with timeblock('byfile'):
			if fobj is None:
				print('No file object for this file.')
			else:
				try:
					objs = objstr.by_file(curfile)
				except KeyError:
					print('No tags or metadata object for this file.')
				else:
					for k, v in _iterdictlist(objs[0]):
						if k in { 'sig', 'hashes' }:
							continue
						print('%s:\t%s' % (k, v))

		if idx == 0:
			print('1) No previous file')
		else:
			print('1) Previous: %s' % repr(str(files[idx - 1])))

		if idx + 1 == len(files):
			print('2) No next file')
		else:
			print('2) Next: %s' % repr(str(files[idx + 1])))

		print('3) List files')
		print('4) Browse directory of file.')
		print('5) Browse original list of files.')
		print('6) Add new tag.')
		print('7) Open file.')
		print('8) Turn auto skip %s' % 'off' if autoskip else 'on')

		tags = cache.tags()

		for pos, (tag, value) in enumerate(tags):
			print('%s) %s=%s' % (string.ascii_lowercase[pos], tag, value))

		print('Q) Quit')

		print('Select option: ')

		inp = sys.stdin.readline().strip()

		if not inp:
			continue
		elif inp == '1':
			idx = max(0, idx - 1)
			continue
		elif inp == '2':
			idx = min(len(files) - 1, idx + 1)
			continue
		elif inp == '3':
			files, idx = getnextfile(files, idx)
			continue
		elif inp == '4':
			files = sorted(curfile.parent.iterdir())
			try:
				idx = files.index(curfile)
			except ValueError:
				print('WARNING: File no longer present.')
				idx = 0
			continue
		elif inp == '5':
			files = [ pathlib.Path(x) for x in options.files ]
			try:
				idx = files.index(curfile)
			except ValueError:
				print('WARNING: File not present.')
				idx = 0
			continue
		elif inp == '6':
			print('Tag?')
			try:
				tag, value = sys.stdin.readline().strip().split('=', 1)
			except ValueError:
				print('Invalid tag, no "=".')
			else:
				cache.add((tag, value))
				metadata = objstr.get_metadata(curfile, persona)[0]

				metadata = metadata.new_version((tag, value))

				objstr.loadobj(metadata)

			continue
		elif inp == '7':
			subprocess.run(('open', curfile))
			continue

		elif inp.lower() == 'q':
			break

		try:
			i = string.ascii_lowercase.index(inp.lower())
			cache.add(tags[i])
		except (ValueError, IndexError):
			pass
		else:
			metadata = objstr.get_metadata(curfile, persona)[0]

			metadata = metadata.new_version(tags[i])

			objstr.loadobj(metadata)

			continue

		print('Invalid selection.')

@init_datastructs
def cmd_dump(options, persona, objstr, cache):
	if options.dump_uuids or options.dump_hashes:
		for i in options.dump_uuids:
			print(objstr.by_id(i).encode('json'))

		for i in options.dump_hashes:
			for j in objstr.by_hash(i):
				print(j.encode('json'))

		return

	print(persona.get_identity().encode('json'))

	for i in objstr:
		print(i.encode('json'))

def cmd_auto(options):
	for i in options.files[:]:
		mf = detect_from_filename(i)

		primary = mf[0].split('/', 1)[0]
		mt = mf[0]
		if primary == 'text':
			mt += '; charset=%s' % mf[1]

		print('Set:')
		print('\tmimetype:\t%s' % mt)
		print()
		print('Apply (y/N)?')

		inp = sys.stdin.readline()

		if inp.strip().lower() in ('y', 'yes'):
			options.modtagvalues = [ '-mimetype', '+mimetype=%s' % mt ]
			options.files = [ i ]
			cmd_modify(options)

@init_datastructs
def cmd_list(options, persona, objstr, cache):
	exit = 0
	for i in options.files:
		try:
			objs = objstr.by_file(i)
		except (ValueError, KeyError):
			# create the file, it may have the same hash
			# as something else
			try:
				fobj = persona.by_file(i)
				objstr.loadobj(fobj)

				objs = objstr.by_file(i)
			except (FileNotFoundError, KeyError) as e:
				if options.empty and options.json:
					print('{}')
					continue

				print('ERROR: file not found: %s' % repr(i), file=sys.stderr)
				exit = 1
				continue

		for j in objstr.by_file(i):
			if options.json:
				print(j.encode('json'))
			else:
				for k, v in _iterdictlist(j):
					print('%s:\t%s' % (k, v))

	if exit:
		sys.exit(exit)

def handle_bittorrent(fname, persona, objstr):
	with open(fname, 'rb') as fp:
		torrent = bencode.bdecode(fp.read())
		bencodedinfo = bencode.bencode(torrent['info'])
		infohash = hashlib.sha1(bencodedinfo).hexdigest()

	# XXX - not entirely happy w/ URI
	uri = 'magnet:?xt=urn:btih:%s&dn=%s' % (infohash,
	    torrent['info']['name'].decode('utf-8'))

	try:
		cont = objstr.by_id(Container.make_id(uri))
	except KeyError:
		pass
	else:
		if 'incomplete' not in cont:
			print('Warning, container already complete, skipping %s.' % repr(fname), file=sys.stderr)
			return

	good, bad = validate_file(fname)

	if bad:
		print('Warning, incomple/invalid files, not added for %s:' %
		    repr(fname), file=sys.stderr)
		print('\n'.join('\t%s' %
		    repr(str(pathlib.Path(*x.parts[1:]))) for x in
		    sorted(bad)), file=sys.stderr)

	files = []
	hashes = []
	for j in sorted(good):
		files.append(str(pathlib.PosixPath(*j.parts[1:])))
		try:
			fobj = objstr.by_file(j, ('file',))[0]
		except:
			fobj = persona.by_file(j)
			objstr.loadobj(fobj)

		# XXX - ensure only one is added?
		hashes.extend(fobj.hashes)

	kwargs = dict(files=files, hashes=hashes,
		    uri=uri)

	if bad:
		kwargs['incomplete'] = True

	# XXX - doesn't combine files/hashes, that is if a
	# Container has one set of good files, and then the
	# next scan has a different set, only the second set
	# will be present, not any from the first set.

	try:
		cont = objstr.by_id(Container.make_id(uri))
		cont = cont.new_version(dels=() if bad
		    else ('incomplete',), replaces=kwargs.items())
	except KeyError:
		cont = persona.Container(**kwargs)

	objstr.loadobj(cont)

def handle_archive(fname, persona, objstr):
	with libarchive.Archive(fname) as arch:
		files = []
		hashes = []

		for i in arch:
			if not i.isfile():
				continue

			files.append(i.pathname)

			with arch.readstream(i.size) as fp:
				hashes.append(_hashfp(fp))

	try:
		fobj = objstr.by_file(fname, ('file',))[0]
	except:
		fobj = persona.by_file(fname)
		objstr.loadobj(fobj)

	uri = _makehashuri(fobj.hashes[0])

	kwargs = dict(files=files, hashes=hashes,
		    uri=uri)
	try:
		cont = objstr.by_id(Container.make_id(uri))
		# XXX - only update when different, check uri
		cont = cont.new_version(replaces=kwargs.items())
	except KeyError:
		cont = persona.Container(**kwargs)

	objstr.loadobj(cont)

_container_mapping = {
	'application/x-bittorrent': handle_bittorrent,
	'application/x-tar': handle_archive,
	'application/zip': handle_archive,
}

@init_datastructs
def cmd_container(options, persona, objstr, cache):
	for i in options.files:
		mf = detect_from_filename(i)
		#_debprint('mf:', repr(mf))
		fun = _container_mapping[mf.mime_type]

		fun(i, persona, objstr)

def _json_objstream(fp):
	inp = fp.read()

	jd = json.JSONDecoder()

	while inp:
		inp = inp.strip()
		jobj, endpos = jd.raw_decode(inp)

		yield jobj

		inp = inp[endpos:]

@init_datastructs
def cmd_import(options, persona, objstr, cache):
	for jobj in _json_objstream(sys.stdin):
		if options.sign:
			cbr = _makeuuid(jobj['created_by_ref'])
			if cbr != persona.uuid:
				# new owner
				jobj['created_by_ref'] = persona.uuid

				# drop old parts
				jobj.pop('uuid', None)
				jobj.pop('modified', None)

		obj = MDBase.create_obj(jobj)

		if options.sign:
			obj = persona.sign(obj)

		objstr.loadobj(obj)

@init_datastructs
def cmd_drop(options, persona, objstr, cache):
	for i in options.uuids:
		objstr.drop_uuid(i)

@init_datastructs
def cmd_search(options, persona, objstr, cache):
	args = options.args.copy()

	_type = args.pop(0)

	if _type == 'tags':
		propmap = aliased(orm.PropertyMapping)
		skeymap = aliased(orm.StringTable)
		svaluemap = aliased(orm.StringTable)

		q = select(skeymap.str, svaluemap.str).where(propmap.keyid == skeymap.id).where(skeymap.str == args[0]).where(propmap.valueid == svaluemap.id).distinct().order_by(svaluemap.str)
		_debprint(repr(q))
		with objstr._ses() as session:
			for k, v in session.execute(q):
				print('%s=%s' % (k, v))
		return

	if _type != 'file':
		print('unknown search type: %s' % repr(_type),
		    file=sys.stderr)
		sys.exit(1)

	searches = [ (x[0], ) + tuple(x[1:].split('=', 1)) for x in args ]
	#_debprint(repr(searches))

	if not searches:
		# just dump everything
		mdofile = aliased(orm.MetaDataObject)

		sel = select(mdofile.data).where(
			# we are operating on files
			mdofile.type == 'file')

		sel = sel.execution_options(yield_per=10)

		with objstr._ses() as session:
			r = ( x[0] for x in session.execute(sel) )

			if _type == 'file':
				mapping = FileObject.prep_mapping(
				    objstr.get_hostmappings())
				r = ( x.get_lcl_name(mapping) for x in r )
			else:
				raise ValueError('unhandled type: %s' % repr(_type))

			for i in r:
				print(i)

		return

	propmap = aliased(orm.PropertyMapping)

	# propobj only returns what can match query
	propobj = select(propmap.obj).distinct()

	onlyexclusions = True

	for i in searches:
		propmapsub = aliased(orm.PropertyMapping)
		skeymap = aliased(orm.StringTable)
		svaluemap = aliased(orm.StringTable)

		origvalue = '='.join(i[1:])

		try:
			op, key, value = i
		except ValueError:
			op, key = i
			value = None

		# handle meta tree
		if key.startswith('meta:'):
			mat = _metaopre.match(origvalue)
			if not mat:
				print('invalid meta: specification: %s' % repr(origvalue),
				    file=sys.stderr)
				sys.exit(1)

			#print(repr(mat), repr(mat.groups()), file=_real_stderr)

			mdo = aliased(orm.MetaDataObject)
			metavalue = _metavalueconv[mat.group('key')](mat.group('val'))
			subq = select(propmapsub.obj).where(propmapsub.obj == mdo.uuid,
			    _metaoplookup[mat.group('op')](mdo.modified, metavalue))

		else:
			subq = select(propmapsub.obj).where(
			    # that match the key
			    propmapsub.keyid == skeymap.id, skeymap.str == key)

			if value is not None:
				subq = subq.where(propmapsub.valueid == svaluemap.id,
				    svaluemap.str == value)

		#subq = subq.subquery()

		if op == '+':
			onlyexclusions = False
			propobj = propobj.where(propmap.obj.in_(subq))
		elif op == '-':
			propobj = propobj.where(propmap.obj.notin_(subq))
		else:
			raise ValueError('unhandled op: %s' % repr(op))

	# propobj should have all the ones we need selected, map back to
	# the object we need

	# base object (file)
	mdofile = aliased(orm.MetaDataObject)
	# hashes of base object
	htfile = aliased(orm.HashTable)

	# hashes to metadata objects
	htmd = aliased(orm.HashTable)
	# metadataobjects
	mdomd = aliased(orm.MetaDataObject)

	# Don't know if distinct is needed/warrented for the
	# in_ sub queries
	sel = select(mdofile.data).where(
		# we are operating on files
		mdofile.type == 'file',
		# we get all the hashes for the files
		mdofile.uuid.in_(
			select(htfile.uuid).where(
				htfile.hash == htmd.hash,
				# we get all the hashes for selected metadata
				htmd.uuid.in_(
					select(mdomd.uuid).where(
						mdomd.type == 'metadata',
						mdomd.uuid.in_(propobj)
					)
				)
			)
		)
	)

	if onlyexclusions:
		# add in all the files that doesn't have metadata

		# base object (file)
		mdofile = aliased(orm.MetaDataObject)
		# hashes of base object
		htfile = aliased(orm.HashTable)

		# hashes to metadata objects
		htmd = aliased(orm.HashTable)
		# metadataobjects
		mdomd = aliased(orm.MetaDataObject)

		# query for hashes w/o metadata
		selwomd = select(mdofile.data).where(
			# we are operating on files
			mdofile.type == 'file',
			# we get all the hashes for the files
			mdofile.uuid == htfile.uuid,
			htfile.hash.notin_(
				select(htmd.hash).where(
					htmd.uuid == mdomd.uuid, mdomd.type == 'metadata'
				)
			)
		)

		sel = sel.union(selwomd)

	sel = sel.execution_options(yield_per=10)

	#_debprint('sel:', _getquery(sel, objstr))
	#print('sel:', _getquery(sel, objstr), file=_real_stderr)

	with objstr._ses() as session:
		r = ( x[0] for x in session.execute(sel) )

		if _type == 'file':
			mapping = FileObject.prep_mapping(
			    objstr.get_hostmappings())
			r = ( x.get_lcl_name(mapping) for x in r )
		else:
			raise ValueError('unhandled type: %s' % repr(_type))

		for i in r:
			print(i)

def main():
	import argparse

	parser = argparse.ArgumentParser()

	parser.add_argument('--debug', action='store_true',
	    default=False, help='enable debug output')

	parser.add_argument('--db', '-d', type=str,
	    help='base name for storage')

	subparsers = parser.add_subparsers(title='subcommands',
	    description='valid subcommands', help='additional help')

	parser_help = subparsers.add_parser('help', help='get help')
	parser_help.set_defaults(func=lambda *args: (parser.print_help(), sys.exit(0)))

	parser_gi = subparsers.add_parser('genident', help='generate identity')
	parser_gi.add_argument('tagvalue', nargs='+',
	    help='add the arg as metadata for the identity, tag=[value]')
	parser_gi.set_defaults(func=cmd_genident)

	parser_i = subparsers.add_parser('ident', help='update identity')
	parser_i.add_argument('tagvalue', nargs='*',
	    help='add the arg as metadata for the identity, tag=[value]')
	parser_i.set_defaults(func=cmd_ident)

	parser_pubkey = subparsers.add_parser('pubkey',
	    help='print public key of identity')
	parser_pubkey.set_defaults(func=cmd_pubkey)

	# used so that - isn't treated as an option
	parser_mod = subparsers.add_parser('modify',
	    help='modify tags on file(s)', prefix_chars='@')
	parser_mod.add_argument('modtagvalues', nargs='+',
	    help='add (+) or delete (-) the tag=[value], for the specified files')
	parser_mod.add_argument('files', nargs='+',
	    help='files to modify')
	parser_mod.set_defaults(func=cmd_modify)

	parser_auto = subparsers.add_parser('auto',
	    help='automatic detection of file properties')
	parser_auto.add_argument('files', nargs='+',
	    help='files to modify')
	parser_auto.set_defaults(func=cmd_auto)

	parser_list = subparsers.add_parser('list', help='list tags on file(s)')
	parser_list.add_argument('--json', action='store_true',
	    help='output data as series of JSON objects')
	parser_list.add_argument('--empty', action='store_true',
	    help='output an empty JSON object if file not found or not metadata')
	parser_list.add_argument('files', nargs='+',
	    help='files to modify')
	parser_list.set_defaults(func=cmd_list)

	parser_container = subparsers.add_parser('container',
	    help='file is examined as a container and the internal files imported as entries')
	parser_container.add_argument('files', nargs='+',
	    help='files to modify')
	parser_container.set_defaults(func=cmd_container)

	parser_hosts = subparsers.add_parser('hosts',
	    help='dump all the hosts, self is always first')
	parser_hosts.set_defaults(func=cmd_hosts)

	parser_mapping = subparsers.add_parser('mapping',
	    help='list mappings, or create a mapping')
	parser_mapping.add_argument('--create', dest='mapping', nargs=2,
	    help='mapping to add, host|hostuuid:path host|hostuuid:path')
	parser_mapping.set_defaults(func=cmd_mapping)

	parser_interactive = subparsers.add_parser('interactive',
	    help='start in interactive mode')
	parser_interactive.add_argument('files', nargs='*',
	    help='files to work with')
	parser_interactive.set_defaults(func=cmd_interactive)

	parser_dump = subparsers.add_parser('dump', help='dump all the objects')
	parser_dump.add_argument('--uuid', dest='dump_uuids', action='append',
	    default=[],
	    help='dump the object with the specified UUID')
	parser_dump.add_argument('--hash', dest='dump_hashes', action='append',
	    default=[],
	    help='dump the object(s) associated w/ the specified hash')
	parser_dump.set_defaults(func=cmd_dump)

	parser_import = subparsers.add_parser('import',
	    help='import objects encoded as json')
	parser_import.add_argument('--sign', action='store_true',
	    help='import as new identity, and sign objects (if created_by_ref is different, new uuid is created)')
	parser_import.set_defaults(func=cmd_import)

	parser_drop = subparsers.add_parser('drop',
	    help='drop the object specified by UUID')
	parser_drop.add_argument('uuids', nargs='+',
	    help='UUID of object to drop')
	parser_drop.set_defaults(func=cmd_drop)

	parser_search = subparsers.add_parser('search',
	    help='find objects', prefix_chars='@')
	parser_search.add_argument('args', nargs='+',
	    help='args')
	parser_search.set_defaults(func=cmd_search)

	options = parser.parse_args()

	if options.debug:
		enable_debug()
	else:
		disable_debug()

	try:
		fun = options.func
	except AttributeError:
		parser.print_help()
		sys.exit(0)

	fun(options)

if __name__ == '__main__':	# pragma: no cover
	main()

class _TestCononicalCoder(unittest.TestCase):
	def test_con(self):
		# make a dict
		obja = {
		    'foo': 23984732, 'a': 5, 'b': 6,
		    'something': '2398472398723498273dfasdfjlaksdfj'
		}

		# reorder the items in it
		objaitems = list(obja.items())
		objaitems.sort()
		objb = dict(objaitems)

		# and they are still the same
		self.assertEqual(obja, objb)

		# This is to make sure that item order changed
		self.assertNotEqual(list(obja.items()), list(objb.items()))

		astr = pasn1.dumps(obja)
		bstr = pasn1.dumps(objb)

		# that they normally will be serialized differently
		self.assertNotEqual(astr, bstr)

		# but w/ the special encoder
		astr = _asn1coder.dumps(obja)
		bstr = _asn1coder.dumps(objb)

		# they are now encoded the same
		self.assertEqual(astr, bstr)

class _TestMigrations(unittest.TestCase):
	def setUp(self):
		self._engine = create_engine('sqlite+pysqlite:///:memory:',
		    echo=_sql_verbose, future=True)

	def test_f2131(self):
		# That an object store generated at the start
		objstr = ObjectStore(self._engine, 'afad01589b76')

		# and a host objects
		hostobj = Host(created_by_ref=uuid.uuid4(), hostuuid=uuid.uuid4())

		# build table metadata from original db
		mdo = sqlalchemy.schema.MetaData()
		mdobjstable = sqlalchemy.Table('metadata_objects', mdo, autoload_with=self._engine)

		with objstr._ses() as session:
			stmt = insert(mdobjstable).values(
			    uuid=hostobj.uuid.hex, modified=hostobj.modified,
			    data=hostobj.encode())
			session.execute(stmt)

			session.commit()

		# migrate the database forward
		objstr._handle_migration('head')

		# make sure we can query it
		self.assertEqual(list(objstr.get_hosts()), [ hostobj ])
		self.assertEqual(list(objstr), [ hostobj ])
		self.assertEqual(list(objstr.get_by_type('file')), [ ])
		self.assertEqual(list(objstr.get_by_type(FileObject)), [ ])
		self.assertEqual(list(objstr.get_by_type(Host)), [ hostobj ])

		#with objstr._ses() as session:
		#	for i in session.query(orm.MetaDataObject).all():
		#		_debprint('c:', repr(i))

	def test_dff0d(self):
		# That an object store generated at the start
		objstr = ObjectStore(self._engine, 'dff0d9ed0be1')

		pers = Persona()
		pers.generate_key()
		objstr.loadobj(pers.get_identity())

		obj = pers.MetaData({ 'other': 'baz'})

		# That has a metadata object
		objstr.loadobj(obj)

		# migrate the database forward
		objstr._handle_migration('head')

		with objstr._ses() as session:
			# that string table has entries
			other = session.execute(select(orm.StringTable.id)
			    .where(orm.StringTable.str == 'other')).first()[0]
			baz = session.execute(select(orm.StringTable.id)
			    .where(orm.StringTable.str == 'baz')).first()[0]

			# that propertymapping was populated
			pm = { (x.obj, x.keyid, x.valueid) for (x,) in
			    session.execute(select(orm.PropertyMapping)) }
			self.assertEqual(pm, { (obj.uuid, other, baz) })

class _TestCases(unittest.TestCase):
	def setUp(self):
		self.fixtures = pathlib.Path('fixtures').resolve()

		d = pathlib.Path(tempfile.mkdtemp()).resolve()
		self.basetempdir = d
		self.tempdir = d / 'subdir'

		self.persona = Persona.load(os.path.join('fixtures',
		    'sample.persona.pasn1'))
		self.created_by_ref = self.persona.get_identity().uuid

		shutil.copytree(self.fixtures / 'testfiles', self.tempdir)
		shutil.copy(self.fixtures / 'sample.data.sqlite3', self.tempdir)

		self.oldcwd = os.getcwd()

	def tearDown(self):
		shutil.rmtree(self.basetempdir)
		self.tempdir = None

		os.chdir(self.oldcwd)

	def test_genstartstop(self):
		self.assertEqual(genstartstop(5, 0), (0, 5))
		self.assertEqual(genstartstop(5, 1), (0, 5))
		self.assertEqual(genstartstop(5, 4), (0, 5))
		self.assertEqual(genstartstop(25, 1), (0, 20))
		self.assertEqual(genstartstop(25, 20), (5, 25))
		self.assertEqual(genstartstop(25, 24), (5, 25))
		self.assertEqual(genstartstop(124, 1), (0, 20))
		self.assertEqual(genstartstop(124, 53), (43, 63))
		self.assertEqual(genstartstop(124, 120), (104, 124))
		self.assertEqual(genstartstop(124, 124), (104, 124))

	def test_fileobject(self):
		os.chdir(self.tempdir)

		engine = create_engine(
		    "sqlite+pysqlite:///memdb1?mode=memory",
		    echo=_sql_verbose, future=True)

		objst = ObjectStore(engine)

		a = self.persona.by_file('test.txt')

		# that the dir is absolute
		self.assertEqual(a.dir[0], '/')

		# make sure the file's hostid is a UUID
		self.assertIsInstance(a.hostid, uuid.UUID)

		# make sure the file's id is a UUID
		self.assertIsInstance(a.id, uuid.UUID)

		objst.loadobj(a)

		#_debprint('a:', repr(a))
		#_debprint('by_id:', objst.by_id(a.uuid))

		# write out the store
		objst.store('teststore.pasn1')

		# load it back in
		objstr = ObjectStore(engine)

		a = objstr.by_id(a['uuid'])

		# make sure the hostid is still a UUID
		self.assertIsInstance(a.hostid, uuid.UUID)

		# make sure the file's id is still a UUID
		self.assertIsInstance(a.id, uuid.UUID)

		# That it can be encoded to json
		jsfo = a.encode('json')

		# that it can be decoded from json
		jsloadedfo = MDBase.decode(jsfo, 'json')

		# and that it is equal
		self.assertEqual(jsloadedfo, a)

	def test_mdbase(self):
		self.assertRaises(ValueError, MDBase, created_by_ref='')
		self.assertRaises(ValueError, MDBase.create_obj,
		    { 'type': 'unknosldkfj' })
		self.assertRaises(ValueError, MDBase.create_obj,
		    { 'type': 'metadata' })

		baseobj = {
			'type': 'metadata',
			'created_by_ref': self.created_by_ref,
		}
		origbase = copy.deepcopy(baseobj)

		# that when an MDBase object is created
		md = MDBase.create_obj(baseobj)

		# it doesn't modify the passed in object (when adding
		# generated properties)
		self.assertEqual(baseobj, origbase)

		# and it has the generted properties
		# Note: cannot mock the functions as they are already
		# referenced at creation time
		self.assertIn('uuid', md)
		self.assertIn('modified', md)

		# That you can create a new version using new_version
		md2 = md.new_version(('dc:creator', 'Jim Bob',))

		# that they are different
		self.assertNotEqual(md, md2)

		# and that the new modified time is different from the old
		self.assertNotEqual(md.modified, md2.modified)

		# and that the modification is present
		self.assertEqual(md2['dc:creator'], [ 'Jim Bob' ])

		# that providing a value from common property
		fvalue = b'fakesig'
		md3 = md.new_version(('sig', fvalue))

		# gets set directly, and is not a list
		self.assertEqual(md3.sig, fvalue)

		# that invalid attribute access raises correct exception
		self.assertRaises(AttributeError, getattr, md,
		    'somerandombogusattribute')

		# that when readding an attribute that already exists
		md3 = md2.new_version(('dc:creator', 'Jim Bob',))

		# that only one exists
		self.assertEqual(md3['dc:creator'], [ 'Jim Bob' ])

	def test_mdbase_encode_decode(self):
		# that an object
		baseobj = {
			'type': 'metadata',
			'created_by_ref': self.created_by_ref,
		}
		obj = MDBase.create_obj(baseobj)

		# can be encoded
		coded = obj.encode()

		# and that the rsults can be decoded
		decobj = MDBase.decode(coded)

		# and that they are equal
		self.assertEqual(obj, decobj)

		# and in the encoded object
		eobj = _asn1coder.loads(coded)

		# the uuid property is a str instance
		self.assertIsInstance(eobj['uuid'], bytes)

		# and has the length of 16
		self.assertEqual(len(eobj['uuid']), 16)

		# and that json can be used to encode
		js = obj.encode('json')

		# and that it is valid json
		jsobj = json.loads(js)

		# and that it can be decoded
		jsdecobj = MDBase.decode(js, 'json')

		# and that it matches
		self.assertEqual(jsdecobj, obj)

		for key, inval in [
			('modified', '2022-08-19T01:27:34.258676'),
			('modified', '2022-08-19T01:27:34Z'),
			('modified', '2022-08-19T01:27:34.258676+00:00'),
			('uuid', 'z5336176-8086-4c21-984f-fda60ddaa172'),
			('uuid', '05336176-8086-421-984f-fda60ddaa172'),
		    ]:
			jsobj['modified'] = inval
			jstest = json.dumps(jsobj)
			self.assertRaises(ValueError, MDBase.decode, jstest, 'json')

	def test_mdbase_wrong_type(self):
		# that created_by_ref can be passed by kw
		obj = MetaData(created_by_ref=self.created_by_ref)

		self.assertRaises(ValueError, FileObject, dict(obj.items(False)))

	def test_makehash(self):
		self.assertRaises(ValueError, ObjectStore.makehash, 'slkj')
		self.assertRaises(ValueError, ObjectStore.makehash, 'sha256:91751cee0a1ab8414400238a761411daa29643ab4b8243e9a91649e25be53ADA')
		self.assertRaises(ValueError, ObjectStore.makehash, 'bogushash:9e0a1ab8414400238a761411daa29643ab4b8243e9a91649e25be53ADA', strict=False)

		self.assertEqual(ObjectStore.makehash('cf83e1357eefb8bdf1542850d66d8007d620e4050b5715dc83f4a921d36ce9ce47d0d13c5d85f2b0ff8318d2877eec2f63b931bd47417a81a538327af927da3e', strict=False), 'sha512:cf83e1357eefb8bdf1542850d66d8007d620e4050b5715dc83f4a921d36ce9ce47d0d13c5d85f2b0ff8318d2877eec2f63b931bd47417a81a538327af927da3e')
		self.assertEqual(ObjectStore.makehash('e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855', strict=False), 'sha256:e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855')

	@staticmethod
	def statmock(fname):
		fname = pathlib.Path(fname)
		fnameparts = fname.parts
		subdiridx = fnameparts.index('subdir')

		_stats = {
			# repr on os.stat_result doesn't work
			#  (mode, ino, dev, nlink, uid, gid, size, atime, mtime, ctime)

			'f': os.stat_result((stat.S_IROTH|stat.S_IXOTH|stat.S_IRGRP|stat.S_IXGRP|stat.S_IXUSR|stat.S_IRUSR|stat.S_IFDIR, 10, 100, 2, 100, 100, 1024, 1654166365, 1558388856.000000, 1663133775)),
			'test.txt': os.stat_result((stat.S_IROTH|stat.S_IRGRP|stat.S_IWUSR|stat.S_IRUSR|stat.S_IFREG, 10, 100, 1, 100, 100, 15, 1654166365, 1558388856.000000, 1663133775)),
			'newfile.txt': os.stat_result((stat.S_IROTH|stat.S_IRGRP|stat.S_IWUSR|stat.S_IRUSR|stat.S_IFREG, 10, 100, 1, 100, 100, 19, 1659652579, 1658982768.041291, 1663133775)),
			'sample.data.sqlite3': os.stat_result((stat.S_IROTH|stat.S_IRGRP|stat.S_IWUSR|stat.S_IRUSR|stat.S_IFREG, 10, 100, 1, 100, 100, 57344, 1663133777, 1663133777.529757, 1663133777)),
			't': os.stat_result((stat.S_IFDIR, 0, 0, 0, 0, 0, 0, 0, 0, 0)),
			'z.jpg': os.stat_result((stat.S_IROTH|stat.S_IRGRP|stat.S_IWUSR|stat.S_IRUSR|stat.S_IFREG, 10, 100, 1, 100, 100, 332, 1661553878, 1661551130.361235, 1663134325)),

		}

		subpath = '/'.join(fnameparts[subdiridx + 1:])
		return _stats[subpath]

	@mock.patch('os.stat')
	def test_enumeratedir(self, statmock):
		statmock.side_effect = self.statmock

		files = enumeratedir(self.tempdir, self.created_by_ref)
		ftest = [ x for x in files if x.filename == 'test.txt' ][0]
		fname = 'test.txt'

		# make sure that they are of type MDBase
		self.assertIsInstance(ftest, MDBase)

		oldid = ftest.id
		self.assertEqual(ftest.filename, fname)
		self.assertEqual(ftest.dir, str(self.tempdir))
		# XXX - do we add host information?
		self.assertEqual(ftest.id, uuid.uuid5(_NAMESPACE_MEDASHARE_PATH,
		    str(hostuuid()) + '/'.join(os.path.split(self.tempdir) +
		    ( fname, ))))
		self.assertEqual(ftest.mtime, datetime.datetime(2019, 5, 20,
		    21, 47, 36, tzinfo=datetime.timezone.utc))
		self.assertEqual(ftest.size, 15)
		self.assertIn('sha512:7d5768d47b6bc27dc4fa7e9732cfa2de506ca262a2749cb108923e5dddffde842bbfee6cb8d692fb43aca0f12946c521cce2633887914ca1f96898478d10ad3f', ftest.hashes)

		# XXX - make sure works w/ relative dirs
		files = enumeratedir(os.path.relpath(self.tempdir),
		    self.created_by_ref)
		self.assertEqual(files[2].filename, 'test.txt')
		self.assertEqual(oldid, files[2].id)

	def test_mdbaseoverlay(self):
		engine = create_engine("sqlite+pysqlite:///:memory:", echo=_sql_verbose, future=True)
		objst = ObjectStore(engine)

		# that a base object
		bid = uuid.uuid4()
		objst.loadobj({
			'type': 'metadata',
			'uuid': bid,
			'modified': datetime.datetime(2019, 6, 10, 14, 3, 10),
			'created_by_ref': self.created_by_ref,
			'hashes': [ 'sha256:91751cee0a1ab8414400238a761411daa29643ab4b8243e9a91649e25be53ada' ],
			'someprop': [ 'somevalue' ],
			'lang': 'en',
		})

		# can have an overlay object
		oid = uuid.uuid4()
		dhash = 'sha256:a7c96262c21db9a06fd49e307d694fd95f624569f9b35bb3ffacd880440f9787'
		objst.loadobj({
			'type': 'metadata',
			'uuid': oid,
			'modified': datetime.datetime(2019, 6, 10, 18, 3, 10),
			'created_by_ref': self.created_by_ref,
			'hashes': [ dhash ],
			'parent_refs': [ bid ],
			'lang': 'en',
		})

		# and that when you get it's properties
		oobj = objst.by_id(oid)
		odict = dict(list(oobj.items()))

		# that is has the overlays property
		self.assertEqual(odict['parent_refs'], [ bid ])

		# that it doesn't have a common property
		self.assertNotIn('type', odict)

		# that when skipcommon is False
		odict = dict(oobj.items(False))

		# that it does have a common property
		self.assertIn('type', odict)

	def test_cryptography_persona(self):
		# Verify that a persona generated by cryptography still works
		persona = Persona.load(self.fixtures / 'cryptography.persona.pasn1')

		realpubkey = 'nFyLw6kB15DrM46ni9eEBRb6QD4rsPuco3ymj3mvz5YM8j3hY6chcjewU7FvqDpWALTSZ3E212SxCNErdYzPjgbxTnrYNyzeYTM2k58krEcKvWW6h'
		pubkey = persona.get_pubkey()

		self.assertEqual(realpubkey, pubkey)

		vpersona = Persona.from_pubkey(realpubkey)

		ident = persona.get_identity()
		vpersona.verify(ident)

		self.assertEqual(ident.uuid, uuid.UUID('52f1a92b-0c92-41e3-b647-356db89fb49c'))

	def test_persona(self):
		# that a newly created persona
		persona = Persona()

		# has an identity object
		idobj = persona.get_identity()

		# and that it has a uuid attribute that matches
		self.assertEqual(persona.uuid, idobj['uuid'])

		# that a key can be generated
		persona.generate_key()

		# that the pubkey property is present
		idobj = persona.get_identity()
		self.assertIsInstance(idobj['pubkey'], bytes)

		# that get_pubkey returns the correct thing
		pubstr = _asn1coder.dumps([ idobj.uuid, idobj['pubkey'] ])
		self.assertEqual(persona.get_pubkey(),
		    base58.b58encode_check(pubstr).decode('ascii'))

		# and that there is a signature
		self.assertIsInstance(idobj['sig'], bytes)

		# and that it can verify itself
		persona.verify(idobj)

		# and that a new persona can be created from the pubkey
		pkpersona = Persona.from_pubkey(persona.get_pubkey())

		# and that it can verify the old identity
		self.assertTrue(pkpersona.verify(idobj))

		# that a second time, it raises an exception
		self.assertRaises(RuntimeError, persona.generate_key)

		# that a file object created by it
		testfname = os.path.join(self.tempdir, 'test.txt')
		testobj = persona.by_file(testfname)

		# has the correct created_by_ref
		self.assertEqual(testobj.created_by_ref, idobj.uuid)

		self.assertEqual(testobj.type, 'file')

		# and has a signature
		self.assertIn('sig', testobj)

		# that a persona created from the identity object
		vpersona = Persona(idobj)

		# can verify the sig
		self.assertTrue(vpersona.verify(testobj))

		# and that a bogus signature
		bogussig = 'somebogussig'
		bogusobj = MDBase.create_obj(testobj)
		bogusobj.sig = bogussig

		# fails to verify
		self.assertRaises(Exception, vpersona.verify, bogusobj)

		# and that a modified object
		otherobj = testobj.new_version(('customprop', 'value'))

		# fails to verify
		self.assertRaises(Exception, vpersona.verify, otherobj)

		# that a persona object can be written
		perpath = os.path.join(self.basetempdir, 'persona.pasn1')
		persona.store(perpath)

		# and that when loaded back
		loadpersona = Persona.load(perpath)

		# the new persona object can sign an object
		nvtestobj = loadpersona.sign(testobj.new_version())

		# and the old persona can verify it.
		self.assertTrue(vpersona.verify(nvtestobj))

	def test_persona_metadata(self):
		# that a persona
		persona = Persona()
		persona.generate_key()

		# can create a metadata object
		hashobj = ['asdlfkj']
		mdobj = persona.MetaData(hashes=hashobj)

		# that the object has the correct created_by_ref
		self.assertEqual(mdobj.created_by_ref, persona.uuid)

		# and has the provided hashes
		self.assertEqual(mdobj.hashes, hashobj)

		# and that it can be verified
		persona.verify(mdobj)

	def test_objectstore(self):
		persona = self.persona
		objst = ObjectStore.load(self.tempdir / 'sample.data.sqlite3')

		lst = objst.by_hash('91751cee0a1ab8414400238a761411daa29643ab4b8243e9a91649e25be53ada')
		self.assertEqual(len(lst), 1)

		objst.loadobj({
			'type': 'metadata',
			'uuid': uuid.UUID('c9a1d1e2-3109-4efd-8948-577dc15e44e7'),
			'modified': datetime.datetime(2019, 5, 31, 14, 3, 10,
			    tzinfo=datetime.timezone.utc),
			'created_by_ref': self.created_by_ref,
			'hashes': [ 'sha256:91751cee0a1ab8414400238a761411daa29643ab4b8243e9a91649e25be53ada' ],
			'lang': 'en',
		})

		lst = objst.by_hash('91751cee0a1ab8414400238a761411daa29643ab4b8243e9a91649e25be53ada')
		self.assertEqual(len(lst), 2)

		byid = objst.by_id('3e466e06-45de-4ecc-84ba-2d2a3d970e96')

		self.assertIsInstance(byid, MetaData)
		self.assertIn(byid, lst)

		r = byid

		self.assertEqual(r.uuid, uuid.UUID('3e466e06-45de-4ecc-84ba-2d2a3d970e96'))
		self.assertEqual(r['dc:creator'], [ 'John-Mark Gurney' ])

		# XXX do we care anymore?
		if False:
			# test storing the object store
			fname = 'testfile.sqlite3'
			objst.store(fname)

			with open(fname, 'rb') as fp:
				objs = _asn1coder.loads(fp.read())

			os.unlink(fname)

			self.assertEqual(len(objs), len(objst))

			self.assertEqual(objs['created_by_ref'], self.created_by_ref.bytes)

			# make sure that the read back data matches
			for i in objs['objects']:
				i['created_by_ref'] = uuid.UUID(bytes=i['created_by_ref'])
				i['uuid'] = uuid.UUID(bytes=i['uuid'])
				self.assertEqual(objst.by_id(i['uuid']), i)

		# that a file
		testfname = os.path.join(self.tempdir, 'test.txt')

		# when registered
		objst.loadobj(persona.by_file(testfname))

		# can be found
		self.assertEqual(objst.by_file(testfname), [ byid ])

		self.assertRaises(KeyError, objst.by_file, '/dev/null')

		# that when a metadata object
		mdouuid = 'c9a1d1e2-3109-4efd-8948-577dc15e44e7'
		origobj = objst.by_id(mdouuid)

		# is updated:
		obj = origobj.new_version(('foo', 'bar'))

		# and stored
		objst.loadobj(obj)

		# that it is the new one
		self.assertEqual(obj, objst.by_id(mdouuid))

		# and that the old one isn't present anymore in by file
		lst = objst.by_hash('91751cee0a1ab8414400238a761411daa29643ab4b8243e9a91649e25be53ada')
		self.assertNotIn(origobj, lst)

		# XXX make sure that object store contains fileobject

	# Tests to add:
	#   Non-duplicates when same metadata is located by multiple hashes.

	def objcompare(self, fullobjs, partialobjs):
		fullobjs = list(fullobjs)
		#_debprint('objs:', repr(fullobjs), repr(partialobjs))
		self.assertEqual(len(fullobjs), len(partialobjs))

		missing = []
		for i in partialobjs:
			for idx, j in enumerate(fullobjs):
				cmpobj = dict((k, v) for k, v in j.items() if k in set(i.keys()))
				if cmpobj == i:
					break
			else: # pragma: no cover
				missing.append(i)
				continue

			fullobjs.pop(idx)

		if missing: # pragma: no cover
			_debprint('remaining objs:', repr(fullobjs))
			self.fail('Unable to find objects %s in dump' % missing)

	def save_db(f): # pragma: no cover
		@functools.wraps(f)
		def wrapper(self, *args, **kwargs):
			try:
				return f(self, *args, **kwargs)
			except:
				shutil.copyfile(self.storefname,
				    '/tmp/failure.sqlite3')
				raise

		return wrapper

	#@save_db
	def run_command_file(self, f):
		with open(f) as fp:
			cmds = json.load(fp)

		# setup object store
		storefname = self.tempdir / 'storefname'
		self.storefname = storefname
		identfname = self.tempdir / 'identfname'
		cachefname = self.tempdir / 'cachefname'

		# setup path mapping
		def expandusermock(arg):
			if arg == '~/.medashare_store.sqlite3':
				return storefname
			elif arg == '~/.medashare_identity.pasn1':
				return identfname
			elif arg == '~/.medashare_cache.pasn1':
				return cachefname

			if True: #pragma: no cover
				raise NotImplementedError(arg)

		# setup test fname
		testfname = os.path.join(self.tempdir, 'test.txt')
		newtestfname = os.path.join(self.tempdir, 'newfile.txt')

		patches = []

		for idx, cmd in enumerate(cmds):
			try:
				if cmd['skip']: # pragma: no cover
					continue
			except KeyError:
				pass

			for i in cmd.get('format', []):
				if i in { 'cmd', 'files' }:
					vars = locals()
					cmd[i] = [ x.format(**vars) for x in cmd[i] ]
				else:
					cmd[i] = cmd[i].format(**locals())

			try:
				special = cmd['special']
			except KeyError:
				pass
			else:
				if special == 'copy newfile.txt to test.txt':
					shutil.copy(newtestfname, testfname)
				elif special == 'change newfile.txt':
					with open(newtestfname, 'w') as fp:
						fp.write('some new contents')
				elif special == 'verify store object cnt':
					objst = ObjectStore.load(storefname)
					objcnt = len(objst)
					self.assertEqual(objcnt, len(list(objst)))
					self.assertEqual(objcnt, cmd['count'])
				elif special == 'set hostid':
					hostidpatch = mock.patch(__name__ + '.hostuuid')
					hid = uuid.UUID(cmd['hostid']) if 'hostid' in cmd else uuid.uuid4()
					hostidpatch.start().return_value = hid
					patches.append(hostidpatch)
				elif special == 'iter is unique':
					objst = ObjectStore.load(storefname)
					uniqobjs = len(set((x['uuid'] for x in objst)))
					self.assertEqual(len(list(objst)), uniqobjs)
				elif special == 'setup bittorrent files':
					# copy in the torrent file
					tor = importlib.resources.files('medashare.btv')
					tor = tor / 'fixtures' / 'somedir.torrent'
					shutil.copy(tor, self.tempdir)

					# partly recreate files
					btfiles = bttestcase.origfiledata.copy()

					if not cmd['complete']:
						btfiles.update(bttestcase.badfiles)

					sd = self.tempdir / bttestcase.dirname
					sd.mkdir(exist_ok=True)

					bttestcase.make_files(sd, btfiles)
				elif special == 'setup mapping paths':
					mappatha = self.tempdir / 'mapa'
					mappatha.mkdir()

					mappathb = self.tempdir / 'mapb'
					mappathb.mkdir()

					filea = mappatha / 'text.txt'
					filea.write_text('abc123\n')
					fileb = mappathb / 'text.txt'
					shutil.copyfile(filea, fileb)
					shutil.copystat(filea, fileb)
				elif special == 'delete files':
					for i in cmd['files']:
						os.unlink(i)
				elif special == 'setup file':
					shutil.copy(self.fixtures /
					    cmd['file'], self.tempdir)
				elif special == 'setup tar file':
					shutil.copy(self.fixtures /
					    'testfile.tar.gz', self.tempdir)
				elif special == 'store now':
					storednow = datetime.datetime.now(datetime.timezone.utc)
				elif special == 'format next cmd':
					modcmd = cmds[idx + 1]
					formatkwargs = dict(nowiso=storednow.isoformat())
					modcmd['cmd'] = [ x.format(**formatkwargs) for x in modcmd['cmd'] ]
				else: # pragma: no cover
					raise ValueError('unhandled special: %s' % repr(special))

				# coverage bug, fixed in 3.10:
				# https://github.com/nedbat/coveragepy/issues/1432#event-7130600158
				if True: # pragma: no cover
					continue

			with self.subTest(file=f, title=cmd['title']), \
			    mock.patch('os.path.expanduser',
			    side_effect=expandusermock) as eu, \
			    mock.patch('sys.stdin', io.StringIO()) as stdin, \
			    mock.patch('sys.stdout', io.StringIO()) as stdout, \
			    mock.patch('sys.stderr', io.StringIO()) as stderr, \
			    mock.patch('sys.argv', [ 'progname', ] +
			    cmd['cmd']) as argv:

				# if there is stdin
				test_stdin = cmd.get('stdin', '')

				# provide it
				stdin.write(test_stdin)
				stdin.seek(0)

				with self.assertRaises(SystemExit) as cm:
					try:
						main()
					except AssertionError: # pragma: no cover
						# used to nab a copy of the
						# failed sqlite3 db
						if False:
							shutil.copyfile(
							    storefname,
							    '/tmp/failure.sqlite3')
						raise

					# XXX - Minor hack till other tests fixed
					sys.exit(0)

				# with the correct output
				self.maxDiff = None
				outeq = cmd.get('stdout')
				outnre = cmd.get('stdout_nre')
				outre = cmd.get('stdout_re')
				outcheck = cmd.get('stdout_check')
				# python3 -c 'import ast, sys; print(ast.literal_eval(sys.stdin.read()))' << EOF | jq '.'
				if outnre:
					self.assertNotRegex(stdout.getvalue(), outnre)
				if outre:
					self.assertRegex(stdout.getvalue(), outre)
				if outeq:
					self.assertEqual(stdout.getvalue(), outeq)
				if outcheck:
					stdout.seek(0)
					self.objcompare(_json_objstream(stdout), outcheck)

				self.assertEqual(stderr.getvalue(), cmd.get('stderr', ''))

				self.assertEqual(cm.exception.code, cmd.get('exit', 0))

		patches.reverse()
		for i in patches:
			i.stop()

	def test_get_paths(self):
		# Test to make sure get paths works as expected.
		with mock.patch('os.path.expanduser') as eu:
			a, b, c = _get_paths(None)

			eu.assert_any_call('~/.medashare_identity.pasn1')
			eu.assert_any_call('~/.medashare_store.sqlite3')
			eu.assert_any_call('~/.medashare_cache.pasn1')

		pathpref = pathlib.Path('/somepath/somewhere')
		with mock.patch.dict(os.environ, dict(MEDASHARE_PATH=str(pathpref))):
			i, s, c = _get_paths(None)

			self.assertEqual(i, str(pathpref / '.medashare_identity.pasn1'))
			self.assertEqual(s, str(pathpref / '.medashare_store.sqlite3'))
			self.assertEqual(c, str(pathpref / '.medashare_cache.pasn1'))

	def test_help(self):
		# that subcommand help is the same as --help

		with mock.patch('sys.stdout', io.StringIO()) as stdout, \
		    mock.patch('sys.argv', [ 'progname', '--help', ]) as argv:
			with self.assertRaises(SystemExit) as cm:
				main()

				# XXX - Minor hack till other tests fixed
				sys.exit(0)

			dashhelp = stdout.getvalue()

		with mock.patch('sys.stdout', io.StringIO()) as stdout, \
		    mock.patch('sys.argv', [ 'progname', 'help', ]) as argv:
			with self.assertRaises(SystemExit) as cm:
				main()

				# XXX - Minor hack till other tests fixed
				sys.exit(0)

			subhelp = stdout.getvalue()

		self.assertEqual(dashhelp, subhelp)

		with mock.patch('sys.stdout', io.StringIO()) as stdout, \
		    mock.patch('sys.argv', [ 'progname', ]) as argv:
			with self.assertRaises(SystemExit) as cm:
				main()

				# XXX - Minor hack till other tests fixed
				sys.exit(0)

			subhelp = stdout.getvalue()

		self.assertEqual(dashhelp, subhelp)

	#@unittest.skip('temp')
	def test_cmds(self):
		cmds = sorted(self.fixtures.glob('cmd.*.json'))

		for i in cmds:
			# make sure each file starts with a clean slate
			self.tearDown()
			self.setUp()

			os.chdir(self.tempdir)
			with self.subTest(file=i):
				self.run_command_file(i)

	# XXX - the following test may no longer be needed
	def test_main(self):
		# Test the main runner, this is only testing things that are
		# specific to running the program, like where the store is
		# created.

		# setup object store
		storefname = self.tempdir / 'storefname'
		identfname = self.tempdir / 'identfname'
		cachefname = self.tempdir / 'cachefname'

		# setup path mapping
		def expandusermock(arg):
			if arg == '~/.medashare_store.sqlite3':
				return storefname
			elif arg == '~/.medashare_identity.pasn1':
				return identfname
			elif arg == '~/.medashare_cache.pasn1':
				return cachefname

		# setup test fname
		testfname = os.path.join(self.tempdir, 'test.txt')
		newtestfname = os.path.join(self.tempdir, 'newfile.txt')

		import itertools

		with mock.patch('os.path.expanduser', side_effect=expandusermock) \
		    as eu, mock.patch('medashare.cli.open') as op:
			# that when opening the store and identity fails
			op.side_effect = FileNotFoundError

			# and there is no identity
			with mock.patch('sys.stderr', io.StringIO()) as stderr, mock.patch('sys.argv', [ 'progname', 'list', 'afile' ]) as argv:
				with self.assertRaises(SystemExit) as cm:
					main()

				# that it fails
				self.assertEqual(cm.exception.code, 1)

				# with the correct error message
				self.assertEqual(stderr.getvalue(),
				    'ERROR: Identity not created, create w/ genident.\n')

		with mock.patch('os.path.expanduser', side_effect=expandusermock) \
		    as eu:
			# that generating a new identity
			with mock.patch('sys.stdout', io.StringIO()) as stdout, mock.patch('sys.argv', [ 'progname', 'genident', 'name=A Test User' ]) as argv:
				main()

				# does not output anything
				self.assertEqual(stdout.getvalue(), '')

			# looks up the correct file
			eu.assert_any_call('~/.medashare_identity.pasn1')
			eu.assert_any_call('~/.medashare_store.sqlite3')
			eu.assert_any_call('~/.medashare_cache.pasn1')

			# and that the identity
			persona = Persona.load(identfname)
			pident = persona.get_identity()

			# has the correct name
			self.assertEqual(pident.name, 'A Test User')

			# that when generating an identity when one already exists
			with mock.patch('sys.stderr', io.StringIO()) as stderr, mock.patch('sys.argv', [ 'progname', 'genident', 'name=A Test User' ]) as argv:
				# that it exits
				with self.assertRaises(SystemExit) as cm:
					main()

				# with error code 1
				self.assertEqual(cm.exception.code, 1)

				# and outputs an error message
				self.assertEqual(stderr.getvalue(),
				    'Error: Identity already created.\n')

			# and looked up the correct file
			eu.assert_any_call('~/.medashare_identity.pasn1')

			# that when updating the identity
			with mock.patch('sys.stdout', io.StringIO()) as stdout, mock.patch('sys.argv', [ 'progname', 'ident', 'name=Changed Name' ]) as argv:
				main()

				# it doesn't output anything
				self.assertEqual(stdout.getvalue(), '')

			# and looked up the correct file
			eu.assert_any_call('~/.medashare_identity.pasn1')

			npersona = Persona.load(identfname)
			nident = npersona.get_identity()

			# and has the new name
			self.assertEqual(nident.name, 'Changed Name')

			# and has the same old uuid
			self.assertEqual(nident.uuid, pident.uuid)

			# and that the modified date has changed
			self.assertNotEqual(pident.modified, nident.modified)

			# and that the old Persona can verify the new one
			self.assertTrue(persona.verify(nident))

		orig_open = open
		with mock.patch('os.path.expanduser', side_effect=expandusermock) \
		    as eu, mock.patch('medashare.cli.open') as op:
			# that when the store fails
			def open_repl(fname, mode):
				#print('or:', repr(fname), repr(mode), file=sys.stderr)
				self.assertIn(mode, ('rb', 'wb'))

				if fname == identfname or mode == 'wb':
					return orig_open(fname, mode)

				#print('foo:', repr(fname), repr(mode), file=sys.stderr)
				if True: #pragma: no cover
					raise FileNotFoundError

			op.side_effect = open_repl

			# and there is no store
			with mock.patch('sys.stderr', io.StringIO()) as stderr, mock.patch('sys.argv', [ 'progname', 'list', 'foo', ]) as argv:
				# that it exits
				with self.assertRaises(SystemExit) as cm:
					main()

				# with error code 1
				self.assertEqual(cm.exception.code, 1)

				# and outputs an error message
				self.assertEqual(stderr.getvalue(),
				    'ERROR: file not found: \'foo\'\n')

# Tests to add:
# that cache is updated on modify cmd
# add tests for dump options uuid and hash
# expand mappings to multiple mappings, that is a -> b, b -> c, implies a -> c
# support host names in --create