community.general/lib/ansible/utils/__init__.py

# (c) 2012, Michael DeHaan <michael.dehaan@gmail.com>
#
# This file is part of Ansible
#
# Ansible is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# Ansible is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with Ansible.  If not, see <http://www.gnu.org/licenses/>.

import sys
import re
import os
import shlex
import yaml
import copy
import optparse
import operator
from ansible import errors
from ansible import __version__
from ansible.utils.plugins import *
from ansible.utils import template
import ansible.constants as C
import time
import StringIO
import stat
import string
import termios
import tty
import pipes
import random
import difflib
import warnings
import traceback
import getpass

import hmac
from Crypto.Cipher import 
from Crypto import Random
from Crypto.Random.random import StrongRandom

VERBOSITY=0

MAX_FILE_SIZE_FOR_DIFF=1*1024*1024

try:
    import json
except ImportError:
    import simplejson as json

try:
    from hashlib import md5 as _md5
    from hashlib import sha1 as _sha1
except ImportError:
    from md5 import md5 as _md5
    from sha1 import sha1 as _sha1

PASSLIB_AVAILABLE = False
try:
    import passlib.hash
    PASSLIB_AVAILABLE = True
except:
    pass

###############################################################
# Abstractions around PyCrypto
###############################################################

class AES256Cipher(object):
    """
    Class abstraction of an AES 256 cipher. This class
    also keeps track of the time since the key was last
    generated, so you know when to rekey. Rekeying would
    be done as follows:

    k = AES256Cipher.gen_key()
    <exchange new key with client securely>
    AES26Cipher.set_key(k)

    From this point on the new key would be used until 
    the lifetime is exceeded.
    """
    def __init__(self, lifetime=60*30, mode=AES.MODE_CFB):
        self.lifetime = lifetime
        self.mode = mode
        self.set_key(self.gen_key())

    def gen_key(self):
        """
        Generates a 256-bit (32 byte) key to be used for the 
        AES block encryption. 
        """
        return b"".join(StrongRandom().sample(string.letters+string.digits+string.punctuation,32))

    def set_key(self,key):
        """
        Sets the internal key to the one provided and resets the
        internal time to now. This key should ONLY be set to one 
        generated by gen_key()
        """
        self.init_time = time.time()
        self.key = key

    def should_rekey(self):
        """
        Returns true if the lifetime of the current key has
        exceeded the set lifetime.
        """
        if (time.time() - self.init_time) > self.lifetime:
            return True
        else:
            return False

    def _pad(self, msg):
        """
        Adds padding to the message so that it is a full
        AES block size. Used during encryption of the message.
        """
        pad = AES.block_size - len(msg) % AES.block_size
        return msg + pad * chr(pad)

    def _unpad(self, msg):
        """
        Strips out the padding that _pad added. Used during
        the decryption of the message.
        """
        pad = ord(msg[-1])
        return msg[:-pad]

    def gen_sig(self, msg):
        """
        Generates an HMAC-SHA1 signature for the message
        """
        return hmac.new(self.key, msg, _sha1).digest()

    def validate_sig(self, msg, sig):
        """
        Verifies the generated signature of the message matches
        the signature provided.
        """
        new_sig = self.gen_sig(msg)
        return (new_sig == sig)

    def encrypt(self, msg):
        """
        Encrypt the message using AES. The signature
        is appended to the end of the message and is
        used to verify the integrity of the IV and data.

        Returns a base64-encoded version of the following:

        rval[0:16]   = initialization vector
        rval[16:-20] = cipher text
        rval[-20:]   = signature
        """
        msg = self._pad(msg)
        iv = Random.new().read(AES.block_size)
        cipher = AES.new(self.key, self.mode, iv)
        data = iv + cipher.encrypt(msg)
        sig = self.gen_sig(data)
        return (data + sig).encode('base64')

    def decrypt(self, msg):
        """
        Decrypt the message using AES. The signature is
        used to verify the IV and data before decoding to 
        ensure the integrity of the message. This is an
        HMAC-SHA1 hash, so it is always 20 characters

        The incoming message format (after base64 decoding)
        is as follows:

        msg[0:16]   = initialization vector
        msg[16:-20] = cipher text
        msg[-20:]   = signature (HMAC-SHA1)

        Returns the plain-text of the cipher.
        """
        msg = msg.decode('base64')
        data = msg[0:-20] # iv + cipher text
        msig = msg[-20:]  # hmac-sha1 hash
        if not self.validate_sig(data,msig):
           raise Exception("Failed to validate the message signature")
        iv = msg[:AES.block_size]
        cipher = AES.new(self.key, self.mode, iv)
        return self._unpad(cipher.decrypt(msg)[AES.block_size:])

###############################################################
# UTILITY FUNCTIONS FOR COMMAND LINE TOOLS
###############################################################

def err(msg):
    ''' print an error message to stderr '''

    print >> sys.stderr, msg

def exit(msg, rc=1):
    ''' quit with an error to stdout and a failure code '''

    err(msg)
    sys.exit(rc)

def jsonify(result, format=False):
    ''' format JSON output (uncompressed or uncompressed) '''

    if result is None:
        return "{}"
    result2 = result.copy()
    for key, value in result2.items():
        if type(value) is str:
            result2[key] = value.decode('utf-8', 'ignore')
    if format:
        return json.dumps(result2, sort_keys=True, indent=4)
    else:
        return json.dumps(result2, sort_keys=True)

def write_tree_file(tree, hostname, buf):
    ''' write something into treedir/hostname '''

    # TODO: might be nice to append playbook runs per host in a similar way
    # in which case, we'd want append mode.
    path = os.path.join(tree, hostname)
    fd = open(path, "w+")
    fd.write(buf)
    fd.close()

def is_failed(result):
    ''' is a given JSON result a failed result? '''

    return ((result.get('rc', 0) != 0) or (result.get('failed', False) in [ True, 'True', 'true']))

def is_changed(result):
    ''' is a given JSON result a changed result? '''

    return (result.get('changed', False) in [ True, 'True', 'true'])

def check_conditional(conditional, basedir, inject, fail_on_undefined=False, jinja2=False):

    if jinja2:
        conditional = "jinja2_compare %s" % conditional

    if conditional.startswith("jinja2_compare"):
        conditional = conditional.replace("jinja2_compare ","")
        # allow variable names
        if conditional in inject:
            conditional = inject[conditional]
        conditional = template.template(basedir, conditional, inject, fail_on_undefined=fail_on_undefined)
        # a Jinja2 evaluation that results in something Python can eval!
        presented = "{%% if %s %%} True {%% else %%} False {%% endif %%}" % conditional
        conditional = template.template(basedir, presented, inject)
        val = conditional.lstrip().rstrip()
        if val == "True":
            return True
        elif val == "False":
            return False
        else:
            raise errors.AnsibleError("unable to evaluate conditional: %s" % conditional)

    if not isinstance(conditional, basestring):
        return conditional

    try:
        conditional = conditional.replace("\n", "\\n")
        result = safe_eval(conditional)
        if result not in [ True, False ]:
            raise errors.AnsibleError("Conditional expression must evaluate to True or False: %s" % conditional)
        return result

    except (NameError, SyntaxError):
        raise errors.AnsibleError("Could not evaluate the expression: (%s)" % conditional)

def is_executable(path):
    '''is the given path executable?'''
    return (stat.S_IXUSR & os.stat(path)[stat.ST_MODE]
            or stat.S_IXGRP & os.stat(path)[stat.ST_MODE]
            or stat.S_IXOTH & os.stat(path)[stat.ST_MODE])

def unfrackpath(path):
    ''' 
    returns a path that is free of symlinks, environment
    variables, relative path traversals and symbols (~)
    example:
    '$HOME/../../var/mail' becomes '/var/spool/mail'
    '''
    return os.path.normpath(os.path.realpath(os.path.expandvars(os.path.expanduser(path))))

def prepare_writeable_dir(tree,mode=0777):
    ''' make sure a directory exists and is writeable '''

    # modify the mode to ensure the owner at least
    # has read/write access to this directory
    mode |= 0700

    # make sure the tree path is always expanded
    # and normalized and free of symlinks
    tree = unfrackpath(tree)

    if not os.path.exists(tree):
        try:
            os.makedirs(tree, mode)
        except (IOError, OSError), e:
            raise errors.AnsibleError("Could not make dir %s: %s" % (tree, e))
    if not os.access(tree, os.W_OK):
        raise errors.AnsibleError("Cannot write to path %s" % tree)
    return tree

def path_dwim(basedir, given):
    '''
    make relative paths work like folks expect.
    '''

    if given.startswith("/"):
        return os.path.abspath(given)
    elif given.startswith("~"):
        return os.path.abspath(os.path.expanduser(given))
    else:
        return os.path.abspath(os.path.join(basedir, given))

def path_dwim_relative(original, dirname, source, playbook_base, check=True):
    ''' find one file in a directory one level up in a dir named dirname relative to current '''
    # (used by roles code)

    basedir = os.path.dirname(original)
    template2 = os.path.join(basedir, '..', dirname, source)
    source2 = path_dwim(basedir, template2)
    if os.path.exists(source2):
        return source2
    obvious_local_path = path_dwim(playbook_base, source)
    if os.path.exists(obvious_local_path):
        return obvious_local_path
    if check:
        raise errors.AnsibleError("input file not found at %s or %s" % (source2, obvious_local_path))
    return source2 # which does not exist

def json_loads(data):
    ''' parse a JSON string and return a data structure '''

    return json.loads(data)

def parse_json(raw_data):
    ''' this version for module return data only '''

    orig_data = raw_data

    # ignore stuff like tcgetattr spewage or other warnings
    data = filter_leading_non_json_lines(raw_data)

    try:
        return json.loads(data)
    except:
        # not JSON, but try "Baby JSON" which allows many of our modules to not
        # require JSON and makes writing modules in bash much simpler
        results = {}
        try:
            tokens = shlex.split(data)
        except:
            print "failed to parse json: "+ data
            raise

        for t in tokens:
            if t.find("=") == -1:
                raise errors.AnsibleError("failed to parse: %s" % orig_data)
            (key,value) = t.split("=", 1)
            if key == 'changed' or 'failed':
                if value.lower() in [ 'true', '1' ]:
                    value = True
                elif value.lower() in [ 'false', '0' ]:
                    value = False
            if key == 'rc':
                value = int(value)
            results[key] = value
        if len(results.keys()) == 0:
            return { "failed" : True, "parsed" : False, "msg" : orig_data }
        return results

def smush_braces(data):
    ''' smush Jinaj2 braces so unresolved templates like {{ foo }} don't get parsed weird by key=value code '''
    while data.find('{{ ') != -1:
        data = data.replace('{{ ', '{{')
    while data.find(' }}') != -1:
        data = data.replace(' }}', '}}')
    return data

def smush_ds(data):
    # things like key={{ foo }} are not handled by shlex.split well, so preprocess any YAML we load
    # so we do not have to call smush elsewhere
    if type(data) == list:
        return [ smush_ds(x) for x in data ]
    elif type(data) == dict:
        for (k,v) in data.items():
            data[k] = smush_ds(v)
        return data
    elif isinstance(data, basestring):
        return smush_braces(data)
    else:
        return data

def parse_yaml(data):
    ''' convert a yaml string to a data structure '''
    return smush_ds(yaml.safe_load(data))

def process_yaml_error(exc, data, path=None):
    if hasattr(exc, 'problem_mark'):
        mark = exc.problem_mark
        if mark.line -1 >= 0:
            before_probline = data.split("\n")[mark.line-1]
        else:
            before_probline = ''
        probline = data.split("\n")[mark.line]
        arrow = " " * mark.column + "^"
        msg = """Syntax Error while loading YAML script, %s
Note: The error may actually appear before this position: line %s, column %s

%s
%s
%s""" % (path, mark.line + 1, mark.column + 1, before_probline, probline, arrow)
    else:
        # No problem markers means we have to throw a generic
        # "stuff messed up" type message. Sry bud.
        if path:
            msg = "Could not parse YAML. Check over %s again." % path
        else:
            msg = "Could not parse YAML."
    raise errors.AnsibleYAMLValidationFailed(msg)


def parse_yaml_from_file(path):
    ''' convert a yaml file to a data structure '''

    try:
        data = file(path).read()
        return parse_yaml(data)
    except IOError:
        raise errors.AnsibleError("file not found: %s" % path)
    except yaml.YAMLError, exc:
        process_yaml_error(exc, data, path)

def parse_kv(args):
    ''' convert a string of key/value items to a dict '''
    options = {}
    if args is not None:
        # attempting to split a unicode here does bad things
        args = args.encode('utf-8')
        vargs = [x.decode('utf-8') for x in shlex.split(args, posix=True)]
        #vargs = shlex.split(str(args), posix=True)
        for x in vargs:
            if x.find("=") != -1:
                k, v = x.split("=",1)
                options[k]=v
    return options

def merge_hash(a, b):
    ''' recursively merges hash b into a
    keys from b take precedence over keys from a '''

    result = copy.deepcopy(a)

    # next, iterate over b keys and values
    for k, v in b.iteritems():
        # if there's already such key in a
        # and that key contains dict
        if k in result and isinstance(result[k], dict):
            # merge those dicts recursively
            result[k] = merge_hash(a[k], v)
        else:
            # otherwise, just copy a value from b to a
            result[k] = v

    return result

def md5s(data):
    ''' Return MD5 hex digest of data. '''

    digest = _md5()
    digest.update(data.encode('utf-8'))
    return digest.hexdigest()

def md5(filename):
    ''' Return MD5 hex digest of local file, or None if file is not present. '''

    if not os.path.exists(filename):
        return None
    digest = _md5()
    blocksize = 64 * 1024
    infile = open(filename, 'rb')
    block = infile.read(blocksize)
    while block:
        digest.update(block)
        block = infile.read(blocksize)
    infile.close()
    return digest.hexdigest()

def default(value, function):
    ''' syntactic sugar around lazy evaluation of defaults '''
    if value is None:
        return function()
    return value

def _gitinfo():
    ''' returns a string containing git branch, commit id and commit date '''
    result = None
    repo_path = os.path.join(os.path.dirname(__file__), '..', '..', '..', '.git')

    if os.path.exists(repo_path):
        # Check if the .git is a file. If it is a file, it means that we are in a submodule structure.
        if os.path.isfile(repo_path):
            try:
                gitdir = yaml.safe_load(open(repo_path)).get('gitdir')
                # There is a posibility the .git file to have an absolute path.
                if os.path.isabs(gitdir):
                    repo_path = gitdir
                else:
                    repo_path = os.path.join(repo_path.split('.git')[0], gitdir)
            except (IOError, AttributeError):
                return ''
        f = open(os.path.join(repo_path, "HEAD"))
        branch = f.readline().split('/')[-1].rstrip("\n")
        f.close()
        branch_path = os.path.join(repo_path, "refs", "heads", branch)
        if os.path.exists(branch_path):
            f = open(branch_path)
            commit = f.readline()[:10]
            f.close()
            date = time.localtime(os.stat(branch_path).st_mtime)
            if time.daylight == 0:
                offset = time.timezone
            else:
                offset = time.altzone
            result = "({0} {1}) last updated {2} (GMT {3:+04d})".format(branch, commit,
                time.strftime("%Y/%m/%d %H:%M:%S", date), offset / -36)
    else:
        result = ''
    return result

def version(prog):
    result = "{0} {1}".format(prog, __version__)
    gitinfo = _gitinfo()
    if gitinfo:
        result = result + " {0}".format(gitinfo)
    return result

def getch():
    ''' read in a single character '''
    fd = sys.stdin.fileno()
    old_settings = termios.tcgetattr(fd)
    try:
        tty.setraw(sys.stdin.fileno())
        ch = sys.stdin.read(1)
    finally:
        termios.tcsetattr(fd, termios.TCSADRAIN, old_settings)
    return ch

####################################################################
# option handling code for /usr/bin/ansible and ansible-playbook
# below this line

class SortedOptParser(optparse.OptionParser):
    '''Optparser which sorts the options by opt before outputting --help'''

    def format_help(self, formatter=None):
        self.option_list.sort(key=operator.methodcaller('get_opt_string'))
        return optparse.OptionParser.format_help(self, formatter=None)

def increment_debug(option, opt, value, parser):
    global VERBOSITY
    VERBOSITY += 1

def base_parser(constants=C, usage="", output_opts=False, runas_opts=False,
    async_opts=False, connect_opts=False, subset_opts=False, check_opts=False, diff_opts=False):
    ''' create an options parser for any ansible script '''

    parser = SortedOptParser(usage, version=version("%prog"))
    parser.add_option('-v','--verbose', default=False, action="callback",
        callback=increment_debug, help="verbose mode (-vvv for more)")

    parser.add_option('-f','--forks', dest='forks', default=constants.DEFAULT_FORKS, type='int',
        help="specify number of parallel processes to use (default=%s)" % constants.DEFAULT_FORKS)
    parser.add_option('-i', '--inventory-file', dest='inventory',
        help="specify inventory host file (default=%s)" % constants.DEFAULT_HOST_LIST,
        default=constants.DEFAULT_HOST_LIST)
    parser.add_option('-k', '--ask-pass', default=False, dest='ask_pass', action='store_true',
        help='ask for SSH password')
    parser.add_option('--private-key', default=C.DEFAULT_PRIVATE_KEY_FILE, dest='private_key_file',
        help='use this file to authenticate the connection')
    parser.add_option('-K', '--ask-sudo-pass', default=False, dest='ask_sudo_pass', action='store_true',
        help='ask for sudo password')
    parser.add_option('--list-hosts', dest='listhosts', action='store_true',
        help='outputs a list of matching hosts; does not execute anything else')
    parser.add_option('-M', '--module-path', dest='module_path',
        help="specify path(s) to module library (default=%s)" % constants.DEFAULT_MODULE_PATH,
        default=None)

    if subset_opts:
        parser.add_option('-l', '--limit', default=constants.DEFAULT_SUBSET, dest='subset',
            help='further limit selected hosts to an additional pattern')

    parser.add_option('-T', '--timeout', default=constants.DEFAULT_TIMEOUT, type='int',
        dest='timeout',
        help="override the SSH timeout in seconds (default=%s)" % constants.DEFAULT_TIMEOUT)

    if output_opts:
        parser.add_option('-o', '--one-line', dest='one_line', action='store_true',
            help='condense output')
        parser.add_option('-t', '--tree', dest='tree', default=None,
            help='log output to this directory')

    if runas_opts:
        parser.add_option("-s", "--sudo", default=constants.DEFAULT_SUDO, action="store_true",
            dest='sudo', help="run operations with sudo (nopasswd)")
        parser.add_option('-U', '--sudo-user', dest='sudo_user', help='desired sudo user (default=root)',
            default=None)   # Can't default to root because we need to detect when this option was given
        parser.add_option('-u', '--user', default=constants.DEFAULT_REMOTE_USER,
            dest='remote_user',
            help='connect as this user (default=%s)' % constants.DEFAULT_REMOTE_USER)

    if connect_opts:
        parser.add_option('-c', '--connection', dest='connection',
                          default=C.DEFAULT_TRANSPORT,
                          help="connection type to use (default=%s)" % C.DEFAULT_TRANSPORT)

    if async_opts:
        parser.add_option('-P', '--poll', default=constants.DEFAULT_POLL_INTERVAL, type='int',
            dest='poll_interval',
            help="set the poll interval if using -B (default=%s)" % constants.DEFAULT_POLL_INTERVAL)
        parser.add_option('-B', '--background', dest='seconds', type='int', default=0,
            help='run asynchronously, failing after X seconds (default=N/A)')

    if check_opts:
        parser.add_option("-C", "--check", default=False, dest='check', action='store_true',
            help="don't make any changes; instead, try to predict some of the changes that may occur"
        )

    if diff_opts:
        parser.add_option("-D", "--diff", default=False, dest='diff', action='store_true',
            help="when changing (small) files and templates, show the differences in those files; works great with --check"
        )


    return parser

def ask_passwords(ask_pass=False, ask_sudo_pass=False):
    sshpass = None
    sudopass = None
    sudo_prompt = "sudo password: "

    if ask_pass:
        sshpass = getpass.getpass(prompt="SSH password: ")
        sudo_prompt = "sudo password [defaults to SSH password]: "

    if ask_sudo_pass:
        sudopass = getpass.getpass(prompt=sudo_prompt)
        if ask_pass and sudopass == '':
            sudopass = sshpass

    return (sshpass, sudopass)

def do_encrypt(result, encrypt, salt_size=None, salt=None):
    if PASSLIB_AVAILABLE:
        try:
            crypt = getattr(passlib.hash, encrypt)
        except:
            raise errors.AnsibleError("passlib does not support '%s' algorithm" % encrypt)

        if salt_size:
            result = crypt.encrypt(result, salt_size=salt_size)
        elif salt:
            result = crypt.encrypt(result, salt=salt)
        else:
            result = crypt.encrypt(result)
    else:
        raise errors.AnsibleError("passlib must be installed to encrypt vars_prompt values")

    return result

def last_non_blank_line(buf):

    all_lines = buf.splitlines()
    all_lines.reverse()
    for line in all_lines:
        if (len(line) > 0):
            return line
    # shouldn't occur unless there's no output
    return ""

def filter_leading_non_json_lines(buf):
    '''
    used to avoid random output from SSH at the top of JSON output, like messages from
    tcagetattr, or where dropbear spews MOTD on every single command (which is nuts).

    need to filter anything which starts not with '{', '[', ', '=' or is an empty line.
    filter only leading lines since multiline JSON is valid.
    '''

    filtered_lines = StringIO.StringIO()
    stop_filtering = False
    for line in buf.splitlines():
        if stop_filtering or "=" in line or line.startswith('{') or line.startswith('['):
            stop_filtering = True
            filtered_lines.write(line + '\n')
    return filtered_lines.getvalue()

def boolean(value):
    val = str(value)
    if val.lower() in [ "true", "t", "y", "1", "yes" ]:
        return True
    else:
        return False

def compile_when_to_only_if(expression):
    '''
    when is a shorthand for writing only_if conditionals.  It requires less quoting
    magic.  only_if is retained for backwards compatibility.
    '''

    # when: set $variable
    # when: unset $variable
    # when: failed $json_result
    # when: changed $json_result
    # when: int $x >= $z and $y < 3
    # when: int $x in $alist
    # when: float $x > 2 and $y <= $z
    # when: str $x != $y
    # when: jinja2_compare asdf  # implies {{ asdf }}

    if type(expression) not in [ str, unicode ]:
        raise errors.AnsibleError("invalid usage of when_ operator: %s" % expression)
    tokens = expression.split()
    if len(tokens) < 2:
        raise errors.AnsibleError("invalid usage of when_ operator: %s" % expression)

    # when_set / when_unset
    if tokens[0] in [ 'set', 'unset' ]:
        tcopy = tokens[1:]
        for (i,t) in enumerate(tokens[1:]):
            if t.find("$") != -1:
                tcopy[i] = "is_%s('''%s''')" % (tokens[0], t)
            else:
                tcopy[i] = t
        return " ".join(tcopy)

    # when_failed / when_changed
    elif tokens[0] in [ 'failed', 'changed' ]:
        tcopy = tokens[1:]
        for (i,t) in enumerate(tokens[1:]):
            if t.find("$") != -1:
                tcopy[i] = "is_%s(%s)" % (tokens[0], t)
            else:
                tcopy[i] = t
        return " ".join(tcopy)

    # when_integer / when_float / when_string
    elif tokens[0] in [ 'integer', 'float', 'string' ]:
        cast = None
        if tokens[0] == 'integer':
            cast = 'int'
        elif tokens[0] == 'string':
            cast = 'str'
        elif tokens[0] == 'float':
            cast = 'float'
        tcopy = tokens[1:]
        for (i,t) in enumerate(tokens[1:]):
            #if re.search(t, r"^\w"):
                # bare word will turn into Jinja2 so all the above
                # casting is really not needed
                #tcopy[i] = "%s('''%s''')" % (cast, t)
            t2 = t.strip()
            if (t2[0].isalpha() or t2[0] == '$') and cast == 'str' and t2 != 'in':
                tcopy[i] = "'%s'" % (t)
            else:
                tcopy[i] = t
        result = " ".join(tcopy)
        return result


    # when_boolean
    elif tokens[0] in [ 'bool', 'boolean' ]:
        tcopy = tokens[1:]
        for (i, t) in enumerate(tcopy):
            if t.find("$") != -1:
                tcopy[i] = "(is_set('''%s''') and '''%s'''.lower() not in ('false', 'no', 'n', 'none', '0', ''))" % (t, t)
        return " ".join(tcopy)

    # the stock 'when' without qualification (new in 1.2), assumes Jinja2 terms
    elif tokens[0] == 'jinja2_compare':
        return " ".join(tokens)
    else:
        raise errors.AnsibleError("invalid usage of when_ operator: %s" % expression)

def make_sudo_cmd(sudo_user, executable, cmd):
    """
    helper function for connection plugins to create sudo commands
    """
    # Rather than detect if sudo wants a password this time, -k makes
    # sudo always ask for a password if one is required.
    # Passing a quoted compound command to sudo (or sudo -s)
    # directly doesn't work, so we shellquote it with pipes.quote()
    # and pass the quoted string to the user's shell.  We loop reading
    # output until we see the randomly-generated sudo prompt set with
    # the -p option.
    randbits = ''.join(chr(random.randint(ord('a'), ord('z'))) for x in xrange(32))
    prompt = '[sudo via ansible, key=%s] password: ' % randbits
    sudocmd = '%s -k && %s %s -S -p "%s" -u %s %s -c %s' % (
        C.DEFAULT_SUDO_EXE, C.DEFAULT_SUDO_EXE, C.DEFAULT_SUDO_FLAGS,
        prompt, sudo_user, executable or '$SHELL', pipes.quote(cmd))
    return ('/bin/sh -c ' + pipes.quote(sudocmd), prompt)

_TO_UNICODE_TYPES = (unicode, type(None))

def to_unicode(value):
    if isinstance(value, _TO_UNICODE_TYPES):
        return value
    return value.decode("utf-8")

def get_diff(diff):
    # called by --diff usage in playbook and runner via callbacks
    # include names in diffs 'before' and 'after' and do diff -U 10

    try:
        with warnings.catch_warnings():
            warnings.simplefilter('ignore')
            ret = []
            if 'dst_binary' in diff:
                ret.append("diff skipped: destination file appears to be binary\n")
            if 'src_binary' in diff:
                ret.append("diff skipped: source file appears to be binary\n")
            if 'dst_larger' in diff:
                ret.append("diff skipped: destination file size is greater than %d\n" % diff['dst_larger'])
            if 'src_larger' in diff:
                ret.append("diff skipped: source file size is greater than %d\n" % diff['src_larger'])
            if 'before' in diff and 'after' in diff:
                if 'before_header' in diff:
                    before_header = "before: %s" % diff['before_header']
                else:
                    before_header = 'before'
                if 'after_header' in diff:
                    after_header = "after: %s" % diff['after_header']
                else:
                    after_header = 'after'
                differ = difflib.unified_diff(to_unicode(diff['before']).splitlines(True), to_unicode(diff['after']).splitlines(True), before_header, after_header, '', '', 10)
                for line in list(differ):
                    ret.append(line)
            return u"".join(ret)
    except UnicodeDecodeError:
        return ">> the files are different, but the diff library cannot compare unicode strings"

def is_list_of_strings(items):
    for x in items:
        if not isinstance(x, basestring):
            return False
    return True

def safe_eval(str):
    '''
    this is intended for allowing things like:
    with_items: {{ a_list_variable }}
    where Jinja2 would return a string
    but we do not want to allow it to call functions (outside of Jinja2, where
    the env is constrained)
    '''
    # FIXME: is there a more native way to do this?

    def is_set(var):
        return not var.startswith("$") and not '{{' in var

    def is_unset(var):
        return var.startswith("$") or '{{' in var

    # do not allow method calls to modules
    if not isinstance(str, basestring):
        # already templated to a datastructure, perhaps?
        return str
    if re.search(r'\w\.\w+\(', str):
        return str
    # do not allow imports
    if re.search(r'import \w+', str):
        return str
    try:
        return eval(str)
    except Exception, e:
        return str


def listify_lookup_plugin_terms(terms, basedir, inject):

    if isinstance(terms, basestring):
        # someone did:
        #    with_items: alist
        # OR
        #    with_items: {{ alist }}

        stripped = terms.strip()
        if not (stripped.startswith('{') or stripped.startswith('[')) and not stripped.startswith("/"):
            # if not already a list, get ready to evaluate with Jinja2
            # not sure why the "/" is in above code :)
            try:
                new_terms = template.template(basedir, "{{ %s }}" % terms, inject)
                if isinstance(new_terms, basestring) and new_terms.find("{{") != -1:
                    pass
                else:
                    terms = new_terms
            except:
                pass

        if '{' in terms or '[' in terms:
            # Jinja2 already evaluated a variable to a list.
            # Jinja2-ified list needs to be converted back to a real type
            # TODO: something a bit less heavy than eval
            return safe_eval(terms)

        if isinstance(terms, basestring):
            terms = [ terms ]

    return terms

def combine_vars(a, b):
    if C.DEFAULT_HASH_BEHAVIOUR == "merge":
        return merge_hash(a, b)
    else:
        return dict(a.items() + b.items())