From f895cb12141281702b34da18f2384deb64c881e7 Mon Sep 17 00:00:00 2001 From: Barry Warsaw Date: Wed, 21 Jan 2015 17:56:53 -0500 Subject: Largely merge lp:~harlowja/cloud-init/py2-3 albeit manually because it seemed to be behind trunk. `tox -e py27` passes full test suite. Now to work on replacing mocker. --- cloudinit/sources/DataSourceSmartOS.py | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) (limited to 'cloudinit/sources/DataSourceSmartOS.py') diff --git a/cloudinit/sources/DataSourceSmartOS.py b/cloudinit/sources/DataSourceSmartOS.py index 2733a2f6..7a975d78 100644 --- a/cloudinit/sources/DataSourceSmartOS.py +++ b/cloudinit/sources/DataSourceSmartOS.py @@ -30,12 +30,12 @@ # Comments with "@datadictionary" are snippets of the definition import base64 +import os +import serial + from cloudinit import log as logging from cloudinit import sources from cloudinit import util -import os -import os.path -import serial LOG = logging.getLogger(__name__) @@ -201,7 +201,7 @@ class DataSourceSmartOS(sources.DataSource): if b64_all is not None: self.b64_all = util.is_true(b64_all) - for ci_noun, attribute in SMARTOS_ATTRIB_MAP.iteritems(): + for ci_noun, attribute in SMARTOS_ATTRIB_MAP.items(): smartos_noun, strip = attribute md[ci_noun] = self.query(smartos_noun, strip=strip) @@ -218,11 +218,12 @@ class DataSourceSmartOS(sources.DataSource): user_script = os.path.join(data_d, 'user-script') u_script_l = "%s/user-script" % LEGACY_USER_D write_boot_content(md.get('user-script'), content_f=user_script, - link=u_script_l, shebang=True, mode=0700) + link=u_script_l, shebang=True, mode=0o700) operator_script = os.path.join(data_d, 'operator-script') write_boot_content(md.get('operator-script'), - content_f=operator_script, shebang=False, mode=0700) + content_f=operator_script, shebang=False, + mode=0o700) # @datadictionary: This key has no defined format, but its value # is written to the file /var/db/mdata-user-data on each boot prior @@ -381,7 +382,7 @@ def dmi_data(): def write_boot_content(content, content_f, link=None, shebang=False, - mode=0400): + mode=0o400): """ Write the content to content_f. Under the following rules: 1. If no content, remove the file -- cgit v1.2.3 From de5974fe93dd717e0c7ba6de17db3192cc258cff Mon Sep 17 00:00:00 2001 From: Barry Warsaw Date: Mon, 26 Jan 2015 14:31:09 -0500 Subject: * More str/bytes fixes. * Temporarily skip the MAAS tests in py3 since they need to be ported to oauthlib. --- cloudinit/sources/DataSourceOpenNebula.py | 12 +++++++++--- cloudinit/sources/DataSourceSmartOS.py | 15 +++++++++++++-- tests/unittests/test_datasource/test_maas.py | 7 ++++++- tests/unittests/test_datasource/test_opennebula.py | 4 ++-- 4 files changed, 30 insertions(+), 8 deletions(-) (limited to 'cloudinit/sources/DataSourceSmartOS.py') diff --git a/cloudinit/sources/DataSourceOpenNebula.py b/cloudinit/sources/DataSourceOpenNebula.py index f9dac29e..691b39f8 100644 --- a/cloudinit/sources/DataSourceOpenNebula.py +++ b/cloudinit/sources/DataSourceOpenNebula.py @@ -379,7 +379,8 @@ def read_context_disk_dir(source_dir, asuser=None): raise BrokenContextDiskDir("configured user '%s' " "does not exist", asuser) try: - with open(os.path.join(source_dir, 'context.sh'), 'r') as f: + path = os.path.join(source_dir, 'context.sh') + with open(path, 'r', encoding='utf-8') as f: content = f.read().strip() context = parse_shell_config(content, asuser=asuser) @@ -426,14 +427,19 @@ def read_context_disk_dir(source_dir, asuser=None): context.get('USER_DATA_ENCODING')) if encoding == "base64": try: - results['userdata'] = base64.b64decode(results['userdata']) + userdata = base64.b64decode(results['userdata']) + # In Python 3 we still expect a str, but b64decode will return + # bytes. Convert to str. + if isinstance(userdata, bytes): + userdata = userdata.decode('utf-8') + results['userdata'] = userdata except TypeError: LOG.warn("Failed base64 decoding of userdata") # generate static /etc/network/interfaces # only if there are any required context variables # http://opennebula.org/documentation:rel3.8:cong#network_configuration - for k in context.keys(): + for k in context: if re.match(r'^ETH\d+_IP$', k): (out, _) = util.subp(['/sbin/ip', 'link']) net = OpenNebulaNetwork(out, context) diff --git a/cloudinit/sources/DataSourceSmartOS.py b/cloudinit/sources/DataSourceSmartOS.py index 7a975d78..d3ed40c5 100644 --- a/cloudinit/sources/DataSourceSmartOS.py +++ b/cloudinit/sources/DataSourceSmartOS.py @@ -30,6 +30,7 @@ # Comments with "@datadictionary" are snippets of the definition import base64 +import binascii import os import serial @@ -350,8 +351,18 @@ def query_data(noun, seed_device, seed_timeout, strip=False, default=None, if b64: try: - return base64.b64decode(resp) - except TypeError: + # Generally, we want native strings in the values. Python 3's + # b64decode will return bytes though, so decode them to utf-8 if + # possible. If that fails, return the bytes. + decoded = base64.b64decode(resp) + try: + if isinstance(decoded, bytes): + return decoded.decode('utf-8') + except UnicodeDecodeError: + pass + return decoded + # Bogus input produces different errors in Python 2 and 3; catch both. + except (TypeError, binascii.Error): LOG.warn("Failed base64 decoding key '%s'", noun) return resp diff --git a/tests/unittests/test_datasource/test_maas.py b/tests/unittests/test_datasource/test_maas.py index 6af0cd82..66fe22ae 100644 --- a/tests/unittests/test_datasource/test_maas.py +++ b/tests/unittests/test_datasource/test_maas.py @@ -4,7 +4,11 @@ import shutil import tempfile import unittest -from cloudinit.sources import DataSourceMAAS +# XXX DataSourceMAAS must be ported to oauthlib for Python 3 +import six +if not six.PY3: + from cloudinit.sources import DataSourceMAAS + from cloudinit import url_helper from ..helpers import populate_dir @@ -14,6 +18,7 @@ except ImportError: import mock +@unittest.skipIf(six.PY3, 'DataSourceMAAS must be ported to oauthlib') class TestMAASDataSource(unittest.TestCase): def setUp(self): diff --git a/tests/unittests/test_datasource/test_opennebula.py b/tests/unittests/test_datasource/test_opennebula.py index 31c6232f..ef534bab 100644 --- a/tests/unittests/test_datasource/test_opennebula.py +++ b/tests/unittests/test_datasource/test_opennebula.py @@ -33,7 +33,7 @@ TEST_VARS = { } INVALID_CONTEXT = ';' -USER_DATA = b'#cloud-config\napt_upgrade: true' +USER_DATA = '#cloud-config\napt_upgrade: true' SSH_KEY = 'ssh-rsa AAAAB3NzaC1....sIkJhq8wdX+4I3A4cYbYP ubuntu@server-460-%i' HOSTNAME = 'foo.example.com' PUBLIC_IP = '10.0.0.3' @@ -300,7 +300,7 @@ class TestParseShellConfig(unittest.TestCase): def populate_context_dir(path, variables): data = "# Context variables generated by OpenNebula\n" - for (k, v) in variables.items(): + for k, v in variables.items(): data += ("%s='%s'\n" % (k.upper(), v.replace(r"'", r"'\''"))) populate_dir(path, {'context.sh': data}) -- cgit v1.2.3 From 6e742d20e9ed56498925c7c850cd5da65d063b4b Mon Sep 17 00:00:00 2001 From: Barry Warsaw Date: Tue, 27 Jan 2015 15:03:52 -0500 Subject: Respond to review: - Refactor both the base64 encoding and decoding into utility functions. Also: - Mechanically fix some other broken untested code. --- cloudinit/config/cc_seed_random.py | 8 +------ cloudinit/config/cc_ssh_authkey_fingerprints.py | 2 +- cloudinit/sources/DataSourceOpenNebula.py | 7 +----- cloudinit/sources/DataSourceSmartOS.py | 11 +-------- cloudinit/util.py | 20 ++++++++++++++++ tests/unittests/test_datasource/test_azure.py | 28 ++++++++-------------- tests/unittests/test_datasource/test_opennebula.py | 11 ++------- tests/unittests/test_datasource/test_smartos.py | 14 ++++------- .../test_handler/test_handler_seed_random.py | 12 ++-------- 9 files changed, 42 insertions(+), 71 deletions(-) (limited to 'cloudinit/sources/DataSourceSmartOS.py') diff --git a/cloudinit/config/cc_seed_random.py b/cloudinit/config/cc_seed_random.py index 981e1b08..bb64b0f5 100644 --- a/cloudinit/config/cc_seed_random.py +++ b/cloudinit/config/cc_seed_random.py @@ -38,13 +38,7 @@ def _decode(data, encoding=None): if not encoding or encoding.lower() in ['raw']: return data elif encoding.lower() in ['base64', 'b64']: - # Try to give us a native string in both Python 2 and 3, and remember - # that b64decode() returns bytes in Python 3. - decoded = base64.b64decode(data) - try: - return decoded.decode('utf-8') - except UnicodeDecodeError: - return decoded + return util.b64d(data) elif encoding.lower() in ['gzip', 'gz']: return util.decomp_gzip(data, quiet=False) else: diff --git a/cloudinit/config/cc_ssh_authkey_fingerprints.py b/cloudinit/config/cc_ssh_authkey_fingerprints.py index 51580633..6ce831bc 100644 --- a/cloudinit/config/cc_ssh_authkey_fingerprints.py +++ b/cloudinit/config/cc_ssh_authkey_fingerprints.py @@ -32,7 +32,7 @@ from cloudinit import util def _split_hash(bin_hash): split_up = [] - for i in xrange(0, len(bin_hash), 2): + for i in range(0, len(bin_hash), 2): split_up.append(bin_hash[i:i + 2]) return split_up diff --git a/cloudinit/sources/DataSourceOpenNebula.py b/cloudinit/sources/DataSourceOpenNebula.py index a0275cda..61709c1b 100644 --- a/cloudinit/sources/DataSourceOpenNebula.py +++ b/cloudinit/sources/DataSourceOpenNebula.py @@ -426,12 +426,7 @@ def read_context_disk_dir(source_dir, asuser=None): context.get('USER_DATA_ENCODING')) if encoding == "base64": try: - userdata = base64.b64decode(results['userdata']) - # In Python 3 we still expect a str, but b64decode will return - # bytes. Convert to str. - if isinstance(userdata, bytes): - userdata = userdata.decode('utf-8') - results['userdata'] = userdata + results['userdata'] = util.b64d(results['userdata']) except TypeError: LOG.warn("Failed base64 decoding of userdata") diff --git a/cloudinit/sources/DataSourceSmartOS.py b/cloudinit/sources/DataSourceSmartOS.py index f59ad3d6..9d48beab 100644 --- a/cloudinit/sources/DataSourceSmartOS.py +++ b/cloudinit/sources/DataSourceSmartOS.py @@ -351,16 +351,7 @@ def query_data(noun, seed_device, seed_timeout, strip=False, default=None, if b64: try: - # Generally, we want native strings in the values. Python 3's - # b64decode will return bytes though, so decode them to utf-8 if - # possible. If that fails, return the bytes. - decoded = base64.b64decode(resp) - try: - if isinstance(decoded, bytes): - return decoded.decode('utf-8') - except UnicodeDecodeError: - pass - return decoded + return util.b64d(resp) # Bogus input produces different errors in Python 2 and 3; catch both. except (TypeError, binascii.Error): LOG.warn("Failed base64 decoding key '%s'", noun) diff --git a/cloudinit/util.py b/cloudinit/util.py index 766f8e32..8916cc11 100644 --- a/cloudinit/util.py +++ b/cloudinit/util.py @@ -44,6 +44,7 @@ import sys import tempfile import time +from base64 import b64decode, b64encode from six.moves.urllib import parse as urlparse import six @@ -90,6 +91,25 @@ def encode_text(text, encoding='utf-8'): return text return text.encode(encoding) + +def b64d(source): + # Base64 decode some data, accepting bytes or unicode/str, and returning + # str/unicode if the result is utf-8 compatible, otherwise returning bytes. + decoded = b64decode(source) + if isinstance(decoded, bytes): + try: + return decoded.decode('utf-8') + except UnicodeDecodeError: + return decoded + +def b64e(source): + # Base64 encode some data, accepting bytes or unicode/str, and returning + # str/unicode if the result is utf-8 compatible, otherwise returning bytes. + if not isinstance(source, bytes): + source = source.encode('utf-8') + return b64encode(source).decode('utf-8') + + # Path for DMI Data DMI_SYS_PATH = "/sys/class/dmi/id" diff --git a/tests/unittests/test_datasource/test_azure.py b/tests/unittests/test_datasource/test_azure.py index 97a53bee..965bce4b 100644 --- a/tests/unittests/test_datasource/test_azure.py +++ b/tests/unittests/test_datasource/test_azure.py @@ -1,5 +1,5 @@ from cloudinit import helpers -from cloudinit.util import load_file +from cloudinit.util import b64e, load_file from cloudinit.sources import DataSourceAzure from ..helpers import TestCase, populate_dir @@ -12,7 +12,6 @@ try: except ImportError: from contextlib2 import ExitStack -import base64 import crypt import os import stat @@ -22,13 +21,6 @@ import tempfile import unittest -def b64(source): - # In Python 3, b64encode only accepts bytes and returns bytes. - if not isinstance(source, bytes): - source = source.encode('utf-8') - return base64.b64encode(source).decode('us-ascii') - - def construct_valid_ovf_env(data=None, pubkeys=None, userdata=None): if data is None: data = {'HostName': 'FOOHOST'} @@ -58,7 +50,7 @@ def construct_valid_ovf_env(data=None, pubkeys=None, userdata=None): content += "<%s%s>%s\n" % (key, attrs, val, key) if userdata: - content += "%s\n" % (b64(userdata)) + content += "%s\n" % (b64e(userdata)) if pubkeys: content += "\n" @@ -189,7 +181,7 @@ class TestAzureDataSource(TestCase): # set dscfg in via base64 encoded yaml cfg = {'agent_command': "my_command"} odata = {'HostName': "myhost", 'UserName': "myuser", - 'dscfg': {'text': b64(yaml.dump(cfg)), + 'dscfg': {'text': b64e(yaml.dump(cfg)), 'encoding': 'base64'}} data = {'ovfcontent': construct_valid_ovf_env(data=odata)} @@ -241,7 +233,7 @@ class TestAzureDataSource(TestCase): def test_userdata_found(self): mydata = "FOOBAR" - odata = {'UserData': b64(mydata)} + odata = {'UserData': b64e(mydata)} data = {'ovfcontent': construct_valid_ovf_env(data=odata)} dsrc = self._get_ds(data) @@ -289,7 +281,7 @@ class TestAzureDataSource(TestCase): 'command': 'my-bounce-command', 'hostname_command': 'my-hostname-command'}} odata = {'HostName': "xhost", - 'dscfg': {'text': b64(yaml.dump(cfg)), + 'dscfg': {'text': b64e(yaml.dump(cfg)), 'encoding': 'base64'}} data = {'ovfcontent': construct_valid_ovf_env(data=odata)} self._get_ds(data).get_data() @@ -304,7 +296,7 @@ class TestAzureDataSource(TestCase): # config specifying set_hostname off should not bounce cfg = {'set_hostname': False} odata = {'HostName': "xhost", - 'dscfg': {'text': b64(yaml.dump(cfg)), + 'dscfg': {'text': b64e(yaml.dump(cfg)), 'encoding': 'base64'}} data = {'ovfcontent': construct_valid_ovf_env(data=odata)} self._get_ds(data).get_data() @@ -333,7 +325,7 @@ class TestAzureDataSource(TestCase): # Make sure that user can affect disk aliases dscfg = {'disk_aliases': {'ephemeral0': '/dev/sdc'}} odata = {'HostName': "myhost", 'UserName': "myuser", - 'dscfg': {'text': b64(yaml.dump(dscfg)), + 'dscfg': {'text': b64e(yaml.dump(dscfg)), 'encoding': 'base64'}} usercfg = {'disk_setup': {'/dev/sdc': {'something': '...'}, 'ephemeral0': False}} @@ -370,7 +362,7 @@ class TestAzureDataSource(TestCase): def test_existing_ovf_same(self): # waagent/SharedConfig left alone if found ovf-env.xml same as cached - odata = {'UserData': b64("SOMEUSERDATA")} + odata = {'UserData': b64e("SOMEUSERDATA")} data = {'ovfcontent': construct_valid_ovf_env(data=odata)} populate_dir(self.waagent_d, @@ -394,9 +386,9 @@ class TestAzureDataSource(TestCase): # 'get_data' should remove SharedConfig.xml in /var/lib/waagent # if ovf-env.xml differs. cached_ovfenv = construct_valid_ovf_env( - {'userdata': b64("FOO_USERDATA")}) + {'userdata': b64e("FOO_USERDATA")}) new_ovfenv = construct_valid_ovf_env( - {'userdata': b64("NEW_USERDATA")}) + {'userdata': b64e("NEW_USERDATA")}) populate_dir(self.waagent_d, {'ovf-env.xml': cached_ovfenv, diff --git a/tests/unittests/test_datasource/test_opennebula.py b/tests/unittests/test_datasource/test_opennebula.py index e5a4bd18..27adf21b 100644 --- a/tests/unittests/test_datasource/test_opennebula.py +++ b/tests/unittests/test_datasource/test_opennebula.py @@ -3,19 +3,12 @@ from cloudinit.sources import DataSourceOpenNebula as ds from cloudinit import util from ..helpers import TestCase, populate_dir -from base64 import b64encode import os import pwd import shutil import tempfile import unittest -def b64(source): - # In Python 3, b64encode only accepts bytes and returns bytes. - if not isinstance(source, bytes): - source = source.encode('utf-8') - return b64encode(source).decode('us-ascii') - TEST_VARS = { 'VAR1': 'single', @@ -186,7 +179,7 @@ class TestOpenNebulaDataSource(TestCase): self.assertEqual(USER_DATA, results['userdata']) def test_user_data_encoding_required_for_decode(self): - b64userdata = b64(USER_DATA) + b64userdata = util.b64e(USER_DATA) for k in ('USER_DATA', 'USERDATA'): my_d = os.path.join(self.tmp, k) populate_context_dir(my_d, {k: b64userdata}) @@ -198,7 +191,7 @@ class TestOpenNebulaDataSource(TestCase): def test_user_data_base64_encoding(self): for k in ('USER_DATA', 'USERDATA'): my_d = os.path.join(self.tmp, k) - populate_context_dir(my_d, {k: b64(USER_DATA), + populate_context_dir(my_d, {k: util.b64e(USER_DATA), 'USERDATA_ENCODING': 'base64'}) results = ds.read_context_disk_dir(my_d) diff --git a/tests/unittests/test_datasource/test_smartos.py b/tests/unittests/test_datasource/test_smartos.py index b5ebf94d..8b62b1b1 100644 --- a/tests/unittests/test_datasource/test_smartos.py +++ b/tests/unittests/test_datasource/test_smartos.py @@ -24,9 +24,9 @@ from __future__ import print_function -import base64 from cloudinit import helpers as c_helpers from cloudinit.sources import DataSourceSmartOS +from cloudinit.util import b64e from .. import helpers import os import os.path @@ -36,12 +36,6 @@ import tempfile import stat import uuid -def b64(source): - # In Python 3, b64encode only accepts bytes and returns bytes. - if not isinstance(source, bytes): - source = source.encode('utf-8') - return base64.b64encode(source).decode('us-ascii') - MOCK_RETURNS = { 'hostname': 'test-host', @@ -239,7 +233,7 @@ class TestSmartOSDataSource(helpers.FilesystemMockingTestCase): my_returns = MOCK_RETURNS.copy() my_returns['base64_all'] = "true" for k in ('hostname', 'cloud-init:user-data'): - my_returns[k] = b64(my_returns[k]) + my_returns[k] = b64e(my_returns[k]) dsrc = self._get_ds(mockdata=my_returns) ret = dsrc.get_data() @@ -260,7 +254,7 @@ class TestSmartOSDataSource(helpers.FilesystemMockingTestCase): my_returns['b64-cloud-init:user-data'] = "true" my_returns['b64-hostname'] = "true" for k in ('hostname', 'cloud-init:user-data'): - my_returns[k] = b64(my_returns[k]) + my_returns[k] = b64e(my_returns[k]) dsrc = self._get_ds(mockdata=my_returns) ret = dsrc.get_data() @@ -276,7 +270,7 @@ class TestSmartOSDataSource(helpers.FilesystemMockingTestCase): my_returns = MOCK_RETURNS.copy() my_returns['base64_keys'] = 'hostname,ignored' for k in ('hostname',): - my_returns[k] = b64(my_returns[k]) + my_returns[k] = b64e(my_returns[k]) dsrc = self._get_ds(mockdata=my_returns) ret = dsrc.get_data() diff --git a/tests/unittests/test_handler/test_handler_seed_random.py b/tests/unittests/test_handler/test_handler_seed_random.py index d3f18fa0..0bcdcb31 100644 --- a/tests/unittests/test_handler/test_handler_seed_random.py +++ b/tests/unittests/test_handler/test_handler_seed_random.py @@ -18,7 +18,6 @@ from cloudinit.config import cc_seed_random -import base64 import gzip import tempfile @@ -38,13 +37,6 @@ import logging LOG = logging.getLogger(__name__) -def b64(source): - # In Python 3, b64encode only accepts bytes and returns bytes. - if not isinstance(source, bytes): - source = source.encode('utf-8') - return base64.b64encode(source).decode('us-ascii') - - class TestRandomSeed(t_help.TestCase): def setUp(self): super(TestRandomSeed, self).setUp() @@ -141,7 +133,7 @@ class TestRandomSeed(t_help.TestCase): self.assertEquals("big-toe", contents) def test_append_random_base64(self): - data = b64('bubbles') + data = util.b64e('bubbles') cfg = { 'random_seed': { 'file': self._seed_file, @@ -154,7 +146,7 @@ class TestRandomSeed(t_help.TestCase): self.assertEquals("bubbles", contents) def test_append_random_b64(self): - data = b64('kit-kat') + data = util.b64e('kit-kat') cfg = { 'random_seed': { 'file': self._seed_file, -- cgit v1.2.3