salt/tests/unit/serializers/test_serializers.py
Erik Johnson 3184168365 Use explicit unicode strings + break up salt.utils
This PR is part of what will be an ongoing effort to use explicit
unicode strings in Salt. Because Python 3 does not suport Python 2's raw
unicode string syntax (i.e. `ur'\d+'`), we must use
`salt.utils.locales.sdecode()` to ensure that the raw string is unicode.
However, because of how `salt/utils/__init__.py` has evolved into the
hulking monstrosity it is today, this means importing a large module in
places where it is not needed, which could negatively impact
performance. For this reason, this PR also breaks out some of the
functions from `salt/utils/__init__.py` into new/existing modules under
`salt/utils/`. The long term goal will be that the modules within this
directory do not depend on importing `salt.utils`.

A summary of the changes in this PR is as follows:

* Moves the following functions from `salt.utils` to new locations
  (including a deprecation warning if invoked from `salt.utils`):
  `to_bytes`, `to_str`, `to_unicode`, `str_to_num`, `is_quoted`,
  `dequote`, `is_hex`, `is_bin_str`, `rand_string`,
  `contains_whitespace`, `clean_kwargs`, `invalid_kwargs`, `which`,
  `which_bin`, `path_join`, `shlex_split`, `rand_str`, `is_windows`,
  `is_proxy`, `is_linux`, `is_darwin`, `is_sunos`, `is_smartos`,
  `is_smartos_globalzone`, `is_smartos_zone`, `is_freebsd`, `is_netbsd`,
  `is_openbsd`, `is_aix`
* Moves the functions already deprecated by @rallytime to the bottom of
  `salt/utils/__init__.py` for better organization, so we can keep the
  deprecated ones separate from the ones yet to be deprecated as we
  continue to break up `salt.utils`
* Updates `salt/*.py` and all files under `salt/client/` to use explicit
  unicode string literals.
* Gets rid of implicit imports of `salt.utils` (e.g. `from salt.utils
  import foo` becomes `import salt.utils.foo as foo`).
* Renames the `test.rand_str` function to `test.random_hash` to more
  accurately reflect what it does
* Modifies `salt.utils.stringutils.random()` (née `salt.utils.rand_string()`)
  such that it returns a string matching the passed size. Previously
  this function would get `size` bytes from `os.urandom()`,
  base64-encode it, and return the result, which would in most cases not
  be equal to the passed size.
2017-08-08 13:33:43 -05:00

349 lines
11 KiB
Python

# -*- coding: utf-8 -*-
# Import python libs
from __future__ import absolute_import
from textwrap import dedent
# Import Salt Testing libs
from tests.support.unit import skipIf, TestCase
# Import 3rd party libs
import jinja2
from salt.ext import six
# Import salt libs
import salt.serializers.configparser as configparser
import salt.serializers.json as json
import salt.serializers.yaml as yaml
import salt.serializers.yamlex as yamlex
import salt.serializers.msgpack as msgpack
import salt.serializers.python as python
from salt.serializers import SerializationError
from salt.utils.odict import OrderedDict
# Import test support libs
from tests.support.helpers import flaky
SKIP_MESSAGE = '%s is unavailable, do prerequisites have been met?'
@flaky(condition=six.PY3)
class TestSerializers(TestCase):
@skipIf(not json.available, SKIP_MESSAGE % 'json')
def test_serialize_json(self):
data = {
"foo": "bar"
}
serialized = json.serialize(data)
assert serialized == '{"foo": "bar"}', serialized
deserialized = json.deserialize(serialized)
assert deserialized == data, deserialized
@skipIf(not yaml.available, SKIP_MESSAGE % 'yaml')
def test_serialize_yaml(self):
data = {
"foo": "bar"
}
serialized = yaml.serialize(data)
assert serialized == '{foo: bar}', serialized
deserialized = yaml.deserialize(serialized)
assert deserialized == data, deserialized
@skipIf(not yaml.available, SKIP_MESSAGE % 'sls')
def test_serialize_sls(self):
data = {
"foo": "bar"
}
serialized = yamlex.serialize(data)
assert serialized == '{foo: bar}', serialized
deserialized = yamlex.deserialize(serialized)
assert deserialized == data, deserialized
@skipIf(not yamlex.available, SKIP_MESSAGE % 'sls')
def test_serialize_complex_sls(self):
data = OrderedDict([
("foo", 1),
("bar", 2),
("baz", True),
])
serialized = yamlex.serialize(data)
assert serialized == '{foo: 1, bar: 2, baz: true}', serialized
deserialized = yamlex.deserialize(serialized)
assert deserialized == data, deserialized
@skipIf(not yaml.available, SKIP_MESSAGE % 'yaml')
@skipIf(not yamlex.available, SKIP_MESSAGE % 'sls')
def test_compare_sls_vs_yaml(self):
src = '{foo: 1, bar: 2, baz: {qux: true}}'
sls_data = yamlex.deserialize(src)
yml_data = yaml.deserialize(src)
# ensure that sls & yaml have the same base
assert isinstance(sls_data, dict)
assert isinstance(yml_data, dict)
assert sls_data == yml_data
# ensure that sls is ordered, while yaml not
assert isinstance(sls_data, OrderedDict)
assert not isinstance(yml_data, OrderedDict)
@skipIf(not yaml.available, SKIP_MESSAGE % 'yaml')
@skipIf(not yamlex.available, SKIP_MESSAGE % 'sls')
def test_compare_sls_vs_yaml_with_jinja(self):
tpl = '{{ data }}'
env = jinja2.Environment()
src = '{foo: 1, bar: 2, baz: {qux: true}}'
sls_src = env.from_string(tpl).render(data=yamlex.deserialize(src))
yml_src = env.from_string(tpl).render(data=yaml.deserialize(src))
sls_data = yamlex.deserialize(sls_src)
yml_data = yaml.deserialize(yml_src)
# ensure that sls & yaml have the same base
assert isinstance(sls_data, dict)
assert isinstance(yml_data, dict)
# The below has been commented out because something the loader test
# is modifying the yaml renderer to render things to unicode. Without
# running the loader test, the below passes. Even reloading the module
# from disk does not reset its internal state (per the Python docs).
##
#assert sls_data == yml_data
# ensure that sls is ordered, while yaml not
assert isinstance(sls_data, OrderedDict)
assert not isinstance(yml_data, OrderedDict)
# prove that yaml does not handle well with OrderedDict
# while sls is jinja friendly.
obj = OrderedDict([
('foo', 1),
('bar', 2),
('baz', {'qux': True})
])
sls_obj = yamlex.deserialize(yamlex.serialize(obj))
try:
yml_obj = yaml.deserialize(yaml.serialize(obj))
except SerializationError:
# BLAAM! yaml was unable to serialize OrderedDict,
# but it's not the purpose of the current test.
yml_obj = obj.copy()
sls_src = env.from_string(tpl).render(data=sls_obj)
yml_src = env.from_string(tpl).render(data=yml_obj)
final_obj = yaml.deserialize(sls_src)
assert obj == final_obj
# BLAAM! yml_src is not valid !
final_obj = OrderedDict(yaml.deserialize(yml_src))
assert obj != final_obj
@skipIf(not yamlex.available, SKIP_MESSAGE % 'sls')
def test_sls_aggregate(self):
src = dedent("""
a: lol
foo: !aggregate hello
bar: !aggregate [1, 2, 3]
baz: !aggregate
a: 42
b: 666
c: the beast
""").strip()
# test that !aggregate is correctly parsed
sls_obj = yamlex.deserialize(src)
assert sls_obj == {
'a': 'lol',
'foo': ['hello'],
'bar': [1, 2, 3],
'baz': {
'a': 42,
'b': 666,
'c': 'the beast'
}
}, sls_obj
assert dedent("""
a: lol
foo: [hello]
bar: [1, 2, 3]
baz: {a: 42, b: 666, c: the beast}
""").strip() == yamlex.serialize(sls_obj), sls_obj
# test that !aggregate aggregates scalars
src = dedent("""
placeholder: !aggregate foo
placeholder: !aggregate bar
placeholder: !aggregate baz
""").strip()
sls_obj = yamlex.deserialize(src)
assert sls_obj == {'placeholder': ['foo', 'bar', 'baz']}, sls_obj
# test that !aggregate aggregates lists
src = dedent("""
placeholder: !aggregate foo
placeholder: !aggregate [bar, baz]
placeholder: !aggregate []
placeholder: !aggregate ~
""").strip()
sls_obj = yamlex.deserialize(src)
assert sls_obj == {'placeholder': ['foo', 'bar', 'baz']}, sls_obj
# test that !aggregate aggregates dicts
src = dedent("""
placeholder: !aggregate {foo: 42}
placeholder: !aggregate {bar: null}
placeholder: !aggregate {baz: inga}
""").strip()
sls_obj = yamlex.deserialize(src)
assert sls_obj == {
'placeholder': {
'foo': 42,
'bar': None,
'baz': 'inga'
}
}, sls_obj
# test that !aggregate aggregates deep dicts
src = dedent("""
placeholder: {foo: !aggregate {foo: 42}}
placeholder: {foo: !aggregate {bar: null}}
placeholder: {foo: !aggregate {baz: inga}}
""").strip()
sls_obj = yamlex.deserialize(src)
assert sls_obj == {
'placeholder': {
'foo': {
'foo': 42,
'bar': None,
'baz': 'inga'
}
}
}, sls_obj
# test that {foo: !aggregate bar} and {!aggregate foo: bar}
# are roughly equivalent.
src = dedent("""
placeholder: {!aggregate foo: {foo: 42}}
placeholder: {!aggregate foo: {bar: null}}
placeholder: {!aggregate foo: {baz: inga}}
""").strip()
sls_obj = yamlex.deserialize(src)
assert sls_obj == {
'placeholder': {
'foo': {
'foo': 42,
'bar': None,
'baz': 'inga'
}
}
}, sls_obj
@skipIf(not yamlex.available, SKIP_MESSAGE % 'sls')
def test_sls_reset(self):
src = dedent("""
placeholder: {!aggregate foo: {foo: 42}}
placeholder: {!aggregate foo: {bar: null}}
!reset placeholder: {!aggregate foo: {baz: inga}}
""").strip()
sls_obj = yamlex.deserialize(src)
assert sls_obj == {
'placeholder': {
'foo': {
'baz': 'inga'
}
}
}, sls_obj
@skipIf(not yamlex.available, SKIP_MESSAGE % 'sls')
def test_sls_repr(self):
"""
Ensure that obj __repr__ and __str__ methods are yaml friendly.
"""
def convert(obj):
return yamlex.deserialize(yamlex.serialize(obj))
sls_obj = convert(OrderedDict([('foo', 'bar'), ('baz', 'qux')]))
# ensure that repr and str are yaml friendly
assert sls_obj.__str__() == '{foo: bar, baz: qux}'
assert sls_obj.__repr__() == '{foo: bar, baz: qux}'
# ensure that repr and str are already quoted
assert sls_obj['foo'].__str__() == '"bar"'
assert sls_obj['foo'].__repr__() == '"bar"'
@skipIf(not yamlex.available, SKIP_MESSAGE % 'sls')
def test_sls_micking_file_merging(self):
def convert(obj):
return yamlex.deserialize(yamlex.serialize(obj))
# let say that we have 2 pillar files
src1 = dedent("""
a: first
b: !aggregate first
c:
subkey1: first
subkey2: !aggregate first
""").strip()
src2 = dedent("""
a: second
b: !aggregate second
c:
subkey2: !aggregate second
subkey3: second
""").strip()
sls_obj1 = yamlex.deserialize(src1)
sls_obj2 = yamlex.deserialize(src2)
sls_obj3 = yamlex.merge_recursive(sls_obj1, sls_obj2)
assert sls_obj3 == {
'a': 'second',
'b': ['first', 'second'],
'c': {
'subkey2': ['first', 'second'],
'subkey3': 'second'
}
}, sls_obj3
@skipIf(not msgpack.available, SKIP_MESSAGE % 'msgpack')
def test_msgpack(self):
data = OrderedDict([
("foo", 1),
("bar", 2),
("baz", True),
])
serialized = msgpack.serialize(data)
deserialized = msgpack.deserialize(serialized)
assert deserialized == data, deserialized
@skipIf(not python.available, SKIP_MESSAGE % 'python')
def test_serialize_python(self):
data = {'foo': 'bar'}
serialized = python.serialize(data)
assert serialized == '{\'foo\': \'bar\'}', serialized
@skipIf(not configparser.available, SKIP_MESSAGE % 'configparser')
def test_configparser(self):
data = {'foo': {'bar': 'baz'}}
# configparser appends empty lines
serialized = configparser.serialize(data).strip()
assert serialized == "[foo]\nbar = baz", serialized
deserialized = configparser.deserialize(serialized)
assert deserialized == data, deserialized