Add test verification to ansible-test. (#22636)

* Add unified git diff parser.
* Add metadata and diff handling.
* Add test confidence/verification to bot output.
pull/22384/head
Matt Clay 7 years ago committed by GitHub
parent 5e9a2b8528
commit 869449e288

@ -0,0 +1,12 @@
.PHONY: all compile sanity units
all: compile sanity units
compile:
./ansible-test compile test/runner/ ${TEST_FLAGS}
sanity:
./ansible-test sanity test/runner/ ${TEST_FLAGS}
units:
PYTHONPATH=. pytest units ${TEST_FLAGS}

@ -69,15 +69,18 @@ class ShippableChanges(object):
if self.is_pr:
self.paths = sorted(git.get_diff_names([self.branch]))
self.diff = git.get_diff([self.branch])
else:
merge_runs = self.get_merge_runs(self.project_id, self.branch)
last_successful_commit = self.get_last_successful_commit(git, merge_runs)
if last_successful_commit:
self.paths = sorted(git.get_diff_names([last_successful_commit, self.commit]))
self.diff = git.get_diff([last_successful_commit, self.commit])
else:
# tracked files (including unchanged)
self.paths = sorted(git.get_file_names(['--cached']))
self.diff = None
def get_merge_runs(self, project_id, branch):
"""
@ -161,6 +164,8 @@ class LocalChanges(object):
self.staged = sorted(git.get_diff_names(['--cached']))
# tracked changes (including deletions) which are not staged
self.unstaged = sorted(git.get_diff_names([]))
# diff of all tracked files from fork point to working copy
self.diff = git.get_diff([self.fork_point])
@staticmethod
def is_official_branch(name):

@ -4,6 +4,7 @@ from __future__ import absolute_import, print_function
import os
import sys
import tempfile
import time
import lib.pytar
@ -46,6 +47,27 @@ def delegate(args, exclude, require):
:type args: EnvironmentConfig
:type exclude: list[str]
:type require: list[str]
:rtype: bool
"""
if isinstance(args, TestConfig):
with tempfile.NamedTemporaryFile(prefix='metadata-', suffix='.json', dir=os.getcwd()) as metadata_fd:
args.metadata_path = os.path.basename(metadata_fd.name)
args.metadata.to_file(args.metadata_path)
try:
return delegate_command(args, exclude, require)
finally:
args.metadata_path = None
else:
return delegate_command(args, exclude, require)
def delegate_command(args, exclude, require):
"""
:type args: EnvironmentConfig
:type exclude: list[str]
:type require: list[str]
:rtype: bool
"""
if args.tox:
delegate_tox(args, exclude, require)
@ -396,6 +418,7 @@ def filter_options(args, argv, options, exclude, require):
'--ignore-unstaged': 0,
'--changed-from': 1,
'--changed-path': 1,
'--metadata': 1,
})
elif isinstance(args, SanityConfig):
options.update({
@ -427,3 +450,7 @@ def filter_options(args, argv, options, exclude, require):
for target in require:
yield '--require'
yield target
if args.metadata_path:
yield '--metadata'
yield args.metadata_path

@ -0,0 +1,253 @@
"""Diff parsing functions and classes."""
from __future__ import absolute_import, print_function
import re
import textwrap
import traceback
from lib.util import (
ApplicationError,
)
def parse_diff(lines):
"""
:type lines: list[str]
:rtype: list[FileDiff]
"""
return DiffParser(lines).files
class FileDiff(object):
"""Parsed diff for a single file."""
def __init__(self, old_path, new_path):
"""
:type old_path: str
:type new_path: str
"""
self.old = DiffSide(old_path, new=False)
self.new = DiffSide(new_path, new=True)
self.headers = [] # list [str]
self.binary = False
def append_header(self, line):
"""
:type line: str
"""
self.headers.append(line)
@property
def is_complete(self):
"""
:rtype: bool
"""
return self.old.is_complete and self.new.is_complete
class DiffSide(object):
"""Parsed diff for a single 'side' of a single file."""
def __init__(self, path, new):
"""
:type path: str
:type new: bool
"""
self.path = path
self.new = new
self.prefix = '+' if self.new else '-'
self.eof_newline = True
self.exists = True
self.lines = [] # type: list [tuple[int, str]]
self.lines_and_context = [] # type: list [tuple[int, str]]
self.ranges = [] # type: list [tuple[int, int]]
self._next_line_number = 0
self._lines_remaining = 0
self._range_start = 0
def set_start(self, line_start, line_count):
"""
:type line_start: int
:type line_count: int
"""
self._next_line_number = line_start
self._lines_remaining = line_count
self._range_start = 0
def append(self, line):
"""
:type line: str
"""
if self._lines_remaining <= 0:
raise Exception('Diff range overflow.')
entry = self._next_line_number, line
if line.startswith(' '):
pass
elif line.startswith(self.prefix):
self.lines.append(entry)
if not self._range_start:
self._range_start = self._next_line_number
else:
raise Exception('Unexpected diff content prefix.')
self.lines_and_context.append(entry)
self._lines_remaining -= 1
if self._range_start:
if self.is_complete:
range_end = self._next_line_number
elif line.startswith(' '):
range_end = self._next_line_number - 1
else:
range_end = 0
if range_end:
self.ranges.append((self._range_start, range_end))
self._range_start = 0
self._next_line_number += 1
@property
def is_complete(self):
"""
:rtype: bool
"""
return self._lines_remaining == 0
def format_lines(self, context=True):
"""
:type context: bool
:rtype: list[str]
"""
if context:
lines = self.lines_and_context
else:
lines = self.lines
return ['%s:%4d %s' % (self.path, line[0], line[1]) for line in lines]
class DiffParser(object):
"""Parse diff lines."""
def __init__(self, lines):
"""
:type lines: list[str]
"""
self.lines = lines
self.files = [] # type: list [FileDiff]
self.action = self.process_start
self.line_number = 0
self.previous_line = None # type: str
self.line = None # type: str
self.file = None # type: FileDiff
for self.line in self.lines:
self.line_number += 1
try:
self.action()
except Exception as ex:
message = textwrap.dedent('''
%s
Line: %d
Previous: %s
Current: %s
%s
''').strip() % (
ex,
self.line_number,
self.previous_line or '',
self.line or '',
traceback.format_exc(),
)
raise ApplicationError(message.strip())
self.previous_line = self.line
self.complete_file()
def process_start(self):
"""Process a diff start line."""
self.complete_file()
match = re.search(r'^diff --git a/(?P<old_path>.*) b/(?P<new_path>.*)$', self.line)
if not match:
raise Exception('Unexpected diff start line.')
self.file = FileDiff(match.group('old_path'), match.group('new_path'))
self.action = self.process_continue
def process_range(self):
"""Process a diff range line."""
match = re.search(r'^@@ -((?P<old_start>[0-9]+),)?(?P<old_count>[0-9]+) \+((?P<new_start>[0-9]+),)?(?P<new_count>[0-9]+) @@', self.line)
if not match:
raise Exception('Unexpected diff range line.')
self.file.old.set_start(int(match.group('old_start') or 1), int(match.group('old_count')))
self.file.new.set_start(int(match.group('new_start') or 1), int(match.group('new_count')))
self.action = self.process_content
def process_continue(self):
"""Process a diff start, range or header line."""
if self.line.startswith('diff '):
self.process_start()
elif self.line.startswith('@@ '):
self.process_range()
else:
self.process_header()
def process_header(self):
"""Process a diff header line."""
if self.line.startswith('Binary files '):
self.file.binary = True
elif self.line == '--- /dev/null':
self.file.old.exists = False
elif self.line == '+++ /dev/null':
self.file.new.exists = False
else:
self.file.append_header(self.line)
def process_content(self):
"""Process a diff content line."""
if self.line == r'\ No newline at end of file':
if self.previous_line.startswith(' '):
self.file.old.eof_newline = False
self.file.new.eof_newline = False
elif self.previous_line.startswith('-'):
self.file.old.eof_newline = False
elif self.previous_line.startswith('+'):
self.file.new.eof_newline = False
else:
raise Exception('Unexpected previous diff content line.')
return
if self.file.is_complete:
self.process_continue()
return
if self.line.startswith(' '):
self.file.old.append(self.line)
self.file.new.append(self.line)
elif self.line.startswith('-'):
self.file.old.append(self.line)
elif self.line.startswith('+'):
self.file.new.append(self.line)
else:
raise Exception('Unexpected diff content line.')
def complete_file(self):
"""Complete processing of the current file, if any."""
if not self.file:
return
self.files.append(self.file)

@ -41,6 +41,7 @@ from lib.util import (
remove_tree,
make_dirs,
is_shippable,
is_binary_file,
)
from lib.test import (
@ -82,6 +83,10 @@ from lib.test import (
TestSkipped,
)
from lib.metadata import (
Metadata,
)
SUPPORTED_PYTHON_VERSIONS = (
'2.6',
'2.7',
@ -919,7 +924,7 @@ def detect_changes(args):
def detect_changes_shippable(args):
"""Initialize change detection on Shippable.
:type args: CommonConfig
:type args: TestConfig
:rtype: list[str]
"""
git = Git(args)
@ -934,6 +939,9 @@ def detect_changes_shippable(args):
display.info('Processing %s for branch %s commit %s' % (job_type, result.branch, result.commit))
if not args.metadata.changes:
args.metadata.populate_changes(result.diff)
return result.paths
@ -977,6 +985,19 @@ def detect_changes_local(args):
if args.unstaged:
names |= set(result.unstaged)
if not args.metadata.changes:
args.metadata.populate_changes(result.diff)
for path in result.untracked:
if is_binary_file(path):
args.metadata.changes[path] = ((0, 0),)
continue
with open(path, 'r') as source_fd:
line_count = len(source_fd.read().splitlines())
args.metadata.changes[path] = ((1, line_count),)
return sorted(names)

@ -18,6 +18,14 @@ class Git(object):
self.args = args
self.git = 'git'
def get_diff(self, args):
"""
:type args: list[str]
:rtype: list[str]
"""
cmd = ['diff'] + args
return self.run_git_split(cmd, '\n')
def get_diff_names(self, args):
"""
:type args: list[str]

@ -0,0 +1,82 @@
"""Test metadata for passing data to delegated tests."""
from __future__ import absolute_import, print_function
import json
from lib.util import (
display,
)
from lib.diff import (
parse_diff,
FileDiff,
)
class Metadata(object):
"""Metadata object for passing data to delegated tests."""
def __init__(self):
"""Initialize metadata."""
self.changes = {} # type: dict [str, tuple[tuple[int, int]]
def populate_changes(self, diff):
"""
:type diff: list[str] | None
"""
patches = parse_diff(diff)
patches = sorted(patches, key=lambda k: k.new.path) # type: list [FileDiff]
self.changes = dict((patch.new.path, tuple(patch.new.ranges)) for patch in patches)
renames = [patch.old.path for patch in patches if patch.old.path != patch.new.path and patch.old.exists and patch.new.exists]
deletes = [patch.old.path for patch in patches if not patch.new.exists]
# make sure old paths which were renamed or deleted are registered in changes
for path in renames + deletes:
if path in self.changes:
# old path was replaced with another file
continue
# failed tests involving deleted files should be using line 0 since there is no content remaining
self.changes[path] = ((0, 0),)
def to_dict(self):
"""
:rtype: dict[str, any]
"""
return dict(
changes=self.changes,
)
def to_file(self, path):
"""
:type path: path
"""
data = self.to_dict()
display.info('>>> Metadata: %s\n%s' % (path, data), verbosity=3)
with open(path, 'w') as data_fd:
json.dump(data, data_fd, sort_keys=True, indent=4)
@staticmethod
def from_file(path):
"""
:type path: str
:rtype: Metadata
"""
with open(path, 'r') as data_fd:
data = json.load(data_fd)
return Metadata.from_dict(data)
@staticmethod
def from_dict(data):
"""
:type data: dict[str, any]
:rtype: Metadata
"""
metadata = Metadata()
metadata.changes = data['changes']
return metadata

@ -45,6 +45,7 @@ from lib.test import (
TestFailure,
TestSkipped,
TestMessage,
calculate_best_confidence,
)
COMMAND = 'sanity'
@ -364,6 +365,7 @@ def command_sanity_pep8(args, targets):
path=PEP8_LEGACY_PATH,
line=line,
column=1,
confidence=calculate_best_confidence(((PEP8_LEGACY_PATH, line), (path, 0)), args.metadata) if args.metadata.changes else None,
))
if path in used_paths and path not in failed_result_paths:
@ -374,6 +376,7 @@ def command_sanity_pep8(args, targets):
path=PEP8_LEGACY_PATH,
line=line,
column=1,
confidence=calculate_best_confidence(((PEP8_LEGACY_PATH, line), (path, 0)), args.metadata) if args.metadata.changes else None,
))
line = 0
@ -389,6 +392,7 @@ def command_sanity_pep8(args, targets):
path=PEP8_SKIP_PATH,
line=line,
column=1,
confidence=calculate_best_confidence(((PEP8_SKIP_PATH, line), (path, 0)), args.metadata) if args.metadata.changes else None,
))
for result in results:
@ -586,7 +590,7 @@ class SanityFailure(TestFailure):
class SanityMessage(TestMessage):
"""Single sanity test message for one file."""
def __init__(self, message, path, line=0, column=0, level='error', code=None):
def __init__(self, message, path, line=0, column=0, level='error', code=None, confidence=None):
"""
:type message: str
:type path: str
@ -594,8 +598,9 @@ class SanityMessage(TestMessage):
:type column: int
:type level: str
:type code: str | None
:type confidence: int | None
"""
super(SanityMessage, self).__init__(message, path, line, column, level, code)
super(SanityMessage, self).__init__(message, path, line, column, level, code, confidence)
class SanityTargets(object):

@ -10,6 +10,50 @@ from lib.util import (
EnvironmentConfig,
)
from lib.metadata import (
Metadata,
)
def calculate_best_confidence(choices, metadata):
"""
:type choices: tuple[tuple[str, int]]
:type metadata: Metadata
:rtype: int
"""
best_confidence = 0
for path, line in choices:
confidence = calculate_confidence(path, line, metadata)
best_confidence = max(confidence, best_confidence)
return best_confidence
def calculate_confidence(path, line, metadata):
"""
:type path: str
:type line: int
:type metadata: Metadata
:rtype: int
"""
ranges = metadata.changes.get(path)
# no changes were made to the file
if not ranges:
return 0
# changes were made to the same file and line
if any(r[0] <= line <= r[1] in r for r in ranges):
return 100
# changes were made to the same file and the line number is unknown
if line == 0:
return 75
# changes were made to the same file and the line number is different
return 50
class TestConfig(EnvironmentConfig):
"""Configuration common to all test commands."""
@ -38,6 +82,9 @@ class TestConfig(EnvironmentConfig):
self.junit = args.junit if 'junit' in args else False # type: bool
self.failure_ok = args.failure_ok if 'failure_ok' in args else False # type: bool
self.metadata = Metadata.from_file(args.metadata) if args.metadata else Metadata()
self.metadata_path = None
class TestResult(object):
"""Base class for test results."""
@ -201,9 +248,18 @@ class TestFailure(TestResult):
if messages:
messages = sorted(messages, key=lambda m: m.sort_key)
self.messages = messages
self.messages = messages or []
self.summary = summary
def write(self, args):
"""
:type args: TestConfig
"""
if args.metadata.changes:
self.populate_confidence(args.metadata)
super(TestFailure, self).write(args)
def write_console(self):
"""Write results to console."""
if self.summary:
@ -217,7 +273,7 @@ class TestFailure(TestResult):
display.error('Found %d %s issue(s)%s which need to be resolved:' % (len(self.messages), self.test or self.command, specifier))
for message in self.messages:
display.error(message)
display.error(message.format(show_confidence=True))
def write_lint(self):
"""Write lint results to stdout."""
@ -253,7 +309,13 @@ class TestFailure(TestResult):
message = self.format_title()
output = self.format_block()
if self.messages:
verified = all((m.confidence or 0) >= 50 for m in self.messages)
else:
verified = False
bot_data = dict(
verified=verified,
results=[
dict(
message=message,
@ -271,6 +333,14 @@ class TestFailure(TestResult):
json.dump(bot_data, bot_fd, indent=4, sort_keys=True)
bot_fd.write('\n')
def populate_confidence(self, metadata):
"""
:type metadata: Metadata
"""
for message in self.messages:
if message.confidence is None:
message.confidence = calculate_confidence(message.path, message.line, metadata)
def format_command(self):
"""
:rtype: str
@ -319,7 +389,7 @@ class TestFailure(TestResult):
class TestMessage(object):
"""Single test message for one file."""
def __init__(self, message, path, line=0, column=0, level='error', code=None):
def __init__(self, message, path, line=0, column=0, level='error', code=None, confidence=None):
"""
:type message: str
:type path: str
@ -327,6 +397,7 @@ class TestMessage(object):
:type column: int
:type level: str
:type code: str | None
:type confidence: int | None
"""
self.path = path
self.line = line
@ -334,13 +405,24 @@ class TestMessage(object):
self.level = level
self.code = code
self.message = message
self.confidence = confidence
def __str__(self):
return self.format()
def format(self, show_confidence=False):
"""
:type show_confidence: bool
:rtype: str
"""
if self.code:
msg = '%s %s' % (self.code, self.message)
else:
msg = self.message
if show_confidence and self.confidence is not None:
msg += ' (%d%%)' % self.confidence
return '%s:%s:%s: %s' % (self.path, self.line, self.column, msg)
@property

@ -165,10 +165,14 @@ def raw_command(cmd, capture=False, env=None, data=None, cwd=None, explain=False
raise
if communicate:
stdout, stderr = process.communicate(data)
encoding = 'utf-8'
data_bytes = data.encode(encoding) if data else None
stdout_bytes, stderr_bytes = process.communicate(data_bytes)
stdout_text = stdout_bytes.decode(encoding) if stdout_bytes else u''
stderr_text = stderr_bytes.decode(encoding) if stderr_bytes else u''
else:
process.wait()
stdout, stderr = None, None
stdout_text, stderr_text = None, None
status = process.returncode
runtime = time.time() - start
@ -176,9 +180,9 @@ def raw_command(cmd, capture=False, env=None, data=None, cwd=None, explain=False
display.info('Command exited with status %s after %s seconds.' % (status, runtime), verbosity=4)
if status == 0:
return stdout, stderr
return stdout_text, stderr_text
raise SubprocessError(cmd, status, stdout, stderr, runtime)
raise SubprocessError(cmd, status, stdout_text, stderr_text, runtime)
def common_environment():
@ -266,6 +270,15 @@ def make_dirs(path):
raise
def is_binary_file(path):
"""
:type path: str
:rtype: bool
"""
with open(path, 'rb') as path_fd:
return b'\0' in path_fd.read(1024)
class Display(object):
"""Manages color console output."""
clear = '\033[0m'

@ -2,5 +2,6 @@ jinja2
mock
pep8
pylint
pytest
voluptuous
yamllint

@ -163,6 +163,9 @@ def parse_args():
action='store_true',
help='analyze code coverage when running tests')
test.add_argument('--metadata',
help=argparse.SUPPRESS)
add_changes(test, argparse)
add_environments(test)

@ -0,0 +1,97 @@
"""Tests for diff module."""
import os
import subprocess
import pytest
from lib.diff import (
parse_diff,
FileDiff,
)
def get_diff(base, head=None):
"""Return a git diff between the base and head revision.
:type base: str
:type head: str | None
:rtype: list[str]
"""
if not head or head == 'HEAD':
head = subprocess.check_output(['git', 'rev-parse', 'HEAD']).strip()
cache = '/tmp/git-diff-cache-%s-%s.log' % (base, head)
if os.path.exists(cache):
with open(cache, 'r') as cache_fd:
lines = cache_fd.read().splitlines()
else:
lines = subprocess.check_output(['git', 'diff', base, head]).splitlines()
with open(cache, 'w') as cache_fd:
cache_fd.write('\n'.join(lines))
assert lines
return lines
def get_parsed_diff(base, head=None):
"""Return a parsed git diff between the base and head revision.
:type base: str
:type head: str | None
:rtype: list[FileDiff]
"""
lines = get_diff(base, head)
items = parse_diff(lines)
assert items
for item in items:
assert item.headers
assert item.is_complete
item.old.format_lines()
item.new.format_lines()
for line_range in item.old.ranges:
assert line_range[1] >= line_range[0] > 0
for line_range in item.new.ranges:
assert line_range[1] >= line_range[0] > 0
return items
RANGES_TO_TEST = (
('f31421576b00f0b167cdbe61217c31c21a41ac02', 'HEAD'),
('b8125ac1a61f2c7d1de821c78c884560071895f1', '32146acf4e43e6f95f54d9179bf01f0df9814217')
)
@pytest.mark.parametrize("base, head", RANGES_TO_TEST)
def test_parse_diff(base, head):
"""Integration test to verify parsing of ansible/ansible history."""
get_parsed_diff(base, head)
def test_parse_delete():
"""Integration test to verify parsing of a deleted file."""
commit = 'ee17b914554861470b382e9e80a8e934063e0860'
items = get_parsed_diff(commit + '~', commit)
deletes = [item for item in items if not item.new.exists]
assert len(deletes) == 1
assert deletes[0].old.path == 'lib/ansible/plugins/connection/nspawn.py'
assert deletes[0].new.path == 'lib/ansible/plugins/connection/nspawn.py'
def test_parse_rename():
"""Integration test to verify parsing of renamed files."""
commit = '16a39639f568f4dd5cb233df2d0631bdab3a05e9'
items = get_parsed_diff(commit + '~', commit)
renames = [item for item in items if item.old.path != item.new.path and item.old.exists and item.new.exists]
assert len(renames) == 2
assert renames[0].old.path == 'test/integration/targets/eos_eapi/tests/cli/badtransport.yaml'
assert renames[0].new.path == 'test/integration/targets/eos_eapi/tests/cli/badtransport.1'
assert renames[1].old.path == 'test/integration/targets/eos_eapi/tests/cli/zzz_reset.yaml'
assert renames[1].new.path == 'test/integration/targets/eos_eapi/tests/cli/zzz_reset.1'

@ -17,10 +17,14 @@ ln -sf x86_64-linux-gnu-gcc-4.9 /usr/bin/x86_64-linux-gnu-gcc
retry.py pip install tox --disable-pip-version-check
echo '{"verified": false, "results": []}' > test/results/bot/ansible-test-failure.json
ansible-test compile --failure-ok --color -v --junit --requirements
ansible-test sanity --failure-ok --color -v --junit --tox --skip-test ansible-doc --python 2.7
ansible-test sanity --failure-ok --color -v --junit --tox --test ansible-doc --coverage
rm test/results/bot/ansible-test-failure.json
if find test/results/bot/ -mindepth 1 -name '.*' -prune -o -print -quit | grep -q .; then
echo "One or more of the above ansible-test commands recorded at least one test failure."
exit 1

Loading…
Cancel
Save