Back to Repositories

Testing Core Type Implementation and Command Processing in thefuck

This test suite validates core type implementations in The Fuck project, focusing on CorrectedCommand, Rule, and Command classes. It ensures proper handling of command corrections, rule processing, and command execution functionality.

Test Coverage Overview

The test suite provides comprehensive coverage of core type classes:
  • CorrectedCommand: Tests equality, hashability, string representation and command execution
  • Rule: Validates rule loading, matching, and command correction generation
  • Command: Verifies command creation from raw scripts and proper execution
Key edge cases include Unicode handling, environment variables, and error conditions.

Implementation Analysis

The testing approach uses pytest fixtures and parametrized tests to validate functionality systematically.
  • Mock objects simulate system interactions and command execution
  • Parametrized tests cover multiple input scenarios
  • Fixtures handle environment setup and teardown
Implementation leverages pytest’s powerful features for dependency injection and test parameterization.

Technical Details

Testing tools and configuration:
  • pytest as the primary testing framework
  • Mock library for test doubles
  • Custom fixtures for environment setup
  • Capsys for capturing stdout/stderr
  • Monkeypatch for modifying runtime behavior

Best Practices Demonstrated

The test suite exemplifies several testing best practices:
  • Isolated test cases with clear separation of concerns
  • Comprehensive assertion coverage
  • Proper test organization by class/functionality
  • Effective use of fixtures and mocks
  • Clear test naming and structure

nvbn/thefuck

tests/test_types.py

            
# -*- coding: utf-8 -*-

import os
from subprocess import PIPE, STDOUT
from mock import Mock
import pytest
from tests.utils import CorrectedCommand, Rule
from thefuck import const
from thefuck.exceptions import EmptyCommand
from thefuck.system import Path
from thefuck.types import Command


class TestCorrectedCommand(object):

    def test_equality(self):
        assert (CorrectedCommand('ls', None, 100) ==
                CorrectedCommand('ls', None, 200))
        assert (CorrectedCommand('ls', None, 100) !=
                CorrectedCommand('ls', lambda *_: _, 100))

    def test_hashable(self):
        assert {CorrectedCommand('ls', None, 100),
                CorrectedCommand('ls', None, 200)} == {CorrectedCommand('ls')}

    def test_representable(self):
        assert '{}'.format(CorrectedCommand('ls', None, 100)) == \
               'CorrectedCommand(script=ls, side_effect=None, priority=100)'
        assert u'{}'.format(CorrectedCommand(u'echo café', None, 100)) == \
               u'CorrectedCommand(script=echo café, side_effect=None, priority=100)'

    @pytest.mark.parametrize('script, printed, override_settings', [
        ('git branch', 'git branch', {'repeat': False, 'debug': False}),
        ('git brunch',
         "git brunch || fuck --repeat --force-command 'git brunch'",
         {'repeat': True, 'debug': False}),
        ('git brunch',
         "git brunch || fuck --repeat --debug --force-command 'git brunch'",
         {'repeat': True, 'debug': True})])
    def test_run(self, capsys, settings, script, printed, override_settings):
        settings.update(override_settings)
        CorrectedCommand(script, None, 1000).run(Command(script, ''))
        out, _ = capsys.readouterr()
        assert out == printed


class TestRule(object):
    def test_from_path_rule_exception(self, mocker):
        load_source = mocker.patch('thefuck.types.load_source',
                                   side_effect=ImportError("No module named foo..."))
        assert Rule.from_path(Path('git.py')) is None
        load_source.assert_called_once_with('git', 'git.py')

    def test_from_path(self, mocker):
        match = object()
        get_new_command = object()
        load_source = mocker.patch(
            'thefuck.types.load_source',
            return_value=Mock(match=match,
                              get_new_command=get_new_command,
                              enabled_by_default=True,
                              priority=900,
                              requires_output=True))
        rule_path = os.path.join(os.sep, 'rules', 'bash.py')
        assert (Rule.from_path(Path(rule_path))
                == Rule('bash', match, get_new_command, priority=900))
        load_source.assert_called_once_with('bash', rule_path)

    def test_from_path_excluded_rule(self, mocker, settings):
        load_source = mocker.patch('thefuck.types.load_source')
        settings.update(exclude_rules=['git'])
        rule_path = os.path.join(os.sep, 'rules', 'git.py')
        assert Rule.from_path(Path(rule_path)) is None
        assert not load_source.called

    @pytest.mark.parametrize('rules, rule, is_enabled', [
        (const.DEFAULT_RULES, Rule('git', enabled_by_default=True), True),
        (const.DEFAULT_RULES, Rule('git', enabled_by_default=False), False),
        ([], Rule('git', enabled_by_default=False), False),
        ([], Rule('git', enabled_by_default=True), False),
        (const.DEFAULT_RULES + ['git'], Rule('git', enabled_by_default=False), True),
        (['git'], Rule('git', enabled_by_default=False), True)])
    def test_is_enabled(self, settings, rules, rule, is_enabled):
        settings.update(rules=rules)
        assert rule.is_enabled == is_enabled

    def test_isnt_match(self):
        assert not Rule('', lambda _: False).is_match(
            Command('ls', ''))

    def test_is_match(self):
        rule = Rule('', lambda x: x.script == 'cd ..')
        assert rule.is_match(Command('cd ..', ''))

    @pytest.mark.usefixtures('no_colors')
    def test_isnt_match_when_rule_failed(self, capsys):
        rule = Rule('test', Mock(side_effect=OSError('Denied')),
                    requires_output=False)
        assert not rule.is_match(Command('ls', ''))
        assert capsys.readouterr()[1].split('
')[0] == '[WARN] Rule test:'

    def test_get_corrected_commands_with_rule_returns_list(self):
        rule = Rule(get_new_command=lambda x: [x.script + '!', x.script + '@'],
                    priority=100)
        assert (list(rule.get_corrected_commands(Command('test', '')))
                == [CorrectedCommand(script='test!', priority=100),
                    CorrectedCommand(script='test@', priority=200)])

    def test_get_corrected_commands_with_rule_returns_command(self):
        rule = Rule(get_new_command=lambda x: x.script + '!',
                    priority=100)
        assert (list(rule.get_corrected_commands(Command('test', '')))
                == [CorrectedCommand(script='test!', priority=100)])


class TestCommand(object):
    @pytest.fixture(autouse=True)
    def Popen(self, monkeypatch):
        Popen = Mock()
        Popen.return_value.stdout.read.return_value = b'output'
        monkeypatch.setattr('thefuck.output_readers.rerun.Popen', Popen)
        return Popen

    @pytest.fixture(autouse=True)
    def prepare(self, monkeypatch):
        monkeypatch.setattr('thefuck.output_readers.rerun._wait_output',
                            lambda *_: True)

    def test_from_script_calls(self, Popen, settings, os_environ):
        settings.env = {}
        assert Command.from_raw_script(
            ['apt-get', 'search', 'vim']) == Command(
            'apt-get search vim', 'output')
        Popen.assert_called_once_with('apt-get search vim',
                                      shell=True,
                                      stdin=PIPE,
                                      stdout=PIPE,
                                      stderr=STDOUT,
                                      env=os_environ)

    @pytest.mark.parametrize('script, result', [
        ([], None),
        ([''], None),
        (['', ''], None),
        (['ls', '-la'], 'ls -la'),
        (['ls'], 'ls'),
        (['echo \\ '], 'echo \\ '),
        (['echo \\
'], 'echo \\
')])
    def test_from_script(self, script, result):
        if result:
            assert Command.from_raw_script(script).script == result
        else:
            with pytest.raises(EmptyCommand):
                Command.from_raw_script(script)