# -*- coding: utf-8 -*- """ Test suite for the util module ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :copyright: Copyright 2006-2012 by the Pygments team, see AUTHORS. :license: BSD, see LICENSE for details. """ import re import unittest from pygments import util class FakeLexer(object): def analyse(text): return float(text) analyse = util.make_analysator(analyse) class UtilTest(unittest.TestCase): def test_getoptions(self): raises = self.assertRaises equals = self.assertEqual equals(util.get_bool_opt({}, 'a', True), True) equals(util.get_bool_opt({}, 'a', 1), True) equals(util.get_bool_opt({}, 'a', 'true'), True) equals(util.get_bool_opt({}, 'a', 'no'), False) raises(util.OptionError, util.get_bool_opt, {}, 'a', []) raises(util.OptionError, util.get_bool_opt, {}, 'a', 'foo') equals(util.get_int_opt({}, 'a', 1), 1) raises(util.OptionError, util.get_int_opt, {}, 'a', []) raises(util.OptionError, util.get_int_opt, {}, 'a', 'bar') equals(util.get_list_opt({}, 'a', [1]), [1]) equals(util.get_list_opt({}, 'a', '1 2'), ['1', '2']) raises(util.OptionError, util.get_list_opt, {}, 'a', 1) def test_docstring_headline(self): def f1(): """ docstring headline other text """ def f2(): """ docstring headline other text """ self.assertEqual(util.docstring_headline(f1), "docstring headline") self.assertEqual(util.docstring_headline(f2), "docstring headline") def test_analysator_returns_float(self): # If an analysator wrapped by make_analysator returns a floating point # number, then that number will be returned by the wrapper. self.assertEqual(FakeLexer.analyse('0.5'), 0.5) def test_analysator_returns_boolean(self): # If an analysator wrapped by make_analysator returns a boolean value, # then the wrapper will return 1.0 if the boolean was True or 0.0 if # it was False. self.assertEqual(FakeLexer.analyse(True), 1.0) self.assertEqual(FakeLexer.analyse(False), 0.0) def test_analysator_raises_exception(self): # If an analysator wrapped by make_analysator raises an exception, # then the wrapper will return 0.0. class ErrorLexer(object): def analyse(text): raise RuntimeError('something bad happened') analyse = util.make_analysator(analyse) self.assertEqual(ErrorLexer.analyse(''), 0.0) def test_analysator_value_error(self): # When converting the analysator's return value to a float a # ValueError may occur. If that happens 0.0 is returned instead. self.assertEqual(FakeLexer.analyse('bad input'), 0.0) def test_analysator_type_error(self): # When converting the analysator's return value to a float a # TypeError may occur. If that happens 0.0 is returned instead. self.assertEqual(FakeLexer.analyse(None), 0.0) def test_shebang_matches(self): self.assertTrue(util.shebang_matches('#!/usr/bin/env python', r'python(2\.\d)?')) self.assertTrue(util.shebang_matches('#!/usr/bin/python2.4', r'python(2\.\d)?')) self.assertTrue(util.shebang_matches('#!/usr/bin/startsomethingwith python', r'python(2\.\d)?')) self.assertTrue(util.shebang_matches('#!C:\\Python2.4\\Python.exe', r'python(2\.\d)?')) self.assertFalse(util.shebang_matches('#!/usr/bin/python-ruby', r'python(2\.\d)?')) self.assertFalse(util.shebang_matches('#!/usr/bin/python/ruby', r'python(2\.\d)?')) self.assertFalse(util.shebang_matches('#!', r'python')) def test_doctype_matches(self): self.assertTrue(util.doctype_matches( ' ', 'html.*')) self.assertFalse(util.doctype_matches( ' ', 'html.*')) self.assertTrue(util.html_doctype_matches( '')) def test_xml(self): self.assertTrue(util.looks_like_xml( '')) self.assertTrue(util.looks_like_xml('abc')) self.assertFalse(util.looks_like_xml('')) def test_unirange(self): first_non_bmp = u'\U00010000' r = re.compile(util.unirange(0x10000, 0x20000)) m = r.match(first_non_bmp) self.assertTrue(m) self.assertEquals(m.end(), len(first_non_bmp)) self.assertFalse(r.match(u'\uffff')) self.assertFalse(r.match(u'xxx')) # Tests that end is inclusive r = re.compile(util.unirange(0x10000, 0x10000) + '+') # Tests that the plus works for the entire unicode point, if narrow # build m = r.match(first_non_bmp * 2) self.assertTrue(m) self.assertEquals(m.end(), len(first_non_bmp) * 2)