#
# Test suite for the textwrap module.
#
# Original tests written by Greg Ward <gward@python.net>.
# Converted to PyUnit by Peter Hansen <peter@engcorp.com>.
# Currently maintained by Greg Ward.
#
# $Id$
#

import unittest

from textwrap import TextWrapper, wrap, fill, dedent, indent, shorten


class BaseTestCase(unittest.TestCase):
    '''Parent class with utility methods for textwrap tests.'''

    def show(self, textin):
        if isinstance(textin, list):
            result = []
            for i in range(len(textin)):
                result.append("  %d: %r" % (i, textin[i]))
            result = "\n".join(result) if result else "  no lines"
        elif isinstance(textin, str):
            result = "  %s\n" % repr(textin)
        return result


    def check(self, result, expect):
        self.assertEqual(result, expect,
            'expected:\n%s\nbut got:\n%s' % (
                self.show(expect), self.show(result)))

    def check_wrap(self, text, width, expect, **kwargs):
        result = wrap(text, width, **kwargs)
        self.check(result, expect)

    def check_split(self, text, expect):
        result = self.wrapper._split(text)
        self.assertEqual(result, expect,
                         "\nexpected %r\n"
                         "but got  %r" % (expect, result))


class WrapTestCase(BaseTestCase):

    def setUp(self):
        self.wrapper = TextWrapper(width=45)

    def test_simple(self):
        # Simple case: just words, spaces, and a bit of punctuation

        text = "Hello there, how are you this fine day?  I'm glad to hear it!"

        self.check_wrap(text, 12,
                        ["Hello there,",
                         "how are you",
                         "this fine",
                         "day?  I'm",
                         "glad to hear",
                         "it!"])
        self.check_wrap(text, 42,
                        ["Hello there, how are you this fine day?",
                         "I'm glad to hear it!"])
        self.check_wrap(text, 80, [text])

    def test_empty_string(self):
        # Check that wrapping the empty string returns an empty list.
        self.check_wrap("", 6, [])
        self.check_wrap("", 6, [], drop_whitespace=False)

    def test_empty_string_with_initial_indent(self):
        # Check that the empty string is not indented.
        self.check_wrap("", 6, [], initial_indent="++")
        self.check_wrap("", 6, [], initial_indent="++", drop_whitespace=False)

    def test_whitespace(self):
        # Whitespace munging and end-of-sentence detection

        text = """\
This is a paragraph that already has
line breaks.  But some of its lines are much longer than the others,
so it needs to be wrapped.
Some lines are \ttabbed too.
What a mess!
"""

        expect = ["This is a paragraph that already has line",
                  "breaks.  But some of its lines are much",
                  "longer than the others, so it needs to be",
                  "wrapped.  Some lines are  tabbed too.  What a",
                  "mess!"]

        wrapper = TextWrapper(45, fix_sentence_endings=True)
        result = wrapper.wrap(text)
        self.check(result, expect)

        result = wrapper.fill(text)
        self.check(result, '\n'.join(expect))

        text = "\tTest\tdefault\t\ttabsize."
        expect = ["        Test    default         tabsize."]
        self.check_wrap(text, 80, expect)

        text = "\tTest\tcustom\t\ttabsize."
        expect = ["    Test    custom      tabsize."]
        self.check_wrap(text, 80, expect, tabsize=4)

    def test_fix_sentence_endings(self):
        wrapper = TextWrapper(60, fix_sentence_endings=True)

        # SF #847346: ensure that fix_sentence_endings=True does the
        # right thing even on input short enough that it doesn't need to
        # be wrapped.
        text = "A short line. Note the single space."
        expect = ["A short line.  Note the single space."]
        self.check(wrapper.wrap(text), expect)

        # Test some of the hairy end cases that _fix_sentence_endings()
        # is supposed to handle (the easy stuff is tested in
        # test_whitespace() above).
        text = "Well, Doctor? What do you think?"
        expect = ["Well, Doctor?  What do you think?"]
        self.check(wrapper.wrap(text), expect)

        text = "Well, Doctor?\nWhat do you think?"
        self.check(wrapper.wrap(text), expect)

        text = 'I say, chaps! Anyone for "tennis?"\nHmmph!'
        expect = ['I say, chaps!  Anyone for "tennis?"  Hmmph!']
        self.check(wrapper.wrap(text), expect)

        wrapper.width = 20
        expect = ['I say, chaps!', 'Anyone for "tennis?"', 'Hmmph!']
        self.check(wrapper.wrap(text), expect)

        text = 'And she said, "Go to hell!"\nCan you believe that?'
        expect = ['And she said, "Go to',
                  'hell!"  Can you',
                  'believe that?']
        self.check(wrapper.wrap(text), expect)

        wrapper.width = 60
        expect = ['And she said, "Go to hell!"  Can you believe that?']
        self.check(wrapper.wrap(text), expect)

        text = 'File stdio.h is nice.'
        expect = ['File stdio.h is nice.']
        self.check(wrapper.wrap(text), expect)

    def test_wrap_short(self):
        # Wrapping to make short lines longer

        text = "This is a\nshort paragraph."

        self.check_wrap(text, 20, ["This is a short",
                                   "paragraph."])
        self.check_wrap(text, 40, ["This is a short paragraph."])


    def test_wrap_short_1line(self):
        # Test endcases

        text = "This is a short line."

        self.check_wrap(text, 30, ["This is a short line."])
        self.check_wrap(text, 30, ["(1) This is a short line."],
                        initial_indent="(1) ")


    def test_hyphenated(self):
        # Test breaking hyphenated words

        text = ("this-is-a-useful-feature-for-"
                "reformatting-posts-from-tim-peters'ly")

        self.check_wrap(text, 40,
                        ["this-is-a-useful-feature-for-",
                         "reformatting-posts-from-tim-peters'ly"])
        self.check_wrap(text, 41,
                        ["this-is-a-useful-feature-for-",
                         "reformatting-posts-from-tim-peters'ly"])
        self.check_wrap(text, 42,
                        ["this-is-a-useful-feature-for-reformatting-",
                         "posts-from-tim-peters'ly"])
        # The test tests current behavior but is not testing parts of the API.
        expect = ("this-|is-|a-|useful-|feature-|for-|"
                  "reformatting-|posts-|from-|tim-|peters'ly").split('|')
        self.check_wrap(text, 1, expect, break_long_words=False)
        self.check_split(text, expect)

        self.check_split('e-mail', ['e-mail'])
        self.check_split('Jelly-O', ['Jelly-O'])
        # The test tests current behavior but is not testing parts of the API.
        self.check_split('half-a-crown', 'half-|a-|crown'.split('|'))

    def test_hyphenated_numbers(self):
        # Test that hyphenated numbers (eg. dates) are not broken like words.
        text = ("Python 1.0.0 was released on 1994-01-26.  Python 1.0.1 was\n"
                "released on 1994-02-15.")

        self.check_wrap(text, 30, ['Python 1.0.0 was released on',
                                   '1994-01-26.  Python 1.0.1 was',
                                   'released on 1994-02-15.'])
        self.check_wrap(text, 40, ['Python 1.0.0 was released on 1994-01-26.',
                                   'Python 1.0.1 was released on 1994-02-15.'])
        self.check_wrap(text, 1, text.split(), break_long_words=False)

        text = "I do all my shopping at 7-11."
        self.check_wrap(text, 25, ["I do all my shopping at",
                                   "7-11."])
        self.check_wrap(text, 27, ["I do all my shopping at",
                                   "7-11."])
        self.check_wrap(text, 29, ["I do all my shopping at 7-11."])
        self.check_wrap(text, 1, text.split(), break_long_words=False)

    def test_em_dash(self):
        # Test text with em-dashes
        text = "Em-dashes should be written -- thus."
        self.check_wrap(text, 25,
                        ["Em-dashes should be",
                         "written -- thus."])

        # Probe the boundaries of the properly written em-dash,
        # ie. " -- ".
        self.check_wrap(text, 29,
                        ["Em-dashes should be written",
                         "-- thus."])
        expect = ["Em-dashes should be written --",
                  "thus."]
        self.check_wrap(text, 30, expect)
        self.check_wrap(text, 35, expect)
        self.check_wrap(text, 36,
                        ["Em-dashes should be written -- thus."])

        # The improperly written em-dash is handled too, because
        # it's adjacent to non-whitespace on both sides.
        text = "You can also do--this or even---this."
        expect = ["You can also do",
                  "--this or even",
                  "---this."]
        self.check_wrap(text, 15, expect)
        self.check_wrap(text, 16, expect)
        expect = ["You can also do--",
                  "this or even---",
                  "this."]
        self.check_wrap(text, 17, expect)
        self.check_wrap(text, 19, expect)
        expect = ["You can also do--this or even",
                  "---this."]
        self.check_wrap(text, 29, expect)
        self.check_wrap(text, 31, expect)
        expect = ["You can also do--this or even---",
                  "this."]
        self.check_wrap(text, 32, expect)
        self.check_wrap(text, 35, expect)

        # All of the above behaviour could be deduced by probing the
        # _split() method.
        text = "Here's an -- em-dash and--here's another---and another!"
        expect = ["Here's", " ", "an", " ", "--", " ", "em-", "dash", " ",
                  "and", "--", "here's", " ", "another", "---",
                  "and", " ", "another!"]
        self.check_split(text, expect)

        text = "and then--bam!--he was gone"
        expect = ["and", " ", "then", "--", "bam!", "--",
                  "he", " ", "was", " ", "gone"]
        self.check_split(text, expect)


    def test_unix_options (self):
        # Test that Unix-style command-line options are wrapped correctly.
        # Both Optik (OptionParser) and Docutils rely on this behaviour!

        text = "You should use the -n option, or --dry-run in its long form."
        self.check_wrap(text, 20,
                        ["You should use the",
                         "-n option, or --dry-",
                         "run in its long",
                         "form."])
        self.check_wrap(text, 21,
                        ["You should use the -n",
                         "option, or --dry-run",
                         "in its long form."])
        expect = ["You should use the -n option, or",
                  "--dry-run in its long form."]
        self.check_wrap(text, 32, expect)
        self.check_wrap(text, 34, expect)
        self.check_wrap(text, 35, expect)
        self.check_wrap(text, 38, expect)
        expect = ["You should use the -n option, or --dry-",
                  "run in its long form."]
        self.check_wrap(text, 39, expect)
        self.check_wrap(text, 41, expect)
        expect = ["You should use the -n option, or --dry-run",
                  "in its long form."]
        self.check_wrap(text, 42, expect)

        # Again, all of the above can be deduced from _split().
        text = "the -n option, or --dry-run or --dryrun"
        expect = ["the", " ", "-n", " ", "option,", " ", "or", " ",
                  "--dry-", "run", " ", "or", " ", "--dryrun"]
        self.check_split(text, expect)

    def test_funky_hyphens (self):
        # Screwy edge cases cooked up by David Goodger.  All reported
        # in SF bug #596434.
        self.check_split("what the--hey!", ["what", " ", "the", "--", "hey!"])
        self.check_split("what the--", ["what", " ", "the--"])
        self.check_split("what the--.", ["what", " ", "the--."])
        self.check_split("--text--.", ["--text--."])

        # When I first read bug #596434, this is what I thought David
        # was talking about.  I was wrong; these have always worked
        # fine.  The real problem is tested in test_funky_parens()
        # below...
        self.check_split("--option", ["--option"])
        self.check_split("--option-opt", ["--option-", "opt"])
        self.check_split("foo --option-opt bar",
                         ["foo", " ", "--option-", "opt", " ", "bar"])

    def test_punct_hyphens(self):
        # Oh bother, SF #965425 found another problem with hyphens --
        # hyphenated words in single quotes weren't handled correctly.
        # In fact, the bug is that *any* punctuation around a hyphenated
        # word was handled incorrectly, except for a leading "--", which
        # was special-cased for Optik and Docutils.  So test a variety
        # of styles of punctuation around a hyphenated word.
        # (Actually this is based on an Optik bug report, #813077).
        self.check_split("the 'wibble-wobble' widget",
                         ['the', ' ', "'wibble-", "wobble'", ' ', 'widget'])
        self.check_split('the "wibble-wobble" widget',
                         ['the', ' ', '"wibble-', 'wobble"', ' ', 'widget'])
        self.check_split("the (wibble-wobble) widget",
                         ['the', ' ', "(wibble-", "wobble)", ' ', 'widget'])
        self.check_split("the ['wibble-wobble'] widget",
                         ['the', ' ', "['wibble-", "wobble']", ' ', 'widget'])

        # The test tests current behavior but is not testing parts of the API.
        self.check_split("what-d'you-call-it.",
                         "what-d'you-|call-|it.".split('|'))

    def test_funky_parens (self):