[Python-checkins] CVS: python/dist/src/Lib/test test_sgmllib.py,NONE,1.1.2.1 test_generators.py,1.17.2.4,1.17.2.5 test_htmlparser.py,1.1.4.1,1.1.4.2 test_support.py,1.21.4.1,1.21.4.2

Tim Peters tim_one@users.sourceforge.net
Mon, 16 Jul 2001 14:39:43 -0700


Update of /cvsroot/python/python/dist/src/Lib/test
In directory usw-pr-cvs1:/tmp/cvs-serv11584/descr/dist/src/Lib/test

Modified Files:
      Tag: descr-branch
	test_generators.py test_htmlparser.py test_support.py 
Added Files:
      Tag: descr-branch
	test_sgmllib.py 
Log Message:
Resuming interrupted merge checkin.


--- NEW FILE: test_sgmllib.py ---
import pprint
import sgmllib
import test_support
import unittest


class EventCollector(sgmllib.SGMLParser):

    def __init__(self):
        self.events = []
        self.append = self.events.append
        sgmllib.SGMLParser.__init__(self)

    def get_events(self):
        # Normalize the list of events so that buffer artefacts don't
        # separate runs of contiguous characters.
        L = []
        prevtype = None
        for event in self.events:
            type = event[0]
            if type == prevtype == "data":
                L[-1] = ("data", L[-1][1] + event[1])
            else:
                L.append(event)
            prevtype = type
        self.events = L
        return L

    # structure markup

    def unknown_starttag(self, tag, attrs):
        self.append(("starttag", tag, attrs))

    def unknown_endtag(self, tag):
        self.append(("endtag", tag))

    # all other markup

    def handle_comment(self, data):
        self.append(("comment", data))

    def handle_charref(self, data):
        self.append(("charref", data))

    def handle_data(self, data):
        self.append(("data", data))

    def handle_decl(self, decl):
        self.append(("decl", decl))

    def handle_entityref(self, data):
        self.append(("entityref", data))

    def handle_pi(self, data):
        self.append(("pi", data))


class CDATAEventCollector(EventCollector):
    def start_cdata(self, attrs):
        self.append(("starttag", "cdata", attrs))
        self.setliteral()


class SGMLParserTestCase(unittest.TestCase):

    collector = EventCollector

    def check_events(self, source, expected_events):
        parser = self.collector()
        for s in source:
            parser.feed(s)
        parser.close()
        events = parser.get_events()
        if events != expected_events:
            self.fail("received events did not match expected events\n"
                      "Expected:\n" + pprint.pformat(expected_events) +
                      "\nReceived:\n" + pprint.pformat(events))

    def check_parse_error(self, source):
        parser = EventCollector()
        try:
            parser.feed(source)
            parser.close()
        except sgmllib.SGMLParseError:
            pass
        else:
            self.fail("expected SGMLParseError for %r\nReceived:\n%s"
                      % (source, pprint.pformat(parser.get_events())))

    def test_underscore_in_attrname(self):
        # SF bug #436621
        """Make sure attribute names with underscores are accepted"""
        self.check_events("<a has_under _under>", [
            ("starttag", "a", [("has_under", "has_under"),
                               ("_under", "_under")]),
            ])

    def test_underscore_in_tagname(self):
        # SF bug #436621
        """Make sure tag names with underscores are accepted"""
        self.check_events("<has_under></has_under>", [
            ("starttag", "has_under", []),
            ("endtag", "has_under"),
            ])

    def test_quotes_in_unquoted_attrs(self):
        # SF bug #436621
        """Be sure quotes in unquoted attributes are made part of the value"""
        self.check_events("<a href=foo'bar\"baz>", [
            ("starttag", "a", [("href", "foo'bar\"baz")]),
            ])

    def test_xhtml_empty_tag(self):
        """Handling of XHTML-style empty start tags"""
        self.check_events("<br />text<i></i>", [
            ("starttag", "br", []),
            ("data", "text"),
            ("starttag", "i", []),
            ("endtag", "i"),
            ])

    def test_processing_instruction_only(self):
        self.check_events("<?processing instruction>", [
            ("pi", "processing instruction"),
            ])

    def test_bad_nesting(self):
        self.check_events("<a><b></a></b>", [
            ("starttag", "a", []),
            ("starttag", "b", []),
            ("endtag", "a"),
            ("endtag", "b"),
            ])

    def test_attr_syntax(self):
        output = [
          ("starttag", "a", [("b", "v"), ("c", "v"), ("d", "v"), ("e", "e")])
          ]
        self.check_events("""<a b='v' c="v" d=v e>""", output)
        self.check_events("""<a  b = 'v' c = "v" d = v e>""", output)
        self.check_events("""<a\nb\n=\n'v'\nc\n=\n"v"\nd\n=\nv\ne>""", output)
        self.check_events("""<a\tb\t=\t'v'\tc\t=\t"v"\td\t=\tv\te>""", output)

    def test_attr_values(self):
        self.check_events("""<a b='xxx\n\txxx' c="yyy\t\nyyy" d='\txyz\n'>""",
                        [("starttag", "a", [("b", "xxx\n\txxx"),
                                            ("c", "yyy\t\nyyy"),
                                            ("d", "\txyz\n")])
                         ])
        self.check_events("""<a b='' c="">""", [
            ("starttag", "a", [("b", ""), ("c", "")]),
            ])

    def test_attr_funky_names(self):
        self.check_events("""<a a.b='v' c:d=v e-f=v>""", [
            ("starttag", "a", [("a.b", "v"), ("c:d", "v"), ("e-f", "v")]),
            ])

    def test_weird_starttags(self):
        self.check_events("<a<a>", [
            ("starttag", "a", []),
            ("starttag", "a", []),
            ])
        self.check_events("</a<a>", [
            ("endtag", "a"),
            ("starttag", "a", []),
            ])

    def test_declaration_junk_chars(self):
        self.check_parse_error("<!DOCTYPE foo $ >")

    def test_get_starttag_text(self):
        s = """<foobar   \n   one="1"\ttwo=2   >"""
        self.check_events(s, [
            ("starttag", "foobar", [("one", "1"), ("two", "2")]),
            ])

    def test_cdata_content(self):
        s = ("<cdata> <!-- not a comment --> &not-an-entity-ref; </cdata>"
             "<notcdata> <!-- comment --> </notcdata>")
        self.collector = CDATAEventCollector
        self.check_events(s, [
            ("starttag", "cdata", []),
            ("data", " <!-- not a comment --> &not-an-entity-ref; "),
            ("endtag", "cdata"),
            ("starttag", "notcdata", []),
            ("data", " "),
            ("comment", " comment "),
            ("data", " "),
            ("endtag", "notcdata"),
            ])
        s = """<cdata> <not a='start tag'> </cdata>"""
        self.check_events(s, [
            ("starttag", "cdata", []),
            ("data", " <not a='start tag'> "),
            ("endtag", "cdata"),
            ])

    # XXX These tests have been disabled by prefixing their names with
    # an underscore.  The first two exercise outstanding bugs in the
    # sgmllib module, and the third exhibits questionable behavior
    # that needs to be carefully considered before changing it.

    def _test_starttag_end_boundary(self):
        self.check_events("""<a b='<'>""", [("starttag", "a", [("b", "<")])])
        self.check_events("""<a b='>'>""", [("starttag", "a", [("b", ">")])])

    def _test_buffer_artefacts(self):
        output = [("starttag", "a", [("b", "<")])]
        self.check_events(["<a b='<'>"], output)
        self.check_events(["<a ", "b='<'>"], output)
        self.check_events(["<a b", "='<'>"], output)
        self.check_events(["<a b=", "'<'>"], output)
        self.check_events(["<a b='<", "'>"], output)
        self.check_events(["<a b='<'", ">"], output)

        output = [("starttag", "a", [("b", ">")])]
        self.check_events(["<a b='>'>"], output)
        self.check_events(["<a ", "b='>'>"], output)
        self.check_events(["<a b", "='>'>"], output)
        self.check_events(["<a b=", "'>'>"], output)
        self.check_events(["<a b='>", "'>"], output)
        self.check_events(["<a b='>'", ">"], output)

    def _test_starttag_junk_chars(self):
        self.check_parse_error("<")
        self.check_parse_error("<>")
        self.check_parse_error("</$>")
        self.check_parse_error("</")
        self.check_parse_error("</a")
        self.check_parse_error("<$")
        self.check_parse_error("<$>")
        self.check_parse_error("<!")
        self.check_parse_error("<a $>")
        self.check_parse_error("<a")
        self.check_parse_error("<a foo='bar'")
        self.check_parse_error("<a foo='bar")
        self.check_parse_error("<a foo='>'")
        self.check_parse_error("<a foo='>")
        self.check_parse_error("<a foo=>")


test_support.run_unittest(SGMLParserTestCase)

Index: test_generators.py
===================================================================
RCS file: /cvsroot/python/python/dist/src/Lib/test/test_generators.py,v
retrieving revision 1.17.2.4
retrieving revision 1.17.2.5
diff -C2 -r1.17.2.4 -r1.17.2.5
*** test_generators.py	2001/07/14 07:47:34	1.17.2.4
--- test_generators.py	2001/07/16 21:39:41	1.17.2.5
***************
*** 1,2 ****
--- 1,4 ----
+ from __future__ import generators
+ 
  tutorial_tests = """
  Let's try a simple generator:

Index: test_htmlparser.py
===================================================================
RCS file: /cvsroot/python/python/dist/src/Lib/test/test_htmlparser.py,v
retrieving revision 1.1.4.1
retrieving revision 1.1.4.2
diff -C2 -r1.1.4.1 -r1.1.4.2
*** test_htmlparser.py	2001/07/07 22:55:28	1.1.4.1
--- test_htmlparser.py	2001/07/16 21:39:41	1.1.4.2
***************
*** 200,204 ****
          self._parse_error("</")
          self._parse_error("</a")
-         self._parse_error("</a")
          self._parse_error("<a<a>")
          self._parse_error("</a<a>")
--- 200,203 ----

Index: test_support.py
===================================================================
RCS file: /cvsroot/python/python/dist/src/Lib/test/test_support.py,v
retrieving revision 1.21.4.1
retrieving revision 1.21.4.2
diff -C2 -r1.21.4.1 -r1.21.4.2
*** test_support.py	2001/07/06 21:20:46	1.21.4.1
--- test_support.py	2001/07/16 21:39:41	1.21.4.2
***************
*** 136,139 ****
      result = runner.run(suite)
      if not result.wasSuccessful():
!         raise TestFailed("errors occurred in %s.%s"
!                          % (testclass.__module__, testclass.__name__))
--- 136,148 ----
      result = runner.run(suite)
      if not result.wasSuccessful():
!         if len(result.errors) == 1 and not result.failures:
!             err = result.errors[0][1]
!         elif len(result.failures) == 1 and not result.errors:
!             err = result.failures[0][1]
!         else:
!             raise TestFailed("errors occurred in %s.%s"
!                              % (testclass.__module__, testclass.__name__))
!         if err[0] is AssertionError:
!             raise TestFailed(str(err[1]))
!         else:
!             raise TestFailed("%s: %s" % err[:2])