diff options
| author | Leonard Richardson <leonard.richardson@canonical.com> | 2011-02-27 22:12:34 -0500 |
|---|---|---|
| committer | Leonard Richardson <leonard.richardson@canonical.com> | 2011-02-27 22:12:34 -0500 |
| commit | 710ee3323074c1432ece18b1eb9a40cacc0c601d (patch) | |
| tree | 0e1aecad650cdfd499878d7399758e86705783fe /tests/test_html5lib.py | |
| parent | 561b294b2f49bd6c752ae090056e2694dae79d49 (diff) | |
| parent | 082a8c84a79fa33ea23c159495005ebe9a39cbf4 (diff) | |
Added a bunch of tests to verify that BS4 fixes various bugs.
Diffstat (limited to 'tests/test_html5lib.py')
| -rw-r--r-- | tests/test_html5lib.py | 9 |
1 files changed, 9 insertions, 0 deletions
diff --git a/tests/test_html5lib.py b/tests/test_html5lib.py index b9bdd1d..ac99832 100644 --- a/tests/test_html5lib.py +++ b/tests/test_html5lib.py @@ -200,6 +200,15 @@ class TestHTML5BuilderInvalidMarkup(TestLXMLBuilderInvalidMarkup): soup = self.soup("<p>foodbaz</p>") self.assertEquals(soup.p.string, "foodbaz") + def test_entity_out_of_range(self): + # An entity that's out of range will be converted to + # REPLACEMENT CHARACTER. + soup = self.soup("<p>�</p>") + self.assertEquals(soup.p.string, u"\N{REPLACEMENT CHARACTER}") + + soup = self.soup("<p>�</p>") + self.assertEquals(soup.p.string, u"\N{REPLACEMENT CHARACTER}") + class TestHTML5LibEncodingConversion(TestLXMLBuilderEncodingConversion): @property |