o2w/talon
5
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

fallback untouched html if we can not parse html tree

Browse Source
This commit is contained in:
Sergey Obukhov
2016-08-19 11:38:12 -07:00
parent 5b1ca33c57
commit 37c95ff97b
4 changed files with 45 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
tests/utils_test.py
View File

@@ -112,5 +112,22 @@ font: 13px 'Lucida Grande', Arial, sans-serif;
def test_comment_no_parent():
s = "<!-- COMMENT 1 --> no comment"
d = html.document_fromstring(s)
d = u.html_document_fromstring(s)
eq_("no comment", u.html_tree_to_text(d))
@patch.object(u.html5parser, 'fromstring', Mock(side_effect=Exception()))
def test_html_fromstring_exception():
eq_(None, u.html_fromstring("<html></html>"))
@patch.object(u.html5parser, 'document_fromstring')
def test_html_document_fromstring_exception(document_fromstring):
document_fromstring.side_effect = Exception()
eq_(None, u.html_document_fromstring("<html></html>"))
@patch.object(u, 'html_fromstring', Mock(return_value=None))
def test_bad_html_to_text():
bad_html = "one<br>two<br>three"
eq_(None, u.html_to_text(bad_html))