diff options
| author | scoder <stefan_ml@behnel.de> | 2017-03-18 09:56:26 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2017-03-18 09:56:26 +0100 |
| commit | 91e1d273549e29510a8f431ba561381514600575 (patch) | |
| tree | e5773b4bd09ddffcbdbff42756fb19eb7b14163d /src/lxml/html | |
| parent | 2802687228fd26c0d87d3aba76fb16570d448a25 (diff) | |
| parent | 62b8c345779bdfb8b261a6f7de3287f4fe800bb6 (diff) | |
| download | python-lxml-91e1d273549e29510a8f431ba561381514600575.tar.gz | |
Merge pull request #233 from ondergetekende/1673355
Fix LP1673355
Diffstat (limited to 'src/lxml/html')
| -rw-r--r-- | src/lxml/html/html5parser.py | 9 | ||||
| -rw-r--r-- | src/lxml/html/tests/test_html5parser.py | 6 |
2 files changed, 14 insertions, 1 deletions
diff --git a/src/lxml/html/html5parser.py b/src/lxml/html/html5parser.py index 7188c7ea..ba9d41b3 100644 --- a/src/lxml/html/html5parser.py +++ b/src/lxml/html/html5parser.py @@ -147,7 +147,14 @@ def fromstring(html, guess_charset=True, parser=None): guess_charset=guess_charset) # document starts with doctype or <html>, full document! - start = html[:50].lstrip().lower() + start = html[:50] + if hasattr(start, 'decode'): + # In python3, we may have been presented with a bytes object. + # Decode in ascii, that also covers latin-1 and utf-8 for the + # characters we need + start = start.decode('ascii', 'replace') + + start = start.lstrip().lower() if start.startswith('<html') or start.startswith('<!doctype'): return doc diff --git a/src/lxml/html/tests/test_html5parser.py b/src/lxml/html/tests/test_html5parser.py index fad45dc4..667c68d2 100644 --- a/src/lxml/html/tests/test_html5parser.py +++ b/src/lxml/html/tests/test_html5parser.py @@ -233,6 +233,12 @@ class Test_fromstring(unittest.TestCase): self.assertEqual(self.call_it('<!DOCTYPE html>', parser=parser), 'the doc') + def test_returns_whole_doc_if_input_is_encoded(self): + parser = DummyParser(root='the doc') + input = '<!DOCTYPE html>'.encode('ascii') + self.assertEqual(self.call_it(input, parser=parser), + 'the doc') + def test_returns_whole_doc_if_head_not_empty(self, use_ns=True): E = HTMLElementMaker(namespaceHTMLElements=use_ns) root = E.html(E.head(E.title())) |