8000 Added support for the PullDOM treewalker in tests (DOCTYPE nodes are … · awesome-python/html5lib-python@17e8245 · GitHub
[go: up one dir, main page]
Skip to content

Commit 17e8245

Browse files
tbroyertbroyer
committed
Added support for the PullDOM treewalker in tests (DOCTYPE nodes are not supported by PULLDOM so they raise NotImplementedError exceptions; as of now, there are 54 parser tests using DOCTYPE so there should be 54 errors in treewalker tests)
--HG-- extra : convert_revision : svn%3Aacbfec75-9323-0410-a652-858a13e371e0/trunk%40621
1 parent f3a281d commit 17e8245

File tree

1 file changed

+51
-18
lines changed

1 file changed

+51
-18
lines changed

‎tests/test_treewalkers.py

Lines changed: 51 additions & 18 deletions
Original file line numberDiff line numberDiff line change
@@ -38,56 +38,87 @@ def load(f):
3838
#from html5lib import html5parser, serializer, treewalkers, treebuilders
3939
#END RELEASE
4040

41+
def PullDOMAdapter(node):
42+
from xml.dom import Node
43+
from xml.dom.pulldom import START_ELEMENT, END_ELEMENT, COMMENT, CHARACTERS
44+
45+
if node.nodeType in (Node.DOCUMENT_NODE, Node.DOCUMENT_FRAGMENT_NODE):
46+
for childNode in node.childNodes:
47+
for event in PullDOMAdapter(childNode):
48+
yield event
49+
50+
elif node.nodeType == Node.DOCUMENT_TYPE_NODE:
51+
raise NotImplementedError("DOCTYPE nodes are not supported by PullDOM")
52+
53+
elif node.nodeType == Node.COMMENT_NODE:
54+
yield COMMENT, node
55+
56+
elif node.nodeType in (Node.TEXT_NODE, Node.CDATA_SECTION_NODE):
57+
yield CHARACTERS, node
58+
59+
elif node.nodeType == Node.ELEMENT_NODE:
60+
yield START_ELEMENT, node
61+
for childNode in node.childNodes:
62+
for event in PullDOMAdapter(childNode):
63+
yield event
64+
yield END_ELEMENT, node
65+
66+
else:
67+
raise NotImplementedError("Node type not supported: " + str(node.nodeType))
68+
4169
treeTypes = {
42-
"simpletree": (treebuilders.getTreeBuilder("simpletree"),
43-
treewalkers.getTreeWalker("simpletree")),
44-
"DOM": (treebuilders.getTreeBuilder("dom"),
45-
treewalkers.getTreeWalker("dom")),
70+
"simpletree": {"builder": treebuilders.getTreeBuilder("simpletree"),
71+
"walker": treewalkers.getTreeWalker("simpletree")},
72+
"DOM": {"builder": treebuilders.getTreeBuilder("dom"),
73+
"walker": treewalkers.getTreeWalker("dom")},
74+
"PullDOM": {"builder": treebuilders.getTreeBuilder("dom"),
75+
"adapter": PullDOMAdapter,
76+
"walker": treewalkers.getTreeWalker("pulldom")},
4677
}
4778

4879
#Try whatever etree implementations are available from a list that are
4980
#"supposed" to work
5081
try:
5182
import xml.etree.ElementTree as ElementTree
5283
treeTypes['ElementTree'] = \
53-
(treebuilders.getTreeBuilder("etree", ElementTree),
54-
treewalkers.getTreeWalker("etree", ElementTree))
84+
{"builder": treebuilders.getTreeBuilder("etree", ElementTree),
85+
"walker": treewalkers.getTreeWalker("etree", ElementTree)}
5586
except ImportError:
5687
try:
5788
import elementtree.ElementTree as ElementTree
5889
treeTypes['ElementTree'] = \
59-
(treebuilders.getTreeBuilder("etree", ElementTree),
60-
treewalkers.getTreeWalker("etree", ElementTree))
90+
{"builder": treebuilders.getTreeBuilder("etree", ElementTree),
91+
"walker": treewalkers.getTreeWalker("etree", ElementTree)}
6192
except ImportError:
6293
pass
6394

6495
try:
6596
import xml.etree.cElementTree as cElementTree
6697
treeTypes['cElementTree'] = \
67-
(treebuilders.getTreeBuilder("etree", cElementTree),
68-
treewalkers.getTreeWalker("etree", cElementTree))
98+
{"builder": treebuilders.getTreeBuilder("etree", cElementTree),
99+
"walker": treewalkers.getTreeWalker("etree", cElementTree)}
69100
except ImportError:
70101
try:
71102
import cElementTree
72103
treeTypes['cElementTree'] = \
73-
(treebuilders.getTreeBuilder("etree", cElementTree),
74-
treewalkers.getTreeWalker("etree", cElementTree))
104+
{"builder": treebuilders.getTreeBuilder("etree", cElementTree),
105+
"walker": treewalkers.getTreeWalker("etree", cElementTree)}
75106
except ImportError:
76107
pass
77108

78109
try:
79110
import lxml.etree as lxml
80111
treeTypes['lxml'] = \
81-
(treebuilders.getTreeBuilder("etree", lxml),
82-
treewalkers.getTreeWalker("etree", lxml))
112+
{"builder": treebuilders.getTreeBuilder("etree", lxml),
113+
"walker": treewalkers.getTreeWalker("etree", lxml)}
83114
except ImportError:
84115
pass
85116

86117
try:
87118
import BeautifulSoup
88119
treeTypes["beautifulsoup"] = \
89-
(treebuilders.getTreeBuilder("beautifulsoup"),
90-
treewalkers.getTreeWalker("beautifulsoup"))
120+
{"builder": treebuilders.getTreeBuilder("beautifulsoup"),
121+
"walker": treewalkers.getTreeWalker("beautifulsoup")}
91122
except ImportError:
92123
pass
93124

@@ -146,12 +177,14 @@ def sortattrs(x):
146177

147178
class TestCase(unittest.TestCase):
148179
def runTest(self, innerHTML, input, expected, errors, treeClass):
149-
p = html5parser.HTMLParser(tree = treeClass[0])
180+
p = html5parser.HTMLParser(tree = treeClass["builder"])
181+
150182
if innerHTML:
151183
document = p.parseFragment(StringIO.StringIO(input), innerHTML)
152184
else:
153185
document = p.parse(StringIO.StringIO(input))
154-
output = convertTokens(treeClass[1]().walk(document))
186+
document = treeClass.get("adapter", lambda x: x)(document)
187+
output = convertTokens(treeClass["walker"]().walk(document))
155188
output = attrlist.sub(sortattrs, output)
156189
expected = attrlist.sub(sortattrs, expected)
157190
errorMsg = "\n".join(["\n\nExpected:", expected,

0 commit comments

Comments
 (0)
0