[Python-checkins] cpython (merge 3.4 -> default): Issue #10590: Added tests for xml.sax.parse() and xml.sax.parseString().
serhiy.storchaka
python-checkins at python.org
Thu Apr 2 22:22:08 CEST 2015
https://hg.python.org/cpython/rev/846c165cf643
changeset: 95404:846c165cf643
parent: 95401:407883c52bf3
parent: 95403:ca8666310eb3
user: Serhiy Storchaka <storchaka at gmail.com>
date: Thu Apr 02 23:07:19 2015 +0300
summary:
Issue #10590: Added tests for xml.sax.parse() and xml.sax.parseString().
files:
Lib/test/test_sax.py | 124 ++++++++++++++++++++++++++++++-
1 files changed, 123 insertions(+), 1 deletions(-)
diff --git a/Lib/test/test_sax.py b/Lib/test/test_sax.py
--- a/Lib/test/test_sax.py
+++ b/Lib/test/test_sax.py
@@ -16,10 +16,11 @@
from xml.sax.xmlreader import InputSource, AttributesImpl, AttributesNSImpl
from io import BytesIO, StringIO
import codecs
+import gc
import os.path
import shutil
from test import support
-from test.support import findfile, run_unittest
+from test.support import findfile, run_unittest, TESTFN
TEST_XMLFILE = findfile("test.xml", subdir="xmltestdata")
TEST_XMLFILE_OUT = findfile("test.xml.out", subdir="xmltestdata")
@@ -95,6 +96,126 @@
self.assertEqual(attrs["attr"], "val")
self.assertEqual(attrs.getQNameByName("attr"), "attr")
+
+def xml_str(doc, encoding=None):
+ if encoding is None:
+ return doc
+ return '<?xml version="1.0" encoding="%s"?>\n%s' % (encoding, doc)
+
+def xml_bytes(doc, encoding, decl_encoding=...):
+ if decl_encoding is ...:
+ decl_encoding = encoding
+ return xml_str(doc, decl_encoding).encode(encoding, 'xmlcharrefreplace')
+
+def make_xml_file(doc, encoding, decl_encoding=...):
+ if decl_encoding is ...:
+ decl_encoding = encoding
+ with open(TESTFN, 'w', encoding=encoding, errors='xmlcharrefreplace') as f:
+ f.write(xml_str(doc, decl_encoding))
+
+
+class ParseTest(unittest.TestCase):
+ data = '<money value="$\xa3\u20ac\U0001017b">$\xa3\u20ac\U0001017b</money>'
+
+ def tearDown(self):
+ support.unlink(TESTFN)
+
+ def check_parse(self, f):
+ from xml.sax import parse
+ result = StringIO()
+ parse(f, XMLGenerator(result, 'utf-8'))
+ self.assertEqual(result.getvalue(), xml_str(self.data, 'utf-8'))
+
+ def test_parse_text(self):
+ encodings = ('us-ascii', 'iso-8859-1', 'utf-8',
+ 'utf-16', 'utf-16le', 'utf-16be')
+ for encoding in encodings:
+ self.check_parse(StringIO(xml_str(self.data, encoding)))
+ make_xml_file(self.data, encoding)
+ with open(TESTFN, 'r', encoding=encoding) as f:
+ self.check_parse(f)
+ self.check_parse(StringIO(self.data))
+ make_xml_file(self.data, encoding, None)
+ with open(TESTFN, 'r', encoding=encoding) as f:
+ self.check_parse(f)
+
+ def test_parse_bytes(self):
+ # UTF-8 is default encoding, US-ASCII is compatible with UTF-8,
+ # UTF-16 is autodetected
+ encodings = ('us-ascii', 'utf-8', 'utf-16', 'utf-16le', 'utf-16be')
+ for encoding in encodings:
+ self.check_parse(BytesIO(xml_bytes(self.data, encoding)))
+ make_xml_file(self.data, encoding)
+ self.check_parse(TESTFN)
+ with open(TESTFN, 'rb') as f:
+ self.check_parse(f)
+ self.check_parse(BytesIO(xml_bytes(self.data, encoding, None)))
+ make_xml_file(self.data, encoding, None)
+ self.check_parse(TESTFN)
+ with open(TESTFN, 'rb') as f:
+ self.check_parse(f)
+ # accept UTF-8 with BOM
+ self.check_parse(BytesIO(xml_bytes(self.data, 'utf-8-sig', 'utf-8')))
+ make_xml_file(self.data, 'utf-8-sig', 'utf-8')
+ self.check_parse(TESTFN)
+ with open(TESTFN, 'rb') as f:
+ self.check_parse(f)
+ self.check_parse(BytesIO(xml_bytes(self.data, 'utf-8-sig', None)))
+ make_xml_file(self.data, 'utf-8-sig', None)
+ self.check_parse(TESTFN)
+ with open(TESTFN, 'rb') as f:
+ self.check_parse(f)
+ # accept data with declared encoding
+ self.check_parse(BytesIO(xml_bytes(self.data, 'iso-8859-1')))
+ make_xml_file(self.data, 'iso-8859-1')
+ self.check_parse(TESTFN)
+ with open(TESTFN, 'rb') as f:
+ self.check_parse(f)
+ # fail on non-UTF-8 incompatible data without declared encoding
+ with self.assertRaises(SAXException):
+ self.check_parse(BytesIO(xml_bytes(self.data, 'iso-8859-1', None)))
+ make_xml_file(self.data, 'iso-8859-1', None)
+ with support.check_warnings(('unclosed file', ResourceWarning)):
+ # XXX Failed parser leaks an opened file.
+ with self.assertRaises(SAXException):
+ self.check_parse(TESTFN)
+ # Collect leaked file.
+ gc.collect()
+ with open(TESTFN, 'rb') as f:
+ with self.assertRaises(SAXException):
+ self.check_parse(f)
+
+ def test_parse_InputSource(self):
+ # accept data without declared but with explicitly specified encoding
+ make_xml_file(self.data, 'iso-8859-1', None)
+ with open(TESTFN, 'rb') as f:
+ input = InputSource()
+ input.setByteStream(f)
+ input.setEncoding('iso-8859-1')
+ self.check_parse(input)
+
+ def check_parseString(self, s):
+ from xml.sax import parseString
+ result = StringIO()
+ parseString(s, XMLGenerator(result, 'utf-8'))
+ self.assertEqual(result.getvalue(), xml_str(self.data, 'utf-8'))
+
+ def test_parseString_bytes(self):
+ # UTF-8 is default encoding, US-ASCII is compatible with UTF-8,
+ # UTF-16 is autodetected
+ encodings = ('us-ascii', 'utf-8', 'utf-16', 'utf-16le', 'utf-16be')
+ for encoding in encodings:
+ self.check_parseString(xml_bytes(self.data, encoding))
+ self.check_parseString(xml_bytes(self.data, encoding, None))
+ # accept UTF-8 with BOM
+ self.check_parseString(xml_bytes(self.data, 'utf-8-sig', 'utf-8'))
+ self.check_parseString(xml_bytes(self.data, 'utf-8-sig', None))
+ # accept data with declared encoding
+ self.check_parseString(xml_bytes(self.data, 'iso-8859-1'))
+ # fail on non-UTF-8 incompatible data without declared encoding
+ with self.assertRaises(SAXException):
+ self.check_parseString(xml_bytes(self.data, 'iso-8859-1', None))
+
class MakeParserTest(unittest.TestCase):
def test_make_parser2(self):
# Creating parsers several times in a row should succeed.
@@ -1115,6 +1236,7 @@
def test_main():
run_unittest(MakeParserTest,
+ ParseTest,
SaxutilsTest,
PrepareInputSourceTest,
StringXmlgenTest,
--
Repository URL: https://hg.python.org/cpython
More information about the Python-checkins
mailing list