Skip to content

Commit 33ebff4

Browse files
serhiy-storchakaserhiy-storchaka
committed
Run the testing code in subprocess.
1 parent 97c8a36 commit 33ebff4

4 files changed

Lines changed: 59 additions & 49 deletions

File tree

‎Doc/library/html.parser.rst‎

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -193,6 +193,14 @@ implementations do nothing (except for :meth:`~HTMLParser.handle_startendtag`):
193193
this method will receive ``'[if IE 9]>IE9-specific content<![endif]'``.
194194

195195

196+
.. method:: HTMLParser.handle_bogus_comment(data)
197+
198+
This method is called when a bogus comment is encountered (e.g. ``<!...>`` or ``</...>``).
199+
The default implementation simply calls :meth:`handle_comment`.
200+
201+
.. versionadded:: next
202+
203+
196204
.. method:: HTMLParser.handle_decl(decl)
197205

198206
This method is called to handle an HTML doctype declaration (e.g.

‎Lib/html/parser.py‎

Lines changed: 8 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -262,8 +262,7 @@ def goahead(self, end):
262262
elif endtagopen.match(rawdata, i): # </ + letter
263263
pass
264264
else:
265-
# bogus comment
266-
self.handle_comment(rawdata[i+2:])
265+
self.handle_bogus_comment(rawdata[i+2:])
267266
elif startswith("<!--", i):
268267
j = n
269268
for suffix in ("--!", "--", "-"):
@@ -275,12 +274,11 @@ def goahead(self, end):
275274
if self._support_cdata:
276275
self.unknown_decl(rawdata[i+3:])
277276
else:
278-
self.handle_comment(rawdata[i+1:])
277+
self.handle_bogus_comment(rawdata[i+1:])
279278
elif rawdata[i:i+9].lower() == '<!doctype':
280279
self.handle_decl(rawdata[i+2:])
281280
elif startswith("<!", i):
282-
# bogus comment
283-
self.handle_comment(rawdata[i+2:])
281+
self.handle_bogus_comment(rawdata[i+2:])
284282
elif startswith("<?", i):
285283
self.handle_pi(rawdata[i+2:])
286284
else:
@@ -394,7 +392,7 @@ def parse_bogus_comment(self, i, report=1):
394392
if pos == -1:
395393
return -1
396394
if report:
397-
self.handle_comment(rawdata[i+2:pos])
395+
self.handle_bogus_comment(rawdata[i+2:pos])
398396
return pos + 1
399397

400398
# Internal -- parse processing instr, return end or -1 if not terminated
@@ -522,6 +520,10 @@ def handle_entityref(self, name):
522520
def handle_data(self, data):
523521
pass
524522

523+
# Overridable -- handle bogus comment
524+
def handle_bogus_comment(self, data):
525+
self.handle_comment(data)
526+
525527
# Overridable -- handle comment
526528
def handle_comment(self, data):
527529
pass

‎Lib/test/test_htmlparser.py‎

Lines changed: 39 additions & 36 deletions
Original file line numberDiff line numberDiff line change
@@ -50,6 +50,9 @@ def handle_endtag(self, tag):
5050

5151
# all other markup
5252

53+
def handle_bogus_comment(self, data):
54+
self.append(("bogus comment", data))
55+
5356
def handle_comment(self, data):
5457
self.append(("comment", data))
5558

@@ -606,14 +609,14 @@ def test_starttag_junk_chars(self):
606609
self._run_check("< ", [('data', '< ')])
607610
self._run_check("</>", [])
608611
self._run_check("<$>", [('data', '<$>')])
609-
self._run_check("</$>", [('comment', '$')])
612+
self._run_check("</$>", [('bogus comment', '$')])
610613
self._run_check("</", [('data', '</')])
611614
self._run_check("</a", [])
612-
self._run_check("</ a>", [('comment', ' a')])
613-
self._run_check("</ a", [('comment', ' a')])
615+
self._run_check("</ a>", [('bogus comment', ' a')])
616+
self._run_check("</ a", [('bogus comment', ' a')])
614617
self._run_check("<a<a>", [('starttag', 'a<a', [])])
615618
self._run_check("</a<a>", [('endtag', 'a<a')])
616-
self._run_check("<!", [('comment', '')])
619+
self._run_check("<!", [('bogus comment', '')])
617620
self._run_check("<a", [])
618621
self._run_check("<a foo='bar'", [])
619622
self._run_check("<a foo='bar", [])
@@ -666,7 +669,7 @@ def test_declaration_junk_chars(self):
666669

667670
def test_illegal_declarations(self):
668671
self._run_check('<!spacer type="block" height="25">',
669-
[('comment', 'spacer type="block" height="25"')])
672+
[('bogus comment', 'spacer type="block" height="25"')])
670673

671674
def test_invalid_end_tags(self):
672675
# A collection of broken end tags. <br> is used as separator.
@@ -681,8 +684,8 @@ def test_invalid_end_tags(self):
681684
# text and attributes are discarded
682685
('endtag', 'div'),
683686
('starttag', 'br', []),
684-
# comment because the first char after </ is not a-zA-Z
685-
('comment', '<h4'),
687+
# bogus comment because the first char after </ is not a-zA-Z
688+
('bogus comment', '<h4'),
686689
('starttag', 'br', []),
687690
# attributes are discarded
688691
('endtag', 'li'),
@@ -771,9 +774,9 @@ def test_eof_in_comments(self):
771774

772775
def test_eof_in_declarations(self):
773776
data = [
774-
('<!', [('comment', '')]),
775-
('<!-', [('comment', '-')]),
776-
('<![', [('comment', '[')]),
777+
('<!', [('bogus comment', '')]),
778+
('<!-', [('bogus comment', '-')]),
779+
('<![', [('bogus comment', '[')]),
777780
('<!DOCTYPE', [('decl', 'DOCTYPE')]),
778781
('<!DOCTYPE ', [('decl', 'DOCTYPE ')]),
779782
('<!DOCTYPE html', [('decl', 'DOCTYPE html')]),
@@ -791,7 +794,7 @@ def test_eof_in_cdata(self, content):
791794
self._run_check('<![CDATA[' + content,
792795
[('unknown decl', 'CDATA[' + content)])
793796
self._run_check('<![CDATA[' + content,
794-
[('comment', '![CDATA[' + content)],
797+
[('bogus comment', '![CDATA[' + content)],
795798
collector=EventCollector(autocdata=True))
796799
self._run_check('<svg><text y="100"><![CDATA[' + content,
797800
[('starttag', 'svg', []),
@@ -814,19 +817,19 @@ def test_bogus_comments(self):
814817
'<![CDATA]]>' # required '[' after CDATA
815818
)
816819
expected = [
817-
('comment', 'ELEMENT br EMPTY'),
818-
('comment', ' not really a comment '),
819-
('comment', ' not a comment either --'),
820-
('comment', ' -- close enough --'),
821-
('comment', ''),
822-
('comment', '<-- this was an empty comment'),
823-
('comment', '!! another bogus comment !!!'),
824-
('comment', '[with square brackets]!'),
825-
('comment', '[\nmultiline\nbogusness\n]!'),
826-
('comment', '[more brackets]-[and a hyphen]!'),
827-
('comment', '[cdata[should be uppercase]]'),
828-
('comment', '[CDATA [whitespaces are not ignored]]'),
829-
('comment', '[CDATA]]'),
820+
('bogus comment', 'ELEMENT br EMPTY'),
821+
('bogus comment', ' not really a comment '),
822+
('bogus comment', ' not a comment either --'),
823+
('bogus comment', ' -- close enough --'),
824+
('bogus comment', ''),
825+
('bogus comment', '<-- this was an empty comment'),
826+
('bogus comment', '!! another bogus comment !!!'),
827+
('bogus comment', '[with square brackets]!'),
828+
('bogus comment', '[\nmultiline\nbogusness\n]!'),
829+
('bogus comment', '[more brackets]-[and a hyphen]!'),
830+
('bogus comment', '[cdata[should be uppercase]]'),
831+
('bogus comment', '[CDATA [whitespaces are not ignored]]'),
832+
('bogus comment', '[CDATA]]'),
830833
]
831834
self._run_check(html, expected)
832835

@@ -840,23 +843,23 @@ def test_broken_condcoms(self):
840843
'<![if !ie 6]><b>foo</b><![endif]>'
841844
'<![if (!IE)|(lt IE 9)]><img src="mammoth.bmp" /><![endif]>')
842845
expected = [
843-
('comment', '[if !(IE)]'),
846+
('bogus comment', '[if !(IE)]'),
844847
('data', 'broken condcom'),
845-
('comment', '[endif]'),
846-
('comment', '[if ! IE]'),
848+
('bogus comment', '[endif]'),
849+
('bogus comment', '[if ! IE]'),
847850
('startendtag', 'link', [('href', 'favicon.tiff')]),
848-
('comment', '[endif]'),
849-
('comment', '[if !IE 6]'),
851+
('bogus comment', '[endif]'),
852+
('bogus comment', '[if !IE 6]'),
850853
('startendtag', 'img', [('src', 'firefox.png')]),
851-
('comment', '[endif]'),
852-
('comment', '[if !ie 6]'),
854+
('bogus comment', '[endif]'),
855+
('bogus comment', '[if !ie 6]'),
853856
('starttag', 'b', []),
854857
('data', 'foo'),
855858
('endtag', 'b'),
856-
('comment', '[endif]'),
857-
('comment', '[if (!IE)|(lt IE 9)]'),
859+
('bogus comment', '[endif]'),
860+
('bogus comment', '[if (!IE)|(lt IE 9)]'),
858861
('startendtag', 'img', [('src', 'mammoth.bmp')]),
859-
('comment', '[endif]')
862+
('bogus comment', '[endif]')
860863
]
861864
self._run_check(html, expected)
862865

@@ -896,14 +899,14 @@ def test_cdata_section(self):
896899
'<svg><text y="100"><![CDATA[foo<br>bar]]></text></svg>'
897900
'<![CDATA[foo<br>bar]]>')
898901
expected = [
899-
('comment', '[CDATA[foo<br'),
902+
('bogus comment', '[CDATA[foo<br'),
900903
('data', 'bar]]>'),
901904
('starttag', 'svg', []),
902905
('starttag', 'text', [('y', '100')]),
903906
('unknown decl', 'CDATA[foo<br>bar'),
904907
('endtag', 'text'),
905908
('endtag', 'svg'),
906-
('comment', '[CDATA[foo<br'),
909+
('bogus comment', '[CDATA[foo<br'),
907910
('data', 'bar]]>'),
908911
]
909912
self._run_check(html, expected, collector=EventCollector(autocdata=True))

‎Lib/test/test_zoneinfo/test_zoneinfo.py‎

Lines changed: 4 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -12,7 +12,6 @@
1212
import re
1313
import shutil
1414
import struct
15-
import sys
1615
import tempfile
1716
import unittest
1817
from datetime import date, datetime, time, timedelta, timezone
@@ -23,6 +22,7 @@
2322
from test.test_zoneinfo import _support as test_support
2423
from test.test_zoneinfo._support import TZPATH_TEST_LOCK, ZoneInfoTestBase
2524
from test.support.import_helper import import_module, CleanImport
25+
from test.support.script_helper import assert_python_ok
2626

2727
lzma = import_module('lzma')
2828
py_zoneinfo, c_zoneinfo = test_support.get_modules()
@@ -1949,17 +1949,14 @@ class CTestModule(TestModule):
19491949

19501950
class MiscTests(unittest.TestCase):
19511951
def test_pydatetime(self):
1952-
with CleanImport('zoneinfo', 'zoneinfo._tzpath', 'zoneinfo._zoneinfo',
1953-
'_zoneinfo', 'datetime', '_pydatetime', '_datetime'):
1952+
assert_python_ok('-c', '''if 1:
1953+
import sys
19541954
sys.modules['_datetime'] = None
19551955
import datetime
19561956
import zoneinfo
1957-
zoneinfo.ZoneInfo.clear_cache()
19581957
tzinfo = zoneinfo.ZoneInfo('Europe/Paris')
19591958
datetime.datetime(2025, 10, 26, 2, 0, tzinfo=tzinfo)
1960-
self.assertIn('_pydatetime', sys.modules)
1961-
self.assertNotIn('_zoneinfo', sys.modules)
1962-
self.assertIn('zoneinfo._zoneinfo', sys.modules)
1959+
''')
19631960

19641961

19651962
class ExtensionBuiltTest(unittest.TestCase):

0 commit comments

Comments
 (0)