Skip to content

gh-143927: Normalize all line endings (CR, CRLF, and LF) in configparser #143929

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
encukou merged 1 commit into from
Jun 23, 2026
+16 −1
Merged

gh-143927: Normalize all line endings (CR, CRLF, and LF) in configparser #143929

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 3 additions & 1 deletion Lib/configparser.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -985,7 +985,9 @@ def _write_section(self, fp, section_name, section_items, delimiter, unnamed=Fal
value = self._interpolation.before_write(self, section_name, key,
value)
if value is not None or not self._allow_no_value:
value = delimiter + str(value).replace('\n', '\n\t')
# Convert all possible line-endings into '\n\t'

@cmaloney cmaloney Jan 17, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure the exact underlying security issue. As is, I think this will break round tripping on Windows; Text I/O used by the open expects to encode/decode newlines in particular formats on Windows by default (https://github.com/sethmlarson/cpython/blob/2a122fd420f5b425fd39848a48f5eb196cee2aa7/Lib/configparser.py#L753). See #143428 (comment) for a recent related case + more background links.

In the open calls can we use newline="" potentially?

@encukou encukou Jun 16, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Byte-for-byte round-tripping is already kind of broken for newlines. This is a text file; relying on newline format (or e.g. trailing whitespace) is a bad idea to begin with.

In the open calls can we use newline="" potentially?

Potentially, but it would change the behavior for all users on Windows.
And it wouldn't solve the issue, since users can pass in a file they opened themselves.

value = (delimiter + str(value).replace('\r\n', '\n')
.replace('\r', '\n').replace('\n', '\n\t'))
Comment on lines +989 to +990

@encukou encukou Jun 16, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

On my system, re is significantly faster than three replaces for reasonable-sized input.

Suggested change
value = (delimiter + str(value).replace('\r\n', '\n')
.replace('\r', '\n').replace('\n', '\n\t'))
value = delimiter + re.sub('\r\n?|\n', '\n\t', str(value))

@hugovk hugovk Jun 22, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Interesting, re is about four times slower than three replacess on my macOS:

❯ p --version --version
Python 3.15.0b2 (v3.15.0b2:94a64bbc6ce, Jun  2 2026, 11:57:29) [Clang 21.0.0 (clang-2100.1.1.101)]

❯ p -m timeit -s "v = 'lorem ipsum dolor sit amet '*5" "v.replace('\r\n', '\n').replace('\r', '\n').replace('\n', '\n\t')"
2000000 loops, best of 5: 189 nsec per loop

❯ p -m timeit -s "import re; v = 'lorem ipsum dolor sit amet '*5" "re.sub('\r\n?|\n', '\n\t', v)"
500000 loops, best of 5: 773 nsec per loop

❯ p -m timeit -s "v = ('lorem ipsum dolor sit amet '*5 + '\n')*5" "v.replace('\r\n', '\n').replace('\r', '\n').replace('\n', '\n\t')"
500000 loops, best of 5: 936 nsec per loop

❯ p -m timeit -s "import re; v = ('lorem ipsum dolor sit amet '*5 + '\n')*5" "re.sub('\r\n?|\n', '\n\t', v)"
100000 loops, best of 5: 3.43 usec per loop

❯ p -m timeit -s "v = 'a\r\nb\nc\rd ' * 200" "v.replace('\r\n', '\n').replace('\r', '\n').replace('\n', '\n\t')"
20000 loops, best of 5: 10.2 usec per loop

❯ p -m timeit -s "import re; v = 'a\r\nb\nc\rd ' * 200" "re.sub('\r\n?|\n', '\n\t', v)"
5000 loops, best of 5: 43.1 usec per loop

❯ p -m timeit -s "v = 'a\r\nb\nc\rd ' * 10000" "v.replace('\r\n', '\n').replace('\r', '\n').replace('\n', '\n\t')"
500 loops, best of 5: 516 usec per loop

❯ p -m timeit -s "import re; v = 'a\r\nb\nc\rd ' * 10000" "re.sub('\r\n?|\n', '\n\t', v)"
100 loops, best of 5: 2.07 msec per loop

@encukou encukou Jun 23, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

OK, let's keep the replaces.

else:
value = ""
fp.write("{}{}\n".format(key, value))
Expand Down
11 changes: 11 additions & 0 deletions Lib/test/test_configparser.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -526,6 +526,17 @@ def test_default_case_sensitivity(self):
cf.get(self.default_section, "Foo"), "Bar",
"could not locate option, expecting case-insensitive defaults")

def test_crlf_normalization(self):
cf = self.newconfig({"key1": "a\nb","key2": "a\rb", "key3": "a\r\nb", "key4": "a\r\nb"})
buf = io.StringIO()
cf.write(buf)
cf_str = buf.getvalue()
self.assertNotIn("\r", cf_str)
self.assertNotIn("\r\n", cf_str)
self.assertEqual(cf_str.count("\n"), 10)
self.assertEqual(cf_str.count("\n\t"), 4)
self.assertTrue(cf_str.endswith("\n\n"))

def test_parse_errors(self):
cf = self.newconfig()
self.parse_error(cf, configparser.ParsingError,
Expand Down
2 changes: 2 additions & 0 deletions Misc/NEWS.d/next/Security/2026-01-16-11-58-19.gh-issue-143927.aviFeG.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,2 @@
Normalize all line endings (CR, CRLF, and LF) to LF+TAB when writing
multi-line configparser values.
Loading