Skip to content

Accept surrogates in _json.encode_basestring{,_ascii} #7673

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
youknowone merged 1 commit into from
Apr 24, 2026
Merged

Accept surrogates in _json.encode_basestring{,_ascii} #7673

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
10 changes: 1 addition & 9 deletions Lib/test/test_json/test_unicode.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -138,14 +138,6 @@ def test_object_pairs_hook_with_unicode(self):

class TestPyUnicode(TestUnicode, PyTest): pass
class TestCUnicode(TestUnicode, CTest):
@unittest.expectedFailure # TODO: RUSTPYTHON
def test_ascii_non_printable_encode(self):
return super().test_ascii_non_printable_encode()

@unittest.skip("TODO: RUSTPYTHON; panics with 'str has surrogates'")
@unittest.skip("TODO: RUSTPYTHON; decode path still uses PyUtf8StrRef")
def test_single_surrogate_decode(self):
return super().test_single_surrogate_decode()

@unittest.skip("TODO: RUSTPYTHON; panics with 'str has surrogates'")
def test_single_surrogate_encode(self):
return super().test_single_surrogate_encode()
24 changes: 15 additions & 9 deletions crates/stdlib/src/json.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -667,24 +667,30 @@ mod _json {
}
}

fn encode_string(s: &str, ascii_only: bool) -> String {
fn encode_string(wtf8: &rustpython_common::wtf8::Wtf8, ascii_only: bool) -> Wtf8Buf {
flame_guard!("_json::encode_string");
let mut buf = Vec::<u8>::with_capacity(s.len() + 2);
machinery::write_json_string(s, ascii_only, &mut buf)
let mut buf = Vec::<u8>::with_capacity(wtf8.len() + 2);
machinery::write_json_string(wtf8, ascii_only, &mut buf)
// SAFETY: writing to a vec can't fail
.unwrap_or_else(|_| unsafe { core::hint::unreachable_unchecked() });
// SAFETY: we only output valid utf8 from write_json_string
unsafe { String::from_utf8_unchecked(buf) }
// write_json_string is designed to produce valid WTF-8 bytes:
// - ASCII control characters and JSON-specials are written as ASCII escapes
// - Valid Unicode scalars are written as UTF-8 (a subset of WTF-8)
// - Lone surrogates (ascii_only=false branch only) pass through as the
// input's WTF-8 byte sequences unchanged
// Use the checked constructor so any violation of that invariant
// surfaces as a panic during testing instead of undefined behavior.
Wtf8Buf::from_bytes(buf).expect("write_json_string produced invalid WTF-8")
}

#[pyfunction]
fn encode_basestring(s: PyUtf8StrRef) -> String {
encode_string(s.as_str(), false)
fn encode_basestring(s: PyStrRef) -> Wtf8Buf {
encode_string(s.as_wtf8(), false)
}

#[pyfunction]
fn encode_basestring_ascii(s: PyUtf8StrRef) -> String {
encode_string(s.as_str(), true)
fn encode_basestring_ascii(s: PyStrRef) -> Wtf8Buf {
encode_string(s.as_wtf8(), true)
}

fn py_decode_error(
Expand Down
53 changes: 37 additions & 16 deletions crates/stdlib/src/json/machinery.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,9 +35,9 @@ use rustpython_common::wtf8::{CodePoint, Wtf8, Wtf8Buf};

static ESCAPE_CHARS: [&str; 0x20] = [
"\\u0000", "\\u0001", "\\u0002", "\\u0003", "\\u0004", "\\u0005", "\\u0006", "\\u0007", "\\b",
"\\t", "\\n", "\\u000", "\\f", "\\r", "\\u000e", "\\u000f", "\\u0010", "\\u0011", "\\u0012",
"\\t", "\\n", "\\u000b", "\\f", "\\r", "\\u000e", "\\u000f", "\\u0010", "\\u0011", "\\u0012",
"\\u0013", "\\u0014", "\\u0015", "\\u0016", "\\u0017", "\\u0018", "\\u0019", "\\u001a",
"\\u001", "\\u001c", "\\u001d", "\\u001e", "\\u001f",
"\\u001b", "\\u001c", "\\u001d", "\\u001e", "\\u001f",
];

// This bitset represents which bytes can be copied as-is to a JSON string (0)
Expand Down Expand Up @@ -72,30 +72,51 @@ fn json_escaped_char(c: u8) -> Option<&'static str> {
}
}

pub fn write_json_string<W: io::Write>(s: &str, ascii_only: bool, w: &mut W) -> io::Result<()> {
pub fn write_json_string<W: io::Write>(wtf8: &Wtf8, ascii_only: bool, w: &mut W) -> io::Result<()> {
w.write_all(b"\"")?;
let mut write_start_idx = 0;
let bytes = s.as_bytes();
let bytes = wtf8.as_bytes();
if ascii_only {
for (idx, c) in s.char_indices() {
if c.is_ascii() {
if let Some(escaped) = json_escaped_char(c as u8) {
for (idx, cp) in wtf8.code_point_indices() {
if let Some(c) = cp.to_char() {
// Valid Unicode scalar.
if c.is_ascii() {
if let Some(escaped) = json_escaped_char(c as u8) {
w.write_all(&bytes[write_start_idx..idx])?;
w.write_all(escaped.as_bytes())?;
write_start_idx = idx + 1;
}
} else {
w.write_all(&bytes[write_start_idx..idx])?;
w.write_all(escaped.as_bytes())?;
write_start_idx = idx + 1;
write_start_idx = idx + c.len_utf8();
// codepoints outside the BMP get 2 '\uxxxx' sequences to represent them
for point in c.encode_utf16(&mut [0; 2]) {
write!(w, "\\u{point:04x}")?;
}
}
} else {
// Lone surrogate code point (U+D800..U+DFFF).
// WTF-8 encodes these as 3-byte sequences; skip those raw bytes
// and emit a \uXXXX escape with the surrogate value.
w.write_all(&bytes[write_start_idx..idx])?;
write_start_idx = idx + c.len_utf8();
// codepoints outside the BMP get 2 '\uxxxx' sequences to represent them
for point in c.encode_utf16(&mut [0; 2]) {
write!(w, "\\u{point:04x}")?;
}
write_start_idx = idx + 3;
write!(w, "\\u{:04x}", cp.to_u32())?;
}
}
} else {
for (idx, c) in s.bytes().enumerate() {
if let Some(escaped) = json_escaped_char(c) {
// ensure_ascii is false: only JSON-required escapes (< 0x20, \, ")
// are applied. 0x7F (DEL) is NOT escaped here, matching CPython's
// py_encode_basestring. Multi-byte UTF-8 characters and WTF-8
// surrogate sequences flow through unchanged via the trailing flush,
// so surrogates round-trip as-is (matching CPython behavior).
for (idx, c) in wtf8.as_bytes().iter().enumerate() {
let escaped_opt: Option<&'static str> = match *c {
x if x < 0x20 => Some(ESCAPE_CHARS[x as usize]),
b'\\' => Some("\\\\"),
b'\"' => Some("\\\""),
_ => None,
};
if let Some(escaped) = escaped_opt {
w.write_all(&bytes[write_start_idx..idx])?;
w.write_all(escaped.as_bytes())?;
write_start_idx = idx + 1;
Expand Down
Loading