eblume/kingfisher
1
0
Fork 0
forked from mirrors/kingfisher
Code Pull requests Activity Actions

Merge pull request #182 from mongodb/main

Browse source 
sync with main
This commit is contained in:
Mick Grove 2026-01-14 17:20:19 -08:00 • committed by GitHub
commit 96f585ffa3
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 98 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

34
src/validation/httpvalidation.rs
View file

@ -298,7 +298,13 @@ fn body_looks_like_html(body: &str, headers: &HeaderMap) -> bool {
.unwrap_or(false);
// ---- 2. early-body scan (<=1024 bytes) --------------------------------
let probe = &body[..body.len().min(1024)];
// Find the last character boundary at or before 1024 bytes to avoid UTF-8 boundary issues
// Walk backward at most 3 bytes (UTF-8 max char size is 4 bytes) to find valid boundary
let mut end = 1024.min(body.len());
while end > 0 && !body.is_char_boundary(end) {
end -= 1;
}
let probe = &body[..end];
// Trim any leading whitespace so we still catch HTML that starts after newlines/indentation.
let trimmed = probe.trim_start_matches(|c: char| c.is_whitespace());
let probe = trimmed.to_ascii_lowercase();
@ -570,4 +576,30 @@ mod tests {
assert!(!ok, "HTML responses should be rejected unless explicitly allowed");
}
#[test]
fn test_body_looks_like_html_utf8_boundary() {
// Test case for UTF-8 boundary issue: multi-byte character at 1024-byte boundary
// This reproduces the bug where slicing at byte 1024 would panic if it's in the middle
// of a multi-byte character (e.g., Chinese character '业' spans bytes 1023..1026)
let mut headers = HeaderMap::new();
headers.insert(header::CONTENT_TYPE, HeaderValue::from_static("text/html; charset=utf-8"));
// HTML at the start, with padding to push a multi-byte char to byte 1024
// This mirrors the real crash: HTML response from Gitee with Chinese chars
let html_start = "<!DOCTYPE html><html lang=\"zh-CN\"><head><title>";
let padding_len = 1023 - html_start.len();
let body = format!(
"{}{}业</title></head><body>Gitee</body></html>",
html_start,
"x".repeat(padding_len)
);
// Verify our test setup: multi-byte char should be at byte 1023
assert_eq!(body.as_bytes()[1023], 0xE4, "Expected first byte of '业' at position 1023");
// This should not panic AND should correctly identify HTML
let result = body_looks_like_html(&body, &headers);
assert!(result, "Should correctly identify HTML even with multi-byte characters at boundary");
}
}