use tracing::debug;
use crate::http::header_processor::HttpResponseHead;
use aria2_protocol::http::content_disposition::parse_content_disposition_with_default_utf8;
pub fn determine_filename(
response_head: &HttpResponseHead,
request_url: &str,
content_disposition_default_utf8: bool,
) -> String {
if let Some(filename) =
content_disposition_filename(response_head, content_disposition_default_utf8)
{
debug!(
filename = %filename,
source = "Content-Disposition",
"Filename determined"
);
return filename;
}
let filename = crate::validation::uri::sanitize_filename_from_uri(request_url);
debug!(filename = %filename, source = "URL", "Filename determined");
filename
}
pub(crate) fn determine_filename_from_response(
response: &reqwest::Response,
content_disposition_default_utf8: bool,
) -> String {
let headers = response
.headers()
.iter()
.filter_map(|(name, value)| {
Some((
name.as_str().to_ascii_lowercase(),
value.to_str().ok()?.to_owned(),
))
})
.collect();
let head = HttpResponseHead::from_parts(
format!("{:?}", response.version()),
response.status().as_u16(),
response
.status()
.canonical_reason()
.unwrap_or_default()
.to_owned(),
headers,
);
determine_filename(
&head,
response.url().as_str(),
content_disposition_default_utf8,
)
}
fn content_disposition_filename(
response_head: &HttpResponseHead,
default_utf8: bool,
) -> Option<String> {
let values = response_head.header_all("content-disposition");
let value = match values.as_slice() {
[] => return None,
[value] => value,
_ => {
debug!(
count = values.len(),
"Ignoring duplicate Content-Disposition headers"
);
return None;
}
};
let candidate = parse_content_disposition_filename(value, default_utf8)?;
crate::validation::uri::sanitize_filename_candidate(&candidate)
}
fn parse_content_disposition_filename(cd_value: &str, default_utf8: bool) -> Option<String> {
let result = parse_content_disposition_with_default_utf8(cd_value, default_utf8);
if result.disposition_type.is_empty() {
return None;
}
let filename = result.filename?;
if filename.contains('/') || filename.contains('\\') {
debug!(
filename = %filename,
"Content-Disposition filename rejected: contains path separator"
);
return None;
}
Some(filename)
}
#[cfg(test)]
mod tests {
use super::*;
use crate::http::header_processor::HttpHeaderProcessor;
use crate::validation::uri::sanitize_filename_candidate as sanitize_filename;
fn parse_head(raw: &[u8]) -> HttpResponseHead {
let mut proc = HttpHeaderProcessor::new();
proc.feed(raw);
proc.get_result().unwrap()
}
#[test]
fn test_filename_from_url_path() {
let head = parse_head(b"HTTP/1.1 200 OK\r\nContent-Length: 100\r\n\r\n");
let filename = determine_filename(&head, "http://example.com/path/to/file.txt", false);
assert_eq!(filename, "file.txt");
}
#[test]
fn test_filename_from_url_trailing_slash() {
let head = parse_head(b"HTTP/1.1 200 OK\r\nContent-Length: 100\r\n\r\n");
let filename = determine_filename(&head, "http://example.com/dir/", false);
assert_eq!(filename, "index.html");
}
#[test]
fn test_filename_from_url_percent_encoded() {
let head = parse_head(b"HTTP/1.1 200 OK\r\nContent-Length: 100\r\n\r\n");
let filename = determine_filename(&head, "http://example.com/path/my%20file.txt", false);
assert_eq!(filename, "my file.txt");
}
#[test]
fn test_filename_from_url_utf8_percent_encoded() {
let head = parse_head(b"HTTP/1.1 200 OK\r\nContent-Length: 100\r\n\r\n");
let filename = determine_filename(
&head,
"http://example.com/%E6%97%A5%E6%9C%AC%E8%AA%9E.txt",
false,
);
assert_eq!(filename, "\u{65e5}\u{672c}\u{8a9e}.txt");
}
#[test]
fn test_filename_from_url_decodes_only_after_extracting_segment() {
let head = parse_head(b"HTTP/1.1 200 OK\r\nContent-Length: 100\r\n\r\n");
let filename = determine_filename(
&head,
"https://example.com/archive/dir%2Freport.txt?token=ignored#fragment",
false,
);
assert_eq!(filename, "report.txt");
}
#[test]
fn test_filename_from_url_preserves_invalid_utf8_as_replacement() {
let head = parse_head(b"HTTP/1.1 200 OK\r\nContent-Length: 100\r\n\r\n");
let filename = determine_filename(&head, "https://example.com/%FF.bin", false);
assert_eq!(filename, "\u{fffd}.bin");
}
#[test]
fn test_filename_from_content_disposition_quoted() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename=\"my file.pdf\"\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/download", false);
assert_eq!(filename, "my file.pdf");
}
#[test]
fn test_filename_from_content_disposition_unquoted() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename=report.csv\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/download", false);
assert_eq!(filename, "report.csv");
}
#[test]
fn test_filename_from_content_disposition_star() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename*=UTF-8''my%20doc.txt\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/download", false);
assert_eq!(filename, "my doc.txt");
}
#[test]
fn test_filename_content_disposition_priority_over_url() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: inline; filename=\"override.txt\"\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/original.txt", false);
assert_eq!(filename, "override.txt");
}
#[test]
fn test_filename_star_priority_over_filename() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename=\"fallback.txt\"; filename*=UTF-8''preferred.txt\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/download", false);
assert_eq!(filename, "preferred.txt");
}
#[test]
fn test_filename_star_with_cjk() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename*=UTF-8''%e3%81%93%e3%82%93%e3%81%ab%e3%81%a1%e3%81%af.txt\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/download", false);
assert_eq!(filename, "\u{3053}\u{3093}\u{306b}\u{3061}\u{306f}.txt");
}
#[test]
fn test_content_disposition_path_separator_rejected() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename=\"subdir/file.txt\"\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/original.txt", false);
assert_eq!(filename, "original.txt");
}
#[test]
fn test_content_disposition_backslash_rejected() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename=\"dir\\\\file.txt\"\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/original.txt", false);
assert_eq!(filename, "original.txt");
}
#[test]
fn test_content_disposition_directory_traversal_rejected() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename=\"../etc/passwd\"\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/original.txt", false);
assert_eq!(filename, "original.txt");
}
#[test]
fn test_duplicate_filename_rejected() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename=first.txt; filename=second.txt\r\n\r\n",
);
let filename = determine_filename(&head, "http://example.com/original.txt", false);
assert_eq!(filename, "original.txt");
}
#[test]
fn test_sanitize_filename() {
assert_eq!(
sanitize_filename("normal.txt"),
Some("normal.txt".to_string())
);
assert_eq!(
sanitize_filename("path/to/file.txt"),
Some("file.txt".to_string())
);
assert_eq!(sanitize_filename("win\\path"), Some("path".to_string()));
assert_eq!(
sanitize_filename("null\0byte"),
Some("nullbyte".to_string())
);
assert_eq!(sanitize_filename("CON.txt"), Some("_CON.txt".to_string()));
assert_eq!(sanitize_filename("CON "), Some("_CON".to_string()));
assert_eq!(sanitize_filename("COM1."), Some("_COM1".to_string()));
assert_eq!(sanitize_filename(".."), None);
}
#[test]
fn test_sanitize_filename_handles_reserved_names_and_utf8_length() {
for (input, expected) in [
("PRN.txt", "_PRN.txt"),
("AUX", "_AUX"),
("NUL.log", "_NUL.log"),
("LPT9.csv", "_LPT9.csv"),
("COM0.txt", "COM0.txt"),
] {
assert_eq!(sanitize_filename(input), Some(expected.to_owned()));
}
let filename = sanitize_filename(&"界".repeat(100)).unwrap();
assert_eq!(filename, "界".repeat(85));
assert_eq!(filename.len(), 255);
}
#[test]
fn test_content_disposition_default_utf8_is_honored() {
let head = parse_head(
"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename=\"café.txt\"\r\n\r\n"
.as_bytes(),
);
assert_eq!(
determine_filename(&head, "https://example.com/fallback.txt", true),
"café.txt"
);
assert_eq!(
determine_filename(&head, "https://example.com/fallback.txt", false),
"café.txt"
);
}
#[test]
fn test_duplicate_content_disposition_headers_fall_back_to_url() {
let head = parse_head(
b"HTTP/1.1 200 OK\r\nContent-Disposition: attachment; filename=first.txt\r\nContent-Disposition: attachment; filename=second.txt\r\n\r\n",
);
assert_eq!(
determine_filename(&head, "https://example.com/fallback.txt", false),
"fallback.txt"
);
}
#[test]
fn test_no_content_disposition_uses_url() {
let head = parse_head(b"HTTP/1.1 200 OK\r\nContent-Length: 100\r\n\r\n");
let filename = determine_filename(&head, "http://example.com/data.csv", false);
assert_eq!(filename, "data.csv");
}
#[test]
fn test_empty_content_disposition_uses_url() {
let head =
parse_head(b"HTTP/1.1 200 OK\r\nContent-Disposition: \r\nContent-Length: 100\r\n\r\n");
let filename = determine_filename(&head, "http://example.com/data.csv", false);
assert_eq!(filename, "data.csv");
}
}