Match Brotli and Zstandard content codings as whole tokens

is_brotli_encoding() and is_zstd_encoding() searched the Content-Encoding
value for "br" and "zstd" as substrings, while is_zlib_encoding() beside them
compared the whole value. So "fibre" and "librarian" were read as Brotli and
"x-zstd-ish" as Zstandard, and "gzip, br" -- a value naming two codings, which
cpp-httplib does not support -- was labeled Brotli and run through a Brotli
decompressor over gzip data.

RFC 9110 8.4.1 defines a content coding as a token, so compare the whole value
case-insensitively as the zlib check already does. A value naming several
codings no longer matches any of them and takes the pass-through path
prepare_content_receiver() already documents for an unrecognized coding.

contains_case_ignore() has no callers left.
This commit is contained in:
yhirose
2026-08-18 21:05:22 -04:00
parent 881842cd72
commit e8887d98e9
2 changed files with 48 additions and 9 deletions
+2 -9
View File
@@ -7642,13 +7642,6 @@ inline bool zstd_decompressor::decompress(const char *data, size_t data_length,
}
#endif
inline bool contains_case_ignore(const std::string &s, const char *token) {
auto token_end = token + std::strlen(token);
return std::search(s.begin(), s.end(), token, token_end, [](char a, char b) {
return case_ignore::to_lower(a) == case_ignore::to_lower(b);
}) != s.end();
}
// Content codings are case-insensitive (RFC 9110 8.4.1). Matching them
// case-sensitively would make a response labeled e.g. "GZIP" look like an
// unknown coding, and its payload would be handed back still compressed.
@@ -7658,11 +7651,11 @@ inline bool is_zlib_encoding(const std::string &encoding) {
}
inline bool is_brotli_encoding(const std::string &encoding) {
return contains_case_ignore(encoding, "br");
return case_ignore::equal(encoding, "br");
}
inline bool is_zstd_encoding(const std::string &encoding) {
return contains_case_ignore(encoding, "zstd");
return case_ignore::equal(encoding, "zstd");
}
// Returns true if the content coding is one cpp-httplib is able to decompress
+46
View File
@@ -11049,6 +11049,52 @@ static const char GZIPPED_HELLO_WORLD[] = {
// A content coding cpp-httplib recognizes but was not built with must be
// reported as such. Handing the still-compressed payload back to the caller
// would silently corrupt it.
// A content coding is a whole token, not something the field value merely
// contains. Matching "br" and "zstd" as substrings made unrelated values such
// as "fibre" look like Brotli, and turned a multi-coding value like
// "gzip, br" into a Brotli-labeled body, so a decompressor was run over data
// it was never meant to see.
TEST(ContentEncodingTest, SubstringOfACodingIsNotTheCoding) {
const std::string body = "\xff\xd8\xff\xe0 not really a jpeg";
Server svr;
svr.Get("/fibre", [&](const Request & /*req*/, Response &res) {
res.set_content(body, "image/jpeg");
res.set_header("Content-Encoding", "fibre");
});
svr.Get("/multi", [&](const Request & /*req*/, Response &res) {
res.set_content(body, "image/jpeg");
res.set_header("Content-Encoding", "gzip, br");
});
svr.Get("/zstdish", [&](const Request & /*req*/, Response &res) {
res.set_content(body, "image/jpeg");
res.set_header("Content-Encoding", "x-zstd-ish");
});
auto port = svr.bind_to_any_port(HOST);
thread t = thread([&]() { svr.listen_after_bind(); });
auto se = detail::scope_exit([&] {
svr.stop();
t.join();
ASSERT_FALSE(svr.is_running());
});
svr.wait_until_ready();
Client cli(HOST, port);
for (const char *path : {"/fibre", "/multi", "/zstdish"}) {
auto res = cli.Get(path);
ASSERT_TRUE(res) << path << " -> " << to_string(res.error());
EXPECT_EQ(StatusCode::OK_200, res->status) << path;
// Not a coding we recognize, so the payload comes back untouched
EXPECT_EQ(body, res->body) << path;
}
}
TEST(ContentEncodingTest, KnownEncodingWithoutSupportIsReported) {
const std::string gzipped(GZIPPED_HELLO_WORLD, sizeof(GZIPPED_HELLO_WORLD));