LibWeb/MimeSniff: Add sniffing in an audio or video context

This commit is contained in:
Kemal Zebari 2023-11-09 15:32:34 -08:00 committed by Andrew Kaster
parent 02ea85da2c
commit 3a820ddbdf
3 changed files with 86 additions and 7 deletions

View file

@ -50,6 +50,16 @@ static void set_image_type_mappings(HashMap<StringView, Vector<StringView>>& mim
mime_type_to_headers_map.set("image/jpeg"sv, { "\xFF\xD8\xFF"sv });
}
static void set_audio_or_video_type_mappings(HashMap<StringView, Vector<StringView>>& mime_type_to_headers_map)
{
mime_type_to_headers_map.set("audio/aiff"sv, { "FORM\x00\x00\x00\x00\x41IFF"sv });
mime_type_to_headers_map.set("audio/mpeg"sv, { "ID3"sv });
mime_type_to_headers_map.set("application/ogg"sv, { "OggS\x00"sv });
mime_type_to_headers_map.set("audio/midi"sv, { "MThd\x00\x00\x00\x06"sv });
mime_type_to_headers_map.set("video/avi"sv, { "RIFF\x00\x00\x00\x00\x41\x56\x49\x20"sv });
mime_type_to_headers_map.set("audio/wave"sv, { "RIFF\x00\x00\x00\x00WAVE"sv });
}
TEST_CASE(determine_computed_mime_type_in_both_none_and_browsing_sniffing_context)
{
HashMap<StringView, Vector<StringView>> mime_type_to_headers_map;
@ -85,13 +95,8 @@ TEST_CASE(determine_computed_mime_type_in_both_none_and_browsing_sniffing_contex
});
set_image_type_mappings(mime_type_to_headers_map);
set_audio_or_video_type_mappings(mime_type_to_headers_map);
mime_type_to_headers_map.set("audio/aiff"sv, { "FORM\x00\x00\x00\x00\x41IFF"sv });
mime_type_to_headers_map.set("audio/mpeg"sv, { "ID3"sv });
mime_type_to_headers_map.set("application/ogg"sv, { "OggS\x00"sv });
mime_type_to_headers_map.set("audio/midi"sv, { "MThd\x00\x00\x00\x06"sv });
mime_type_to_headers_map.set("video/avi"sv, { "RIFF\x00\x00\x00\x00\x41\x56\x49\x20"sv });
mime_type_to_headers_map.set("audio/wave"sv, { "RIFF\x00\x00\x00\x00WAVE"sv });
mime_type_to_headers_map.set("application/x-gzip"sv, { "\x1F\x8B\x08"sv });
mime_type_to_headers_map.set("application/zip"sv, { "PK\x03\x04"sv });
mime_type_to_headers_map.set("application/x-rar-compressed"sv, { "Rar\x20\x1A\x07\x00"sv });
@ -159,3 +164,41 @@ TEST_CASE(determine_computed_mime_type_in_image_sniffing_context)
EXPECT_EQ(mime_type, computed_mime_type.essence());
}
TEST_CASE(determine_computed_mime_type_in_audio_or_video_sniffing_context)
{
// Cover case where supplied type is an XML MIME type.
auto mime_type = "application/rss+xml"sv;
auto supplied_type = MUST(Web::MimeSniff::MimeType::parse(mime_type)).release_value();
auto computed_mime_type = MUST(Web::MimeSniff::Resource::sniff(""sv.bytes(), Web::MimeSniff::SniffingConfiguration {
.sniffing_context = Web::MimeSniff::SniffingContext::AudioOrVideo,
.supplied_type = supplied_type,
}));
EXPECT_EQ(mime_type, MUST(computed_mime_type.serialized()));
HashMap<StringView, Vector<StringView>> mime_type_to_headers_map;
set_audio_or_video_type_mappings(mime_type_to_headers_map);
// Also consider a resource that is not an audio or video.
mime_type_to_headers_map.set("application/octet-stream"sv, { "\x00"sv });
for (auto const& mime_type_to_headers : mime_type_to_headers_map) {
auto mime_type = mime_type_to_headers.key;
for (auto const& header : mime_type_to_headers.value) {
auto computed_mime_type = MUST(Web::MimeSniff::Resource::sniff(header.bytes(), Web::MimeSniff::SniffingConfiguration { .sniffing_context = Web::MimeSniff::SniffingContext::AudioOrVideo }));
EXPECT_EQ(mime_type, computed_mime_type.essence());
}
}
// Cover case where we aren't dealing with an audio or video MIME type.
mime_type = "text/html"sv;
supplied_type = MUST(Web::MimeSniff::MimeType::parse("text/html"sv)).release_value();
computed_mime_type = MUST(Web::MimeSniff::Resource::sniff(""sv.bytes(), Web::MimeSniff::SniffingConfiguration {
.sniffing_context = Web::MimeSniff::SniffingContext::AudioOrVideo,
.supplied_type = supplied_type,
}));
EXPECT_EQ(mime_type, computed_mime_type.essence());
}

View file

@ -540,6 +540,8 @@ ErrorOr<void> Resource::context_specific_sniffing_algorithm(SniffingContext snif
if (sniffing_context == SniffingContext::Image)
return rules_for_sniffing_images_specifically();
if (sniffing_context == SniffingContext::AudioOrVideo)
return rules_for_sniffing_audio_or_video_specifically();
return {};
}
@ -576,4 +578,36 @@ ErrorOr<void> Resource::rules_for_sniffing_images_specifically()
return {};
}
// https://mimesniff.spec.whatwg.org/#sniffing-in-an-audio-or-video-context
ErrorOr<void> Resource::rules_for_sniffing_audio_or_video_specifically()
{
// 1. If the supplied MIME type is an XML MIME type, the computed MIME type is the supplied MIME type.
// Abort these steps.
// NOTE: Non-standard but due to the mime type detection algorithm we need this sanity check.
if (m_supplied_mime_type.has_value() && m_supplied_mime_type->is_xml()) {
m_computed_mime_type = m_supplied_mime_type.value();
return {};
}
// 2. Let audio-or-video-type-matched be the result of executing the audio or video type pattern matching
// algorithm with the resource header as the byte sequence to be matched.
auto audio_or_video_type_matched = TRY(match_an_audio_or_video_type_pattern(resource_header()));
// 3. If audio-or-video-type-matched is not undefined, the computed MIME type is audio-or-video-type-matched.
// Abort these steps.
if (audio_or_video_type_matched.has_value()) {
m_computed_mime_type = audio_or_video_type_matched.release_value();
return {};
}
// 4. The computed MIME type is the supplied MIME type.
// NOTE: Non-standard but due to the mime type detection algorithm we need this sanity check.
if (m_supplied_mime_type.has_value()) {
m_computed_mime_type = m_supplied_mime_type.value();
}
// NOTE: Non-standard but if the supplied mime type is undefined, we use computed mime type's default value.
return {};
}
}

View file

@ -13,7 +13,8 @@ namespace Web::MimeSniff {
enum class SniffingContext {
None,
Browsing,
Image
Image,
AudioOrVideo
};
struct SniffingConfiguration {
@ -42,6 +43,7 @@ private:
ErrorOr<void> mime_type_sniffing_algorithm();
ErrorOr<void> context_specific_sniffing_algorithm(SniffingContext sniffing_context);
ErrorOr<void> rules_for_sniffing_images_specifically();
ErrorOr<void> rules_for_sniffing_audio_or_video_specifically();
// https://mimesniff.spec.whatwg.org/#supplied-mime-type
// A supplied MIME type, the MIME type determined by the supplied MIME type detection algorithm.