ladybird/Libraries/LibGfx/ImageFormats/TIFFLoader.cpp

810 lines
32 KiB
C++

/*
* Copyright (c) 2023, Lucas Chollet <lucas.chollet@serenityos.org>
*
* SPDX-License-Identifier: BSD-2-Clause
*/
#include "TIFFLoader.h"
#include <AK/ConstrainedStream.h>
#include <AK/Debug.h>
#include <AK/Endian.h>
#include <AK/String.h>
#include <LibCompress/Lzw.h>
#include <LibCompress/PackBitsDecoder.h>
#include <LibCompress/Zlib.h>
#include <LibGfx/CMYKBitmap.h>
#include <LibGfx/ImageFormats/CCITTDecoder.h>
#include <LibGfx/ImageFormats/ExifOrientedBitmap.h>
#include <LibGfx/ImageFormats/TIFFMetadata.h>
namespace Gfx {
namespace {
CCITT::Group3Options parse_t4_options(u32 bit_field)
{
// Section 11: CCITT Bilevel Encodings
CCITT::Group3Options options {};
if (bit_field & 0b001)
options.dimensions = CCITT::Group3Options::Mode::TwoDimensions;
if (bit_field & 0b010)
options.compression = CCITT::Group3Options::Compression::Uncompressed;
if (bit_field & 0b100)
options.use_fill_bits = CCITT::Group3Options::UseFillBits::Yes;
return options;
}
bool is_bilevel(TIFF::PhotometricInterpretation interpretation)
{
return interpretation == TIFF::PhotometricInterpretation::WhiteIsZero || interpretation == TIFF::PhotometricInterpretation::BlackIsZero;
}
}
namespace TIFF {
class TIFFLoadingContext {
public:
enum class State {
NotDecoded = 0,
Error,
HeaderDecoded,
FrameDecoded,
};
TIFFLoadingContext(NonnullOwnPtr<FixedMemoryStream> stream)
: m_stream(move(stream))
{
}
ErrorOr<void> decode_image_header()
{
TRY(read_image_file_header());
TRY(read_next_image_file_directory());
m_state = State::HeaderDecoded;
return {};
}
ErrorOr<void> ensure_conditional_tags_are_correct() const
{
if (m_metadata.photometric_interpretation() == PhotometricInterpretation::RGBPalette && !m_metadata.color_map().has_value())
return Error::from_string_literal("TIFFImageDecoderPlugin: RGBPalette image doesn't contain a color map");
if (m_metadata.tile_width() == 0u || m_metadata.tile_length() == 0u)
return Error::from_string_literal("TIFFImageDecoderPlugin: Null value in tile's dimensions");
return {};
}
Optional<Vector<u32>> segment_offsets() const
{
return m_metadata.strip_offsets().has_value() ? m_metadata.strip_offsets() : m_metadata.tile_offsets();
}
Optional<Vector<u32>> segment_byte_counts() const
{
return m_metadata.strip_byte_counts().has_value() ? m_metadata.strip_byte_counts() : m_metadata.tile_byte_counts();
}
bool is_tiled() const
{
return m_metadata.tile_width().has_value() && m_metadata.tile_length().has_value();
}
ErrorOr<void> ensure_baseline_tags_are_correct() const
{
if (!segment_offsets().has_value())
return Error::from_string_literal("TIFFImageDecoderPlugin: Missing Offsets tag");
if (!segment_byte_counts().has_value())
return Error::from_string_literal("TIFFImageDecoderPlugin: Missing ByteCounts tag");
if (segment_offsets()->size() != segment_byte_counts()->size())
return Error::from_string_literal("TIFFImageDecoderPlugin: StripsOffset and StripByteCount have different sizes");
if (!m_metadata.rows_per_strip().has_value() && segment_byte_counts()->size() != 1 && !is_tiled())
return Error::from_string_literal("TIFFImageDecoderPlugin: RowsPerStrip is not provided and impossible to deduce");
if (!is_bilevel(*m_metadata.photometric_interpretation())) {
if (!m_metadata.bits_per_sample().has_value())
return Error::from_string_literal("TIFFImageDecoderPlugin: Tag BitsPerSample is missing");
if (!m_metadata.samples_per_pixel().has_value())
return Error::from_string_literal("TIFFImageDecoderPlugin: Tag SamplesPerPixel is missing");
if (any_of(*m_metadata.bits_per_sample(), [](auto bit_depth) { return bit_depth == 0 || bit_depth > 32; }))
return Error::from_string_literal("TIFFImageDecoderPlugin: Invalid value in BitsPerSample");
if (m_metadata.bits_per_sample()->size() != m_metadata.samples_per_pixel())
return Error::from_string_literal("TIFFImageDecoderPlugin: Invalid number of values in BitsPerSample");
if (*m_metadata.samples_per_pixel() < samples_for_photometric_interpretation())
return Error::from_string_literal("TIFFImageDecoderPlugin: Not enough values in BitsPerSample for given PhotometricInterpretation");
}
return {};
}
void cache_values()
{
if (m_metadata.photometric_interpretation().has_value())
m_photometric_interpretation = m_metadata.photometric_interpretation().value();
if (m_metadata.bits_per_sample().has_value())
m_bits_per_sample = m_metadata.bits_per_sample().value();
else if (is_bilevel(m_photometric_interpretation))
m_bits_per_sample.append(1);
if (m_metadata.image_width().has_value())
m_image_width = m_metadata.image_width().value();
if (m_metadata.predictor().has_value())
m_predictor = m_metadata.predictor().value();
m_alpha_channel_index = alpha_channel_index();
}
ErrorOr<void> decode_frame()
{
TRY(ensure_baseline_tags_are_present(m_metadata));
TRY(ensure_baseline_tags_are_correct());
TRY(ensure_conditional_tags_are_correct());
cache_values();
auto maybe_error = decode_frame_impl();
if (maybe_error.is_error()) {
m_state = State::Error;
return maybe_error.release_error();
}
return {};
}
IntSize size() const
{
return ExifOrientedBitmap::oriented_size({ *m_metadata.image_width(), *m_metadata.image_length() }, *m_metadata.orientation());
}
ExifMetadata const& metadata() const
{
return m_metadata;
}
State state() const
{
return m_state;
}
RefPtr<CMYKBitmap> cmyk_bitmap() const
{
return m_cmyk_bitmap;
}
RefPtr<Bitmap> bitmap() const
{
return m_bitmap;
}
private:
enum class ByteOrder {
LittleEndian,
BigEndian,
};
static ErrorOr<u8> read_component(BigEndianInputBitStream& stream, u8 bits)
{
// FIXME: This function truncates everything to 8-bits
auto const value = TRY(stream.read_bits<u32>(bits));
if (bits > 8)
return value >> (bits - 8);
return NumericLimits<u8>::max() * value / ((1 << bits) - 1);
}
u8 samples_for_photometric_interpretation() const
{
switch (m_photometric_interpretation) {
case PhotometricInterpretation::WhiteIsZero:
case PhotometricInterpretation::BlackIsZero:
case PhotometricInterpretation::RGBPalette:
return 1;
case PhotometricInterpretation::RGB:
return 3;
case PhotometricInterpretation::CMYK:
return 4;
default:
TODO();
}
}
Optional<u8> alpha_channel_index() const
{
if (m_metadata.extra_samples().has_value()) {
auto const extra_samples = m_metadata.extra_samples().value();
for (u8 i = 0; i < extra_samples.size(); ++i) {
if (extra_samples[i] == ExtraSample::UnassociatedAlpha)
return i + samples_for_photometric_interpretation();
}
}
return OptionalNone {};
}
ErrorOr<u8> manage_extra_channels(BigEndianInputBitStream& stream) const
{
// Section 7: Additional Baseline TIFF Requirements
// Some TIFF files may have more components per pixel than you think. A Baseline TIFF reader must skip over
// them gracefully, using the values of the SamplesPerPixel and BitsPerSample fields.
// Both unknown and alpha channels are considered as extra channels, so let's iterate over
// them, conserve the alpha value (if any) and discard everything else.
auto const number_base_channels = samples_for_photometric_interpretation();
Optional<u8> alpha {};
for (u8 i = number_base_channels; i < m_bits_per_sample.size(); ++i) {
if (m_alpha_channel_index == i)
alpha = TRY(read_component(stream, m_bits_per_sample[i]));
else
TRY(read_component(stream, m_bits_per_sample[i]));
}
return alpha.value_or(NumericLimits<u8>::max());
}
ErrorOr<Color> read_color(BigEndianInputBitStream& stream)
{
if (m_photometric_interpretation == PhotometricInterpretation::RGB) {
auto const first_component = TRY(read_component(stream, m_bits_per_sample[0]));
auto const second_component = TRY(read_component(stream, m_bits_per_sample[1]));
auto const third_component = TRY(read_component(stream, m_bits_per_sample[2]));
auto const alpha = TRY(manage_extra_channels(stream));
return Color(first_component, second_component, third_component, alpha);
}
if (m_photometric_interpretation == PhotometricInterpretation::RGBPalette) {
auto const index = TRY(stream.read_bits<u16>(m_bits_per_sample[0]));
auto const alpha = TRY(manage_extra_channels(stream));
// SamplesPerPixel == 1 is a requirement for RGBPalette
// From description of PhotometricInterpretation in Section 8: Baseline Field Reference Guide
// "In a TIFF ColorMap, all the Red values come first, followed by the Green values,
// then the Blue values."
u64 const size = 1ul << m_bits_per_sample[0];
u64 const red_offset = 0 * size;
u64 const green_offset = 1 * size;
u64 const blue_offset = 2 * size;
auto const color_map = *m_metadata.color_map();
if (blue_offset + index >= color_map.size())
return Error::from_string_literal("TIFFImageDecoderPlugin: Color index is out of range");
// FIXME: ColorMap's values are always 16-bits, stop truncating them when we support 16 bits bitmaps
return Color(
color_map[red_offset + index] >> 8,
color_map[green_offset + index] >> 8,
color_map[blue_offset + index] >> 8,
alpha);
}
if (m_photometric_interpretation == PhotometricInterpretation::WhiteIsZero
|| m_photometric_interpretation == PhotometricInterpretation::BlackIsZero) {
auto luminosity = TRY(read_component(stream, m_bits_per_sample[0]));
if (m_photometric_interpretation == PhotometricInterpretation::WhiteIsZero)
luminosity = ~luminosity;
auto const alpha = TRY(manage_extra_channels(stream));
return Color(luminosity, luminosity, luminosity, alpha);
}
return Error::from_string_literal("Unsupported value for PhotometricInterpretation");
}
ErrorOr<CMYK> read_color_cmyk(BigEndianInputBitStream& stream)
{
VERIFY(m_photometric_interpretation == PhotometricInterpretation::CMYK);
auto const first_component = TRY(read_component(stream, m_bits_per_sample[0]));
auto const second_component = TRY(read_component(stream, m_bits_per_sample[1]));
auto const third_component = TRY(read_component(stream, m_bits_per_sample[2]));
auto const fourth_component = TRY(read_component(stream, m_bits_per_sample[3]));
// FIXME: We probably won't encounter CMYK images with an alpha channel, but if
// we do: the first step to support them is not dropping the value here!
[[maybe_unused]] auto const alpha = TRY(manage_extra_channels(stream));
return CMYK { first_component, second_component, third_component, fourth_component };
}
template<CallableAs<ErrorOr<ReadonlyBytes>, u32, IntSize> SegmentDecoder>
ErrorOr<void> loop_over_pixels(SegmentDecoder&& segment_decoder)
{
auto const offsets = *segment_offsets();
auto const byte_counts = *segment_byte_counts();
auto const segment_length = m_metadata.tile_length().value_or(m_metadata.rows_per_strip().value_or(*m_metadata.image_length()));
auto const segment_width = m_metadata.tile_width().value_or(m_image_width);
auto const segment_per_rows = m_metadata.tile_width().map([&](u32 w) { return ceil_div(m_image_width, w); }).value_or(1);
Variant<ExifOrientedBitmap, ExifOrientedCMYKBitmap> oriented_bitmap = TRY(([&]() -> ErrorOr<Variant<ExifOrientedBitmap, ExifOrientedCMYKBitmap>> {
if (m_photometric_interpretation == PhotometricInterpretation::CMYK)
return ExifOrientedCMYKBitmap::create(*metadata().orientation(), { m_image_width, *metadata().image_length() });
return ExifOrientedBitmap::create(*metadata().orientation(), { m_image_width, *metadata().image_length() }, BitmapFormat::BGRA8888);
}()));
for (u32 segment_index = 0; segment_index < offsets.size(); ++segment_index) {
TRY(m_stream->seek(offsets[segment_index]));
auto const rows_in_segment = segment_index < offsets.size() - 1 ? segment_length : *m_metadata.image_length() - segment_length * segment_index;
auto const decoded_bytes = TRY(segment_decoder(byte_counts[segment_index], { segment_width, rows_in_segment }));
auto decoded_segment = make<FixedMemoryStream>(decoded_bytes);
auto decoded_stream = make<BigEndianInputBitStream>(move(decoded_segment));
for (u32 row = 0; row < segment_length; row++) {
auto const image_row = row + segment_length * (segment_index / segment_per_rows);
if (image_row >= *m_metadata.image_length())
break;
Optional<Color> last_color {};
for (u32 column = 0; column < segment_width; ++column) {
// If image_length % segment_length != 0, the last tile will be padded.
// This variable helps us to skip these last columns. Note that we still
// need to read the sample from the stream.
auto const image_column = column + segment_width * (segment_index % segment_per_rows);
if (m_photometric_interpretation == PhotometricInterpretation::CMYK) {
auto const cmyk = TRY(read_color_cmyk(*decoded_stream));
if (image_column >= m_image_width)
continue;
oriented_bitmap.get<ExifOrientedCMYKBitmap>().set_pixel(image_column, image_row, cmyk);
} else {
auto color = TRY(read_color(*decoded_stream));
// FIXME: We should do the differencing at the byte-stream level, that would make it
// compatible with both LibPDF and all color formats.
if (m_predictor == Predictor::HorizontalDifferencing && last_color.has_value()) {
color.set_red(last_color->red() + color.red());
color.set_green(last_color->green() + color.green());
color.set_blue(last_color->blue() + color.blue());
if (m_alpha_channel_index.has_value())
color.set_alpha(last_color->alpha() + color.alpha());
}
last_color = color;
if (image_column >= m_image_width)
continue;
oriented_bitmap.get<ExifOrientedBitmap>().set_pixel(image_column, image_row, color.value());
}
}
decoded_stream->align_to_byte_boundary();
}
}
if (m_photometric_interpretation == PhotometricInterpretation::CMYK)
m_cmyk_bitmap = oriented_bitmap.get<ExifOrientedCMYKBitmap>().bitmap();
else
m_bitmap = oriented_bitmap.get<ExifOrientedBitmap>().bitmap();
return {};
}
ErrorOr<void> ensure_tags_are_correct_for_ccitt() const
{
// Section 8: Baseline Field Reference Guide
// BitsPerSample must be 1, since this type of compression is defined only for bilevel images.
if (m_bits_per_sample.size() > 1)
return Error::from_string_literal("TIFFImageDecoderPlugin: CCITT image with BitsPerSample greater than one");
if (!is_bilevel(*m_metadata.photometric_interpretation()))
return Error::from_string_literal("TIFFImageDecoderPlugin: CCITT compression is used on a non bilevel image");
return {};
}
ErrorOr<ByteBuffer> read_bytes_considering_fill_order(u32 bytes_to_read) const
{
auto const reverse_byte = [](u8 b) {
b = (b & 0xF0) >> 4 | (b & 0x0F) << 4;
b = (b & 0xCC) >> 2 | (b & 0x33) << 2;
b = (b & 0xAA) >> 1 | (b & 0x55) << 1;
return b;
};
auto const bytes = TRY(m_stream->read_in_place<u8 const>(bytes_to_read));
auto copy = TRY(ByteBuffer::copy(bytes));
if (m_metadata.fill_order() == FillOrder::RightToLeft) {
for (auto& byte : copy.bytes())
byte = reverse_byte(byte);
}
return copy;
}
ErrorOr<void> decode_frame_impl()
{
switch (*m_metadata.compression()) {
case Compression::NoCompression: {
auto identity = [&](u32 num_bytes, IntSize) {
return m_stream->read_in_place<u8 const>(num_bytes);
};
TRY(loop_over_pixels(move(identity)));
break;
}
case Compression::CCITTRLE: {
TRY(ensure_tags_are_correct_for_ccitt());
ByteBuffer decoded_bytes {};
auto decode_ccitt_rle_segment = [&](u32 num_bytes, IntSize segment_size) -> ErrorOr<ReadonlyBytes> {
auto const encoded_bytes = TRY(read_bytes_considering_fill_order(num_bytes));
decoded_bytes = TRY(CCITT::decode_ccitt_rle(encoded_bytes, segment_size.width(), segment_size.height()));
return decoded_bytes;
};
TRY(loop_over_pixels(move(decode_ccitt_rle_segment)));
break;
}
case Compression::Group3Fax: {
TRY(ensure_tags_are_correct_for_ccitt());
auto const parameters = parse_t4_options(*m_metadata.t4_options());
ByteBuffer decoded_bytes {};
auto decode_group3_segment = [&](u32 num_bytes, IntSize segment_size) -> ErrorOr<ReadonlyBytes> {
auto const encoded_bytes = TRY(read_bytes_considering_fill_order(num_bytes));
decoded_bytes = TRY(CCITT::decode_ccitt_group3(encoded_bytes, segment_size.width(), segment_size.height(), parameters));
return decoded_bytes;
};
TRY(loop_over_pixels(move(decode_group3_segment)));
break;
}
case Compression::Group4Fax: {
TRY(ensure_tags_are_correct_for_ccitt());
// FIXME: We need to parse T6 options
ByteBuffer decoded_bytes {};
auto decode_group3_segment = [&](u32 num_bytes, IntSize segment_size) -> ErrorOr<ReadonlyBytes> {
auto const encoded_bytes = TRY(read_bytes_considering_fill_order(num_bytes));
decoded_bytes = TRY(CCITT::decode_ccitt_group4(encoded_bytes, segment_size.width(), segment_size.height()));
return decoded_bytes;
};
TRY(loop_over_pixels(move(decode_group3_segment)));
break;
}
case Compression::LZW: {
ByteBuffer decoded_bytes {};
auto decode_lzw_segment = [&](u32 num_bytes, IntSize) -> ErrorOr<ReadonlyBytes> {
auto const encoded_bytes = TRY(m_stream->read_in_place<u8 const>(num_bytes));
if (encoded_bytes.is_empty())
return Error::from_string_literal("TIFFImageDecoderPlugin: Unable to read from empty LZW segment");
// Note: AFAIK, there are two common ways to use LZW compression:
// - With a LittleEndian stream and no Early-Change, this is used in the GIF format
// - With a BigEndian stream and an EarlyChange of 1, this is used in the PDF format
// The fun begins when they decided to change from the former to the latter when moving
// from TIFF 5.0 to 6.0, and without including a way for files to be identified.
// Fortunately, as the first byte of a LZW stream is a constant we can guess the endianess
// and deduce the version from it. The first code is 0x100 (9-bits).
if (encoded_bytes[0] == 0x00)
decoded_bytes = TRY(Compress::LzwDecompressor<LittleEndianInputBitStream>::decompress_all(encoded_bytes, 8, 0));
else
decoded_bytes = TRY(Compress::LzwDecompressor<BigEndianInputBitStream>::decompress_all(encoded_bytes, 8, -1));
return decoded_bytes;
};
TRY(loop_over_pixels(move(decode_lzw_segment)));
break;
}
case Compression::AdobeDeflate:
case Compression::PixarDeflate: {
// This is an extension from the Technical Notes from 2002:
// https://web.archive.org/web/20160305055905/http://partners.adobe.com/public/developer/en/tiff/TIFFphotoshop.pdf
ByteBuffer decoded_bytes {};
auto decode_zlib = [&](u32 num_bytes, IntSize) -> ErrorOr<ReadonlyBytes> {
auto stream = make<ConstrainedStream>(MaybeOwned<Stream>(*m_stream), num_bytes);
auto decompressed_stream = TRY(Compress::ZlibDecompressor::create(move(stream)));
decoded_bytes = TRY(decompressed_stream->read_until_eof(4096));
return decoded_bytes;
};
TRY(loop_over_pixels(move(decode_zlib)));
break;
}
case Compression::PackBits: {
// Section 9: PackBits Compression
ByteBuffer decoded_bytes {};
auto decode_packbits_segment = [&](u32 num_bytes, IntSize) -> ErrorOr<ReadonlyBytes> {
auto const encoded_bytes = TRY(m_stream->read_in_place<u8 const>(num_bytes));
decoded_bytes = TRY(Compress::PackBits::decode_all(encoded_bytes));
return decoded_bytes;
};
TRY(loop_over_pixels(move(decode_packbits_segment)));
break;
}
default:
return Error::from_string_literal("This compression type is not supported yet :^)");
}
return {};
}
template<typename T>
ErrorOr<T> read_value()
{
if (m_byte_order == ByteOrder::LittleEndian)
return TRY(m_stream->read_value<LittleEndian<T>>());
if (m_byte_order == ByteOrder::BigEndian)
return TRY(m_stream->read_value<BigEndian<T>>());
VERIFY_NOT_REACHED();
}
ErrorOr<void> set_next_ifd(u32 ifd_offset)
{
if (ifd_offset != 0) {
if (ifd_offset < TRY(m_stream->tell()))
return Error::from_string_literal("TIFFImageDecoderPlugin: Can not accept an IFD pointing to previous data");
m_next_ifd = Optional<u32> { ifd_offset };
} else {
m_next_ifd = OptionalNone {};
}
return {};
}
ErrorOr<void> read_next_idf_offset()
{
auto const next_block_position = TRY(read_value<u32>());
TRY(set_next_ifd(next_block_position));
return {};
}
ErrorOr<void> read_image_file_header()
{
// Section 2: TIFF Structure - Image File Header
auto const byte_order = TRY(m_stream->read_value<u16>());
switch (byte_order) {
case 0x4949:
m_byte_order = ByteOrder::LittleEndian;
break;
case 0x4D4D:
m_byte_order = ByteOrder::BigEndian;
break;
default:
return Error::from_string_literal("TIFFImageDecoderPlugin: Invalid byte order");
}
auto const magic_number = TRY(read_value<u16>());
if (magic_number != 42)
return Error::from_string_literal("TIFFImageDecoderPlugin: Invalid magic number");
TRY(read_next_idf_offset());
return {};
}
ErrorOr<void> read_next_image_file_directory()
{
// Section 2: TIFF Structure - Image File Directory
if (!m_next_ifd.has_value())
return Error::from_string_literal("TIFFImageDecoderPlugin: Missing an Image File Directory");
dbgln_if(TIFF_DEBUG, "Reading image file directory at offset {}", m_next_ifd);
TRY(m_stream->seek(m_next_ifd.value()));
auto const number_of_field = TRY(read_value<u16>());
auto next_tag_offset = TRY(m_stream->tell());
for (u16 i = 0; i < number_of_field; ++i) {
if (auto maybe_error = read_tag(); maybe_error.is_error() && TIFF_DEBUG)
dbgln("Unable to decode tag {}/{}", i + 1, number_of_field);
// Section 2: TIFF Structure
// IFD Entry
// Size of tag(u16) + type(u16) + count(u32) + value_or_offset(u32) = 12
next_tag_offset += 12;
TRY(m_stream->seek(next_tag_offset));
}
TRY(read_next_idf_offset());
return {};
}
ErrorOr<Vector<Value, 1>> read_tiff_value(Type type, u32 count, u32 offset)
{
auto const old_offset = TRY(m_stream->tell());
ScopeGuard reset_offset { [this, old_offset]() { MUST(m_stream->seek(old_offset)); } };
TRY(m_stream->seek(offset));
if (size_of_type(type) * count > m_stream->remaining())
return Error::from_string_literal("TIFFImageDecoderPlugin: Tag size claims to be bigger that remaining bytes");
auto const read_every_values = [this, count]<typename T>() -> ErrorOr<Vector<Value>> {
Vector<Value, 1> result {};
TRY(result.try_ensure_capacity(count));
if constexpr (IsSpecializationOf<T, Rational>) {
for (u32 i = 0; i < count; ++i)
result.empend(T { TRY(read_value<typename T::Type>()), TRY(read_value<typename T::Type>()) });
} else {
for (u32 i = 0; i < count; ++i)
result.empend(typename TypePromoter<T>::Type(TRY(read_value<T>())));
}
return result;
};
switch (type) {
case Type::Byte:
case Type::Undefined: {
Vector<Value, 1> result;
auto buffer = TRY(ByteBuffer::create_uninitialized(count));
TRY(m_stream->read_until_filled(buffer));
result.append(move(buffer));
return result;
}
case Type::ASCII:
case Type::UTF8: {
Vector<Value, 1> result;
// NOTE: No need to include the null terminator
if (count > 0)
--count;
auto string_data = TRY(ByteBuffer::create_uninitialized(count));
TRY(m_stream->read_until_filled(string_data));
result.empend(TRY(String::from_utf8(StringView { string_data.bytes() })));
return result;
}
case Type::UnsignedShort:
return read_every_values.template operator()<u16>();
case Type::IFD:
case Type::UnsignedLong:
return read_every_values.template operator()<u32>();
case Type::UnsignedRational:
return read_every_values.template operator()<Rational<u32>>();
case Type::SignedLong:
return read_every_values.template operator()<i32>();
case Type::SignedRational:
return read_every_values.template operator()<Rational<i32>>();
case Type::Float:
return read_every_values.template operator()<float>();
case Type::Double:
return read_every_values.template operator()<double>();
default:
VERIFY_NOT_REACHED();
}
}
ErrorOr<void> read_tag()
{
auto const tag = TRY(read_value<u16>());
auto const raw_type = TRY(read_value<u16>());
auto const type = TRY(tiff_type_from_u16(raw_type));
auto const count = TRY(read_value<u32>());
Checked<u32> checked_size = size_of_type(type);
checked_size *= count;
if (checked_size.has_overflow())
return Error::from_string_literal("TIFFImageDecoderPlugin: Invalid tag with too large data");
auto tiff_value = TRY(([=, this]() -> ErrorOr<Vector<Value>> {
if (checked_size.value() <= 4) {
auto value = TRY(read_tiff_value(type, count, TRY(m_stream->tell())));
TRY(m_stream->discard(4));
return value;
}
auto const offset = TRY(read_value<u32>());
return read_tiff_value(type, count, offset);
}()));
auto subifd_handler = [&](u32 ifd_offset) -> ErrorOr<void> {
if (auto result = set_next_ifd(ifd_offset); result.is_error()) {
dbgln("{}", result.error());
return {};
}
TRY(read_next_image_file_directory());
return {};
};
TRY(handle_tag(move(subifd_handler), m_metadata, tag, type, count, move(tiff_value)));
return {};
}
NonnullOwnPtr<FixedMemoryStream> m_stream;
State m_state {};
RefPtr<Bitmap> m_bitmap {};
RefPtr<CMYKBitmap> m_cmyk_bitmap {};
ByteOrder m_byte_order {};
Optional<u32> m_next_ifd {};
ExifMetadata m_metadata {};
// These are caches for m_metadata values
PhotometricInterpretation m_photometric_interpretation {};
Vector<u32, 4> m_bits_per_sample {};
u32 m_image_width {};
Predictor m_predictor {};
Optional<u8> m_alpha_channel_index {};
};
}
TIFFImageDecoderPlugin::TIFFImageDecoderPlugin(NonnullOwnPtr<FixedMemoryStream> stream)
{
m_context = make<TIFF::TIFFLoadingContext>(move(stream));
}
TIFFImageDecoderPlugin::~TIFFImageDecoderPlugin() = default;
bool TIFFImageDecoderPlugin::sniff(ReadonlyBytes bytes)
{
if (bytes.size() < 4)
return false;
bool const valid_little_endian = bytes[0] == 0x49 && bytes[1] == 0x49 && bytes[2] == 0x2A && bytes[3] == 0x00;
bool const valid_big_endian = bytes[0] == 0x4D && bytes[1] == 0x4D && bytes[2] == 0x00 && bytes[3] == 0x2A;
return valid_little_endian || valid_big_endian;
}
IntSize TIFFImageDecoderPlugin::size()
{
return m_context->size();
}
ErrorOr<NonnullOwnPtr<ImageDecoderPlugin>> TIFFImageDecoderPlugin::create(ReadonlyBytes data)
{
auto stream = TRY(try_make<FixedMemoryStream>(data));
auto plugin = TRY(adopt_nonnull_own_or_enomem(new (nothrow) TIFFImageDecoderPlugin(move(stream))));
TRY(plugin->m_context->decode_image_header());
return plugin;
}
ErrorOr<ImageFrameDescriptor> TIFFImageDecoderPlugin::frame(size_t index, Optional<IntSize>)
{
if (index > 0)
return Error::from_string_literal("TIFFImageDecoderPlugin: Invalid frame index");
if (m_context->state() == TIFF::TIFFLoadingContext::State::Error)
return Error::from_string_literal("TIFFImageDecoderPlugin: Decoding failed");
if (m_context->state() < TIFF::TIFFLoadingContext::State::FrameDecoded)
TRY(m_context->decode_frame());
if (m_context->cmyk_bitmap())
return ImageFrameDescriptor { TRY(m_context->cmyk_bitmap()->to_low_quality_rgb()), 0 };
return ImageFrameDescriptor { m_context->bitmap(), 0 };
}
Optional<Metadata const&> TIFFImageDecoderPlugin::metadata()
{
return m_context->metadata();
}
ErrorOr<Optional<ReadonlyBytes>> TIFFImageDecoderPlugin::icc_data()
{
return m_context->metadata().icc_profile().map([](auto const& buffer) -> ReadonlyBytes { return buffer.bytes(); });
}
ErrorOr<NonnullOwnPtr<ExifMetadata>> TIFFImageDecoderPlugin::read_exif_metadata(ReadonlyBytes data)
{
auto stream = TRY(try_make<FixedMemoryStream>(data));
auto plugin = TRY(adopt_nonnull_own_or_enomem(new (nothrow) TIFFImageDecoderPlugin(move(stream))));
TRY(plugin->m_context->decode_image_header());
return try_make<ExifMetadata>(plugin->m_context->metadata());
}
}