xref: /dports/audio/libopenmpt/libopenmpt-0.6.0+release.autotools/src/mpt/io_read/filereader.hpp

/* SPDX-License-Identifier: BSL-1.0 OR BSD-3-Clause */

#ifndef MPT_IO_READ_FILEREADER_HPP
#define MPT_IO_READ_FILEREADER_HPP



#include "mpt/base/alloc.hpp"
#include "mpt/base/bit.hpp"
#include "mpt/base/integer.hpp"
#include "mpt/base/memory.hpp"
#include "mpt/base/namespace.hpp"
#include "mpt/base/span.hpp"
#include "mpt/base/utility.hpp"
#include "mpt/io/base.hpp"
#include "mpt/endian/floatingpoint.hpp"
#include "mpt/endian/integer.hpp"
#include "mpt/string/utility.hpp"

#include <algorithm>
#include <array>
#include <limits>
#include <string>
#include <vector>

#include <cassert>
#include <cstddef>
#include <cstring>



namespace mpt {
inline namespace MPT_INLINE_NS {



namespace IO {



namespace FileReader {

// Read a "T" object from the stream.
// If not enough bytes can be read, false is returned.
// If successful, the file cursor is advanced by the size of "T".
template <typename T, typename TFileCursor>
bool Read(TFileCursor & f, T & target) {
	// cppcheck false-positive
	// cppcheck-suppress uninitvar
	mpt::byte_span dst = mpt::as_raw_memory(target);
	if (dst.size() != f.GetRaw(dst).size()) {
		return false;
	}
	f.Skip(dst.size());
	return true;
}

// Read an array of binary-safe T values.
// If successful, the file cursor is advanced by the size of the array.
// Otherwise, the target is zeroed.
template <typename T, std::size_t destSize, typename TFileCursor>
bool ReadArray(TFileCursor & f, T (&destArray)[destSize]) {
	static_assert(mpt::is_binary_safe<T>::value);
	if (!f.CanRead(sizeof(destArray))) {
		mpt::reset(destArray);
		return false;
	}
	f.ReadRaw(mpt::as_raw_memory(destArray));
	return true;
}

// Read an array of binary-safe T values.
// If successful, the file cursor is advanced by the size of the array.
// Otherwise, the target is zeroed.
template <typename T, std::size_t destSize, typename TFileCursor>
bool ReadArray(TFileCursor & f, std::array<T, destSize> & destArray) {
	static_assert(mpt::is_binary_safe<T>::value);
	if (!f.CanRead(sizeof(destArray))) {
		destArray.fill(T{});
		return false;
	}
	f.ReadRaw(mpt::as_raw_memory(destArray));
	return true;
}

// Read destSize elements of binary-safe type T into a vector.
// If successful, the file cursor is advanced by the size of the vector.
// Otherwise, the vector is resized to destSize, but possibly existing contents are not cleared.
template <typename T, typename TFileCursor>
bool ReadVector(TFileCursor & f, std::vector<T> & destVector, size_t destSize) {
	static_assert(mpt::is_binary_safe<T>::value);
	destVector.resize(destSize);
	if (!f.CanRead(sizeof(T) * destSize)) {
		return false;
	}
	f.ReadRaw(mpt::as_raw_memory(destVector));
	return true;
}

template <typename T, std::size_t destSize, typename TFileCursor>
std::array<T, destSize> ReadArray(TFileCursor & f) {
	std::array<T, destSize> destArray;
	ReadArray(f, destArray);
	return destArray;
}

// Read some kind of integer in little-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename T, typename TFileCursor>
T ReadIntLE(TFileCursor & f) {
	static_assert(std::numeric_limits<T>::is_integer == true, "Target type is a not an integer");
	typename mpt::make_le<T>::type target;
	if (!FileReader::Read(f, target)) {
		return 0;
	}
	return target;
}

// Read some kind of integer in big-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename T, typename TFileCursor>
T ReadIntBE(TFileCursor & f) {
	static_assert(std::numeric_limits<T>::is_integer == true, "Target type is a not an integer");
	typename mpt::make_be<T>::type target;
	if (!FileReader::Read(f, target)) {
		return 0;
	}
	return target;
}

// Read a integer in little-endian format which has some of its higher bytes not stored in file.
// If successful, the file cursor is advanced by the given size.
template <typename T, typename TFileCursor>
T ReadTruncatedIntLE(TFileCursor & f, typename TFileCursor::pos_type size) {
	static_assert(std::numeric_limits<T>::is_integer == true, "Target type is a not an integer");
	assert(sizeof(T) >= size);
	if (size == 0) {
		return 0;
	}
	if (!f.CanRead(size)) {
		return 0;
	}
	uint8 buf[sizeof(T)];
	bool negative = false;
	for (std::size_t i = 0; i < sizeof(T); ++i) {
		uint8 byte = 0;
		if (i < size) {
			FileReader::Read(f, byte);
			negative = std::numeric_limits<T>::is_signed && ((byte & 0x80) != 0x00);
		} else {
			// sign or zero extend
			byte = negative ? 0xff : 0x00;
		}
		buf[i] = byte;
	}
	return mpt::bit_cast<typename mpt::make_le<T>::type>(buf);
}

// Read a supplied-size little endian integer to a fixed size variable.
// The data is properly sign-extended when fewer bytes are stored.
// If more bytes are stored, higher order bytes are silently ignored.
// If successful, the file cursor is advanced by the given size.
template <typename T, typename TFileCursor>
T ReadSizedIntLE(TFileCursor & f, typename TFileCursor::pos_type size) {
	static_assert(std::numeric_limits<T>::is_integer == true, "Target type is a not an integer");
	if (size == 0) {
		return 0;
	}
	if (!f.CanRead(size)) {
		return 0;
	}
	if (size < sizeof(T)) {
		return FileReader::ReadTruncatedIntLE<T>(f, size);
	}
	T retval = FileReader::ReadIntLE<T>(f);
	f.Skip(size - sizeof(T));
	return retval;
}

// Read unsigned 32-Bit integer in little-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
uint32 ReadUint32LE(TFileCursor & f) {
	return FileReader::ReadIntLE<uint32>(f);
}

// Read unsigned 32-Bit integer in big-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
uint32 ReadUint32BE(TFileCursor & f) {
	return FileReader::ReadIntBE<uint32>(f);
}

// Read signed 32-Bit integer in little-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
int32 ReadInt32LE(TFileCursor & f) {
	return FileReader::ReadIntLE<int32>(f);
}

// Read signed 32-Bit integer in big-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
int32 ReadInt32BE(TFileCursor & f) {
	return FileReader::ReadIntBE<int32>(f);
}

// Read unsigned 24-Bit integer in little-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
uint32 ReadUint24LE(TFileCursor & f) {
	const auto arr = FileReader::ReadArray<uint8, 3>(f);
	return arr[0] | (arr[1] << 8) | (arr[2] << 16);
}

// Read unsigned 24-Bit integer in big-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
uint32 ReadUint24BE(TFileCursor & f) {
	const auto arr = FileReader::ReadArray<uint8, 3>(f);
	return (arr[0] << 16) | (arr[1] << 8) | arr[2];
}

// Read unsigned 16-Bit integer in little-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
uint16 ReadUint16LE(TFileCursor & f) {
	return FileReader::ReadIntLE<uint16>(f);
}

// Read unsigned 16-Bit integer in big-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
uint16 ReadUint16BE(TFileCursor & f) {
	return FileReader::ReadIntBE<uint16>(f);
}

// Read signed 16-Bit integer in little-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
int16 ReadInt16LE(TFileCursor & f) {
	return FileReader::ReadIntLE<int16>(f);
}

// Read signed 16-Bit integer in big-endian format.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
int16 ReadInt16BE(TFileCursor & f) {
	return FileReader::ReadIntBE<int16>(f);
}

// Read a single 8bit character.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
char ReadChar(TFileCursor & f) {
	char target;
	if (!FileReader::Read(f, target)) {
		return 0;
	}
	return target;
}

// Read unsigned 8-Bit integer.
// If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
uint8 ReadUint8(TFileCursor & f) {
	uint8 target;
	if (!FileReader::Read(f, target)) {
		return 0;
	}
	return target;
}

// Read signed 8-Bit integer. If successful, the file cursor is advanced by the size of the integer.
template <typename TFileCursor>
int8 ReadInt8(TFileCursor & f) {
	int8 target;
	if (!FileReader::Read(f, target)) {
		return 0;
	}
	return target;
}

// Read 32-Bit float in little-endian format.
// If successful, the file cursor is advanced by the size of the float.
template <typename TFileCursor>
float ReadFloatLE(TFileCursor & f) {
	IEEE754binary32LE target;
	if (!FileReader::Read(f, target)) {
		return 0.0f;
	}
	return target;
}

// Read 32-Bit float in big-endian format.
// If successful, the file cursor is advanced by the size of the float.
template <typename TFileCursor>
float ReadFloatBE(TFileCursor & f) {
	IEEE754binary32BE target;
	if (!FileReader::Read(f, target)) {
		return 0.0f;
	}
	return target;
}

// Read 64-Bit float in little-endian format.
// If successful, the file cursor is advanced by the size of the float.
template <typename TFileCursor>
double ReadDoubleLE(TFileCursor & f) {
	IEEE754binary64LE target;
	if (!FileReader::Read(f, target)) {
		return 0.0;
	}
	return target;
}

// Read 64-Bit float in big-endian format.
// If successful, the file cursor is advanced by the size of the float.
template <typename TFileCursor>
double ReadDoubleBE(TFileCursor & f) {
	IEEE754binary64BE target;
	if (!FileReader::Read(f, target)) {
		return 0.0;
	}
	return target;
}

// Read a struct.
// If successful, the file cursor is advanced by the size of the struct. Otherwise, the target is zeroed.
template <typename T, typename TFileCursor>
bool ReadStruct(TFileCursor & f, T & target) {
	static_assert(mpt::is_binary_safe<T>::value);
	if (!FileReader::Read(f, target)) {
		mpt::reset(target);
		return false;
	}
	return true;
}

// Allow to read a struct partially (if there's less memory available than the struct's size, fill it up with zeros).
// The file cursor is advanced by "partialSize" bytes.
template <typename T, typename TFileCursor>
typename TFileCursor::pos_type ReadStructPartial(TFileCursor & f, T & target, typename TFileCursor::pos_type partialSize = sizeof(T)) {
	static_assert(mpt::is_binary_safe<T>::value);
	typename TFileCursor::pos_type copyBytes = std::min(partialSize, sizeof(T));
	if (!f.CanRead(copyBytes)) {
		copyBytes = f.BytesLeft();
	}
	f.GetRaw(mpt::span(mpt::as_raw_memory(target).data(), copyBytes));
	std::memset(mpt::as_raw_memory(target).data() + copyBytes, 0, sizeof(target) - copyBytes);
	f.Skip(partialSize);
	return copyBytes;
}

// Read a null-terminated string into a std::string
template <typename TFileCursor>
bool ReadNullString(TFileCursor & f, std::string & dest, const typename TFileCursor::pos_type maxLength = std::numeric_limits<typename TFileCursor::pos_type>::max()) {
	dest.clear();
	if (!f.CanRead(1)) {
		return false;
	}
	char buffer[mpt::IO::BUFFERSIZE_MINUSCULE];
	typename TFileCursor::pos_type avail = 0;
	while ((avail = std::min(f.GetRaw(mpt::as_span(buffer)).size(), maxLength - dest.length())) != 0) {
		auto end = std::find(buffer, buffer + avail, '\0');
		dest.insert(dest.end(), buffer, end);
		f.Skip(end - buffer);
		if (end < buffer + avail) {
			// Found null char
			f.Skip(1);
			break;
		}
	}
	return dest.length() != 0;
}

// Read a string up to the next line terminator into a std::string
template <typename TFileCursor>
bool ReadLine(TFileCursor & f, std::string & dest, const typename TFileCursor::pos_type maxLength = std::numeric_limits<typename TFileCursor::pos_type>::max()) {
	dest.clear();
	if (!f.CanRead(1)) {
		return false;
	}
	char buffer[mpt::IO::BUFFERSIZE_MINUSCULE];
	char c = '\0';
	typename TFileCursor::pos_type avail = 0;
	while ((avail = std::min(f.GetRaw(mpt::as_span(buffer)).size(), maxLength - dest.length())) != 0) {
		auto end = std::find_if(buffer, buffer + avail, mpt::is_any_line_ending<char>);
		dest.insert(dest.end(), buffer, end);
		f.Skip(end - buffer);
		if (end < buffer + avail) {
			// Found line ending
			f.Skip(1);
			// Handle CRLF line ending
			if (*end == '\r') {
				if (FileReader::Read(f, c) && c != '\n') {
					f.SkipBack(1);
				}
			}
			break;
		}
	}
	return true;
}

// Compare a magic string with the current stream position.
// Returns true if they are identical and advances the file cursor by the the length of the "magic" string.
// Returns false if the string could not be found. The file cursor is not advanced in this case.
template <size_t N, typename TFileCursor>
bool ReadMagic(TFileCursor & f, const char (&magic)[N]) {
	assert(magic[N - 1] == '\0');
	for (std::size_t i = 0; i < N - 1; ++i) {
		assert(magic[i] != '\0');
	}
	constexpr typename TFileCursor::pos_type magicLength = N - 1;
	std::byte buffer[magicLength] = {};
	if (f.GetRaw(mpt::span(buffer, magicLength)).size() != magicLength) {
		return false;
	}
	if (std::memcmp(buffer, magic, magicLength)) {
		return false;
	}
	f.Skip(magicLength);
	return true;
}

// Read variable-length unsigned integer (as found in MIDI files).
// If successful, the file cursor is advanced by the size of the integer and true is returned.
// False is returned if not enough bytes were left to finish reading of the integer or if an overflow happened (source doesn't fit into target integer).
// In case of an overflow, the target is also set to the maximum value supported by its data type.
template <typename T, typename TFileCursor>
bool ReadVarInt(TFileCursor & f, T & target) {
	static_assert(std::numeric_limits<T>::is_integer == true && std::numeric_limits<T>::is_signed == false, "Target type is not an unsigned integer");

	if (f.NoBytesLeft()) {
		target = 0;
		return false;
	}

	std::byte bytes[16]; // More than enough for any valid VarInt
	typename TFileCursor::pos_type avail = f.GetRaw(mpt::as_span(bytes)).size();
	typename TFileCursor::pos_type readPos = 1;

	uint8 b = mpt::byte_cast<uint8>(bytes[0]);
	target = (b & 0x7F);
	std::size_t writtenBits = static_cast<std::size_t>(mpt::bit_width(target)); // Bits used in the most significant byte

	while (readPos < avail && (b & 0x80) != 0) {
		b = mpt::byte_cast<uint8>(bytes[readPos++]);
		target <<= 7;
		target |= (b & 0x7F);
		writtenBits += 7;
		if (readPos == avail) {
			f.Skip(readPos);
			avail = f.GetRaw(mpt::as_span(bytes)).size();
			readPos = 0;
		}
	}
	f.Skip(readPos);

	if (writtenBits > sizeof(target) * 8u) {
		// Overflow
		target = std::numeric_limits<T>::max();
		return false;
	} else if ((b & 0x80) != 0) {
		// Reached EOF
		return false;
	}
	return true;
}

template <typename Tid, typename Tsize>
struct ChunkHeader {
	using id_type = Tid;
	using size_type = Tsize;
	friend constexpr bool declare_binary_safe(const ChunkHeader &) noexcept {
		return true;
	}
	id_type id{};
	size_type size{};
	id_type GetID() const {
		return id;
	}
	size_type GetLength() const {
		return size;
	}
};

template <typename TChunkHeader, typename TFileCursor>
struct Chunk {
	TChunkHeader header;
	TFileCursor data;
	TChunkHeader GetHeader() const {
		return header;
	}
	TFileCursor GetData() const {
		return data;
	}
};

template <typename TChunkHeader, typename TFileCursor>
struct ChunkList {

	using id_type = decltype(TChunkHeader().GetID());
	using size_type = decltype(TChunkHeader().GetLength());

	std::vector<Chunk<TChunkHeader, TFileCursor>> chunks;

	// Check if the list contains a given chunk.
	bool ChunkExists(id_type id) const {
		return std::find_if(chunks.begin(), chunks.end(), [id](const Chunk<TChunkHeader, TFileCursor> & chunk) { return chunk.GetHeader().GetID() == id; }) != chunks.end();
	}

	// Retrieve the first chunk with a given ID.
	TFileCursor GetChunk(id_type id) const {
		auto chunk = std::find_if(chunks.begin(), chunks.end(), [id](const Chunk<TChunkHeader, TFileCursor> & chunk) { return chunk.GetHeader().GetID() == id; });
		if (chunk == chunks.end()) {
			return TFileCursor();
		}
		return chunk->GetData();
	}

	// Retrieve all chunks with a given ID.
	std::vector<TFileCursor> GetAllChunks(id_type id) const {
		std::vector<TFileCursor> result;
		for (const auto & chunk : chunks) {
			if (chunk.GetHeader().GetID() == id) {
				result.push_back(chunk.GetData());
			}
		}
		return result;
	}
};

// Read a single "TChunkHeader" chunk.
// T is required to have the methods GetID() and GetLength().
// GetLength() must return the chunk size in bytes, and GetID() the chunk ID.
template <typename TChunkHeader, typename TFileCursor>
Chunk<TChunkHeader, TFileCursor> ReadNextChunk(TFileCursor & f, typename TFileCursor::pos_type alignment) {
	Chunk<TChunkHeader, TFileCursor> result;
	if (!FileReader::Read(f, result.header)) {
		return Chunk<TChunkHeader, TFileCursor>();
	}
	typename TFileCursor::pos_type dataSize = result.header.GetLength();
	result.data = f.ReadChunk(dataSize);
	if (alignment > 1) {
		if ((dataSize % alignment) != 0) {
			f.Skip(alignment - (dataSize % alignment));
		}
	}
	return result;
}

// Read a series of "TChunkHeader" chunks until the end of file is reached.
// T is required to have the methods GetID() and GetLength().
// GetLength() must return the chunk size in bytes, and GetID() the chunk ID.
template <typename TChunkHeader, typename TFileCursor>
ChunkList<TChunkHeader, TFileCursor> ReadChunks(TFileCursor & f, typename TFileCursor::pos_type alignment) {
	ChunkList<TChunkHeader, TFileCursor> result;
	while (f.CanRead(sizeof(TChunkHeader))) {
		result.chunks.push_back(FileReader::ReadNextChunk<TChunkHeader, TFileCursor>(f, alignment));
	}
	return result;
}

// Read a series of "TChunkHeader" chunks until a given chunk ID is found.
// T is required to have the methods GetID() and GetLength().
// GetLength() must return the chunk size in bytes, and GetID() the chunk ID.
template <typename TChunkHeader, typename TFileCursor>
ChunkList<TChunkHeader, TFileCursor> ReadChunksUntil(TFileCursor & f, typename TFileCursor::pos_type alignment, decltype(TChunkHeader().GetID()) lastID) {
	ChunkList<TChunkHeader, TFileCursor> result;
	while (f.CanRead(sizeof(TChunkHeader))) {
		result.chunks.push_back(FileReader::ReadNextChunk<TChunkHeader, TFileCursor>(f, alignment));
		if (result.chunks.back().GetHeader().GetID() == lastID) {
			break;
		}
	}
	return result;
}

} // namespace FileReader



} // namespace IO



} // namespace MPT_INLINE_NS
} // namespace mpt



#endif // MPT_IO_READ_FILEREADER_HPP