diff --git a/TODO.md b/TODO.md index 7196b66..85d8a2c 100644 --- a/TODO.md +++ b/TODO.md @@ -8,7 +8,11 @@ - Add/retrive all types of data - Keeps no memory except for caching - [x] Refactor bytecode code -- [ ] Output bytecode format + +Improvements: +- [ ] Fixed int type (based on opcode) +- [ ] Constant type (only floats and strings for now) +- [ ] Debug output bytecode format After some additional development: - [ ] Bytecode debugging info diff --git a/doc/OPCODES b/doc/OPCODES index fd38433..d9b57f1 100644 --- a/doc/OPCODES +++ b/doc/OPCODES @@ -1,17 +1,23 @@ Operations ---------- +Operations take either 0 or 1 parameter. The ones that take a parameter, it can be either a int8, int16 or int32. + +The binary of the opcode is: XXYY.YYYY, where XX defines the parameter type, and YY.YYYY is the instruction. For the XX values: + +00 - no parameter +01 - int8 +10 - int16 +11 - int32 + Stack operations: (0x00~0x1f) pushn [int] Push int - pushr [float] Push float (real) - pshcn [index] Push int from constant list - pshcr [index] Push float from constant list - pshcs [index] Push string from constant list + pushc [index] Push constant pushf [function] Push function id pushz Push zero (or false) pusht Push true - newa [array] Push (create) empty array - newt [table] Push (create) empty table + newa Push (create) empty array + newt Push (create) empty table pop dup diff --git a/src/bytecode/bytearray.cc b/src/bytecode/bytearray.cc index fcd4858..4bfcfdd 100644 --- a/src/bytecode/bytearray.cc +++ b/src/bytecode/bytearray.cc @@ -14,14 +14,23 @@ void ByteArray::set_byte(uint32_t addr, uint8_t byte) } } -void ByteArray::set_int(uint32_t addr, int32_t value) +void ByteArray::set_int8(uint32_t addr, int8_t value) { - uint32_t zz = ((uint32_t)(value << 1)) ^ ((uint32_t)(value >> 31)); - while (zz > 0x7F) { - set_byte(addr++, (zz & 0x7F) | 0x80); - zz >>= 7; - } - set_byte(addr, zz & 0x7F); + set_byte(addr, (uint8_t) value); +} + +void ByteArray::set_int16(uint32_t addr, int16_t value) +{ + set_byte(addr, (uint8_t) (value)); + set_byte(addr+1, (uint8_t) (value >> 8)); +} + +void ByteArray::set_int32(uint32_t addr, int32_t value) +{ + set_byte(addr, (uint8_t) (value)); + set_byte(addr+1, (uint8_t) (value >> 8)); + set_byte(addr+2, (uint8_t) (value >> 16)); + set_byte(addr+3, (uint8_t) (value >> 24)); } void ByteArray::set_uint16(uint32_t addr, uint16_t value) @@ -80,20 +89,23 @@ uint32_t ByteArray::get_uint32(uint32_t addr) const | (uint32_t) get_byte(addr+3) << 24; } -std::pair ByteArray::get_int(uint32_t addr) const +int8_t ByteArray::get_int8(uint32_t addr) const { - uint32_t zz = 0; - int shift = 0; - for (size_t i = 0; shift < 35; i++) { - uint8_t byte = get_byte(addr++); - zz |= (uint32_t)(byte & 0x7F) << shift; - if (!(byte & 0x80)) { - int32_t value = (int32_t)((zz >> 1) ^ -(zz & 1)); - return std::make_pair(value, (int)(i + 1)); - } - shift += 7; - } - throw BytecodeParsingError("Error parsing int32 at position " + std::to_string(addr)); + return std::bit_cast(get_byte(addr)); +} + +int16_t ByteArray::get_int16(uint32_t addr) const +{ + return (uint16_t) get_byte(addr) + | (uint16_t) get_byte(addr+1) << 8; +} + +int32_t ByteArray::get_int32(uint32_t addr) const +{ + return std::bit_cast((uint32_t) get_byte(addr) + | (uint32_t) get_byte(addr+1) << 8 + | (uint32_t) get_byte(addr+2) << 16 + | (uint32_t) get_byte(addr+3) << 24); } float ByteArray::get_float(uint32_t addr) const diff --git a/src/bytecode/bytearray.hh b/src/bytecode/bytearray.hh index 99984ce..fa45591 100644 --- a/src/bytecode/bytearray.hh +++ b/src/bytecode/bytearray.hh @@ -16,23 +16,29 @@ public: void set_byte(uint32_t addr, uint8_t byte); void set_uint16(uint32_t addr, uint16_t value); void set_uint32(uint32_t addr, uint32_t value); - void set_int(uint32_t addr, int32_t value); + void set_int8(uint32_t addr, int8_t value); + void set_int16(uint32_t addr, int16_t value); + void set_int32(uint32_t addr, int32_t value); void set_float(uint32_t addr, float value); void set_string(uint32_t addr, std::string const& str); void set_bytearray(uint32_t addr, ByteArray const& bytearray); - void append_byte(uint8_t byte) { set_byte(data_.size(), byte); } - void append_uint16(uint16_t value) { set_uint16(data_.size(), value); } - void append_uint32(uint32_t value) { set_uint32(data_.size(), value); } - void append_int(int32_t value) { set_int(data_.size(), value); } - void append_float(float value) { set_float(data_.size(), value); } - void append_string(std::string const& str) { set_string(data_.size(), str); } + void append_byte(uint8_t byte) { set_byte(data_.size(), byte); } + void append_uint16(uint16_t value) { set_uint16(data_.size(), value); } + void append_uint32(uint32_t value) { set_uint32(data_.size(), value); } + void append_int8(int8_t value) { set_int8(data_.size(), value); } + void append_int16(int16_t value) { set_int16(data_.size(), value); } + void append_int32(int32_t value) { set_int32(data_.size(), value); } + void append_float(float value) { set_float(data_.size(), value); } + void append_string(std::string const& str) { set_string(data_.size(), str); } void append_bytearray(ByteArray const& bytearray); [[nodiscard]] uint8_t get_byte(uint32_t addr) const; [[nodiscard]] uint16_t get_uint16(uint32_t addr) const; [[nodiscard]] uint32_t get_uint32(uint32_t addr) const; - [[nodiscard]] std::pair get_int(uint32_t addr) const; + [[nodiscard]] int8_t get_int8(uint32_t addr) const; + [[nodiscard]] int16_t get_int16(uint32_t addr) const; + [[nodiscard]] int32_t get_int32(uint32_t addr) const; [[nodiscard]] float get_float(uint32_t addr) const; [[nodiscard]] std::pair get_string(uint32_t addr) const; diff --git a/src/bytecode/bytecode.cc b/src/bytecode/bytecode.cc index b5a3c5c..5dbe173 100644 --- a/src/bytecode/bytecode.cc +++ b/src/bytecode/bytecode.cc @@ -39,12 +39,6 @@ uint32_t Bytecode::n_functions() const return cache_.n_functions; } -int32_t Bytecode::get_constant_int(uint32_t idx) const -{ - uint32_t constant_idx = byte_array_.get_uint32(cache_.constants_idx_addr + (idx * CONST_RECORD_SZ)); - return byte_array_.get_int(cache_.constants_start_addr + constant_idx).first; -} - float Bytecode::get_constant_float(uint32_t idx) const { uint32_t constant_idx = byte_array_.get_uint32(cache_.constants_idx_addr + (idx * CONST_RECORD_SZ)); @@ -76,14 +70,19 @@ uint8_t Bytecode::get_code_byte(uint32_t function_id, uint32_t idx) const return byte_array_.get_byte(cache_.function_addr.at(function_id) + idx); } -std::pair Bytecode::get_code_int(uint32_t function_id, uint32_t idx) const +int8_t Bytecode::get_code_int8(uint32_t function_id, uint32_t idx) const { - return byte_array_.get_int(cache_.function_addr.at(function_id) + idx); + return byte_array_.get_int8(cache_.function_addr.at(function_id) + idx); } -float Bytecode::get_code_float(uint32_t function_id, uint32_t idx) const +int16_t Bytecode::get_code_int16(uint32_t function_id, uint32_t idx) const { - return byte_array_.get_float(cache_.function_addr.at(function_id) + idx); + return byte_array_.get_int16(cache_.function_addr.at(function_id) + idx); +} + +int32_t Bytecode::get_code_int32(uint32_t function_id, uint32_t idx) const +{ + return byte_array_.get_int32(cache_.function_addr.at(function_id) + idx); } ByteArray Bytecode::generate(BytecodePrototype const& bp) @@ -101,7 +100,7 @@ ByteArray Bytecode::generate(BytecodePrototype const& bp) for (auto const& constant: bp.constants) { constant_indexes.append_uint32(idx); std::visit(overloaded { - [&](int32_t i) { raw_constants.append_int(i); }, + [&](int32_t i) { raw_constants.append_int32(i); }, [&](float f) { raw_constants.append_float(f); }, [&](std::string const& s) { raw_constants.append_string(s); }, }, constant); diff --git a/src/bytecode/bytecode.hh b/src/bytecode/bytecode.hh index fa3dc98..6fb489f 100644 --- a/src/bytecode/bytecode.hh +++ b/src/bytecode/bytecode.hh @@ -13,7 +13,6 @@ public: [[nodiscard]] uint32_t n_constants() const; [[nodiscard]] uint32_t n_functions() const; - [[nodiscard]] int32_t get_constant_int(uint32_t idx) const; [[nodiscard]] float get_constant_float(uint32_t idx) const; [[nodiscard]] std::string get_constant_string(uint32_t idx) const; @@ -21,9 +20,10 @@ public: [[nodiscard]] FunctionDef get_function_def(uint32_t function_id) const; [[nodiscard]] uint32_t get_function_sz(uint32_t function_id) const; - [[nodiscard]] uint8_t get_code_byte(uint32_t function_id, uint32_t idx) const; - [[nodiscard]] std::pair get_code_int(uint32_t function_id, uint32_t idx) const; - [[nodiscard]] float get_code_float(uint32_t function_id, uint32_t idx) const; + [[nodiscard]] uint8_t get_code_byte(uint32_t function_id, uint32_t idx) const; + [[nodiscard]] int8_t get_code_int8(uint32_t function_id, uint32_t idx) const; + [[nodiscard]] int16_t get_code_int16(uint32_t function_id, uint32_t idx) const; + [[nodiscard]] int32_t get_code_int32(uint32_t function_id, uint32_t idx) const; // TODO - debugging info diff --git a/src/bytecode/tests.cc b/src/bytecode/tests.cc index a696065..5838123 100644 --- a/src/bytecode/tests.cc +++ b/src/bytecode/tests.cc @@ -26,11 +26,11 @@ TEST(ByteArray, ByteArray) ByteArray ba; ba.set_byte(1, 0xab); ASSERT_EQ(ba.get_byte(1), 0xab); - ba.set_int(1, 12); ASSERT_EQ(ba.get_int(1), std::make_pair(12, 1)); - ba.set_int(1, -12); ASSERT_EQ(ba.get_int(1), std::make_pair(-12, 1)); - ba.set_int(1, 5000); ASSERT_EQ(ba.get_int(1), std::make_pair(5000, 2)); - ba.set_int(1, 5000300); ASSERT_EQ(ba.get_int(1), std::make_pair(5000300, 4)); - ba.set_int(1, -5000300); ASSERT_EQ(ba.get_int(1), std::make_pair(-5000300, 4)); + ba.set_int8(1, 12); ASSERT_EQ(ba.get_int8(1), 12); + ba.set_int8(1, -12); ASSERT_EQ(ba.get_int8(1), -12); + ba.set_int16(1, 5000); ASSERT_EQ(ba.get_int16(1), 5000); + ba.set_int32(1, 5000300); ASSERT_EQ(ba.get_int32(1), 5000300); + ba.set_int32(1, -5000300); ASSERT_EQ(ba.get_int32(1), -5000300); ba.set_float(1, 3.14); ASSERT_FLOAT_EQ(ba.get_float(1), 3.14); ba.set_float(1, -3.14); ASSERT_FLOAT_EQ(ba.get_float(1), -3.14); @@ -82,7 +82,7 @@ TEST(Bytecode, Code) BytecodePrototype bp; auto& f = bp.functions.emplace_back(0, 0); f.code.append_byte(0x68); - f.code.append_int(42); + f.code.append_int32(42); auto& f2 = bp.functions.emplace_back(2, 1); f2.code.append_byte(0x42); @@ -128,7 +128,7 @@ TEST(Bytecode, Parsing) auto& f = bp.functions.emplace_back(0, 0); f.code.append_byte(0x68); - f.code.append_int(42); + f.code.append_int8(42); auto& ff = bp.functions.emplace_back(2, 1); ff.code.append_byte(0x42); @@ -140,10 +140,9 @@ TEST(Bytecode, Parsing) Bytecode bc(std::move(ba)); - ASSERT_EQ(bc.n_constants(), 3); + ASSERT_EQ(bc.n_constants(), 2); ASSERT_EQ(bc.n_functions(), 2); - ASSERT_EQ(bc.get_constant_int(0), 42); ASSERT_FLOAT_EQ(bc.get_constant_float(1), 3.14f); ASSERT_EQ(bc.get_constant_string(2), "HELLO"); @@ -156,7 +155,7 @@ TEST(Bytecode, Parsing) ASSERT_EQ(f2.locals, 1); ASSERT_EQ(bc.get_code_byte(0, 0), 0x68); - ASSERT_EQ(bc.get_code_int(0, 1), std::make_pair(42, 1)); + ASSERT_EQ(bc.get_code_int8(0, 1), 42); ASSERT_EQ(bc.get_code_byte(1, 0), 0x42); }