diff options
| author | Amlal El Mahrouss <amlal@nekernel.org> | 2025-07-30 08:50:15 +0100 |
|---|---|---|
| committer | Amlal El Mahrouss <amlal@nekernel.org> | 2025-07-30 08:50:15 +0100 |
| commit | 1c8c5cff67b20d86c442b0917d6c1fc6407140df (patch) | |
| tree | 53ebea660bef14cdc2ff5b7ebefb4049f705f997 /dev/CompilerKit/src | |
| parent | 073811d89c98d6e1c078a032ca2eedefebf80384 (diff) | |
feat! Breaking API changes of NeCTI's LibCompiler and LibDebugger.
what:
- They've now become CompilerKit and DebuggerKit.
- Expanding XCoff for NeBoot PowerPC backend.
Signed-off-by: Amlal El Mahrouss <amlal@nekernel.org>
Diffstat (limited to 'dev/CompilerKit/src')
| -rw-r--r-- | dev/CompilerKit/src/Backend/Assembler32x0.cc | 39 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Backend/Assembler64x0.cc | 874 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Backend/AssemblerAMD64.cc | 1206 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Backend/AssemblerARM64.cc | 587 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Backend/AssemblerPowerPC.cc | 911 | ||||
| -rw-r--r-- | dev/CompilerKit/src/BasicString.cc | 207 | ||||
| -rw-r--r-- | dev/CompilerKit/src/CodeGen.cc | 52 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Frontend.cc | 51 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Frontend/CCompiler64x0.cc | 1286 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Frontend/CCompilerARM64.cc | 1284 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Frontend/CCompilerPower64.cc | 1303 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Frontend/CPlusPlusCompilerAMD64.cc | 892 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Linker/DynamicLinker64PEF.cc | 672 | ||||
| -rw-r--r-- | dev/CompilerKit/src/Macro/CPlusPlusCompilerPreProcessor.cc | 893 |
14 files changed, 10257 insertions, 0 deletions
diff --git a/dev/CompilerKit/src/Backend/Assembler32x0.cc b/dev/CompilerKit/src/Backend/Assembler32x0.cc new file mode 100644 index 0000000..e298d2d --- /dev/null +++ b/dev/CompilerKit/src/Backend/Assembler32x0.cc @@ -0,0 +1,39 @@ +/* ------------------------------------------- + + Copyright (C) 2024-2025 Amlal EL Mahrouss, all rights reserved + +------------------------------------------- */ + +/// bugs: 0 + +///////////////////////////////////////////////////////////////////////////////////////// + +// @file 32asm.cxx +// @author EL Mahrouss Amlal +// @brief 32x0 Assembler. + +// REMINDER: when dealing with an undefined symbol use (string +// size):LinkerFindSymbol:(string) so that ld will look for it. + +///////////////////////////////////////////////////////////////////////////////////////// + +#ifndef __ASM_NEED_32x0__ +#define __ASM_NEED_32x0__ 1 +#endif + +#include <CompilerKit/AE.h> +#include <CompilerKit/detail/32x0.h> +#include <CompilerKit/Frontend.h> +#include <CompilerKit/PEF.h> +#include <CompilerKit/utils/CompilerUtils.h> + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief 32x0 Assembler entrypoint, the program/module starts here. + +///////////////////////////////////////////////////////////////////////////////////////// + +LIBCOMPILER_MODULE(NEAssemblerMain32000) { + CompilerKit::install_signal(SIGSEGV, Detail::drvi_crash_handler); + return EXIT_SUCCESS; +} diff --git a/dev/CompilerKit/src/Backend/Assembler64x0.cc b/dev/CompilerKit/src/Backend/Assembler64x0.cc new file mode 100644 index 0000000..32b199d --- /dev/null +++ b/dev/CompilerKit/src/Backend/Assembler64x0.cc @@ -0,0 +1,874 @@ +/* ------------------------------------------- + + Copyright (C) 2024-2025 Amlal EL Mahrouss, all rights reserved + +------------------------------------------- */ + +/// bugs: 0 + +///////////////////////////////////////////////////////////////////////////////////////// + +// @file Assembler64x0.cxx +// @author EL Mahrouss Amlal +// @brief 64x000 Assembler. + +// REMINDER: when dealing with an undefined symbol use (string +// size):LinkerFindSymbol:(string) so that ld will look for it. + +///////////////////////////////////////////////////////////////////////////////////////// + +#ifndef __ASM_NEED_64x0__ +#define __ASM_NEED_64x0__ 1 +#endif + +#include <CompilerKit/AE.h> +#include <CompilerKit/detail/64x0.h> +#include <CompilerKit/Frontend.h> +#include <CompilerKit/PEF.h> +#include <CompilerKit/utils/CompilerUtils.h> +#include <algorithm> +#include <filesystem> +#include <fstream> +#include <iostream> +#include <vector> + +///////////////////// + +// ANSI ESCAPE CODES + +///////////////////// + +static char kOutputArch = CompilerKit::kPefArch64000; + +constexpr auto k64x0IPAlignment = 0x1U; + +static std::size_t kCounter = 1UL; + +static std::uintptr_t kOrigin = kPefBaseOrigin; +static std::vector<std::pair<std::string, std::uintptr_t>> kOriginLabel; + +static std::vector<e64k_num_t> kBytes; + +static CompilerKit::AERecordHeader kCurrentRecord{ + .fName = "", .fKind = CompilerKit::kPefCode, .fSize = 0, .fOffset = 0}; + +static std::vector<CompilerKit::AERecordHeader> kRecords; +static std::vector<std::string> kUndefinedSymbols; + +static const std::string kUndefinedSymbol = ":UndefinedSymbol:"; +static const std::string kRelocSymbol = ":RuntimeSymbol:"; + +// \brief forward decl. +static bool asm_read_attributes(std::string line); + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief 64x0 assembler entrypoint, the program/module starts here. + +///////////////////////////////////////////////////////////////////////////////////////// + +LIBCOMPILER_MODULE(AssemblerMain64x0) { + CompilerKit::install_signal(SIGSEGV, Detail::drvi_crash_handler); + + for (size_t i = 1; i < argc; ++i) { + if (argv[i][0] == '-') { + if (strcmp(argv[i], "-64x0-ver") == 0 || strcmp(argv[i], "-64x0-v") == 0) { + kStdOut + << "Assembler64x0: 64x0 Assembler.\nAssembler64x0: v1.10\nAssembler64x0: Copyright (c) " + "Amlal El Mahrouss\n"; + return 0; + } else if (strcmp(argv[i], "-64x0-h") == 0) { + kStdOut << "Assembler64x0: 64x0 Assembler.\nAssembler64x0: Copyright (c) 2024 Mahrouss " + "Logic.\n"; + kStdOut << "--version: Print program version.\n"; + kStdOut << "--verbose: Print verbose output.\n"; + kStdOut << "--binary: Output as flat binary.\n"; + kStdOut << "--64xxx: Compile for a subset of the X64000.\n"; + + return 0; + } else if (strcmp(argv[i], "-64x0-binary") == 0) { + kOutputAsBinary = true; + continue; + } else if (strcmp(argv[i], "-64x0-verbose") == 0) { + kVerbose = true; + continue; + } + + kStdOut << "Assembler64x0: ignore " << argv[i] << "\n"; + continue; + } + + if (!std::filesystem::exists(argv[i])) { + kStdOut << "Assembler64x0: can't open: " << argv[i] << std::endl; + goto asm_fail_exit; + } + + std::string object_output(argv[i]); + + for (auto& ext : kAsmFileExts) { + if (object_output.find(ext) != std::string::npos) { + object_output.erase(object_output.find(ext), std::strlen(ext)); + } + } + + object_output += kOutputAsBinary ? kBinaryFileExt : kObjectFileExt; + + std::ifstream file_ptr(argv[i]); + std::ofstream file_ptr_out(object_output, std::ofstream::binary); + + if (file_ptr_out.bad()) { + if (kVerbose) { + kStdOut << "Assembler64x0: error: " << strerror(errno) << "\n"; + } + } + + std::string line; + + CompilerKit::AEHeader hdr{0}; + + memset(hdr.fPad, kAENullType, kAEPad); + + hdr.fMagic[0] = kAEMag0; + hdr.fMagic[1] = kAEMag1; + hdr.fSize = sizeof(CompilerKit::AEHeader); + hdr.fArch = kOutputArch; + + ///////////////////////////////////////////////////////////////////////////////////////// + + // COMPILATION LOOP + + ///////////////////////////////////////////////////////////////////////////////////////// + + CompilerKit::Encoder64x0 asm64; + + while (std::getline(file_ptr, line)) { + if (auto ln = asm64.CheckLine(line, argv[i]); !ln.empty()) { + Detail::print_error(ln, argv[i]); + continue; + } + + try { + asm_read_attributes(line); + asm64.WriteLine(line, argv[i]); + } catch (const std::exception& e) { + if (kVerbose) { + std::string what = e.what(); + Detail::print_warning("exit because of: " + what, "CompilerKit"); + } + + std::filesystem::remove(object_output); + goto asm_fail_exit; + } + } + + if (!kOutputAsBinary) { + if (kVerbose) { + kStdOut << "Assembler64x0: Writing object file...\n"; + } + + // this is the final step, write everything to the file. + + auto pos = file_ptr_out.tellp(); + + hdr.fCount = kRecords.size() + kUndefinedSymbols.size(); + + file_ptr_out << hdr; + + if (kRecords.empty()) { + kStdErr << "Assembler64x0: At least one record is needed to write an object " + "file.\nAssembler64x0: Make one using `public_segment .code64 foo_bar`.\n"; + + std::filesystem::remove(object_output); + return 1; + } + + kRecords[kRecords.size() - 1].fSize = kBytes.size(); + + std::size_t record_count = 0UL; + + for (auto& rec : kRecords) { + if (kVerbose) kStdOut << "Assembler64x0: Wrote record " << rec.fName << " to file...\n"; + + rec.fFlags |= CompilerKit::kKindRelocationAtRuntime; + rec.fOffset = record_count; + ++record_count; + + file_ptr_out << rec; + } + + // increment once again, so that we won't lie about the kUndefinedSymbols. + ++record_count; + + for (auto& sym : kUndefinedSymbols) { + CompilerKit::AERecordHeader _record_hdr{0}; + + if (kVerbose) kStdOut << "Assembler64x0: Wrote symbol " << sym << " to file...\n"; + + _record_hdr.fKind = kAENullType; + _record_hdr.fSize = sym.size(); + _record_hdr.fOffset = record_count; + + ++record_count; + + memset(_record_hdr.fPad, kAENullType, kAEPad); + memcpy(_record_hdr.fName, sym.c_str(), sym.size()); + + file_ptr_out << _record_hdr; + + ++kCounter; + } + + auto pos_end = file_ptr_out.tellp(); + + file_ptr_out.seekp(pos); + + hdr.fStartCode = pos_end; + hdr.fCodeSize = kBytes.size(); + + file_ptr_out << hdr; + + file_ptr_out.seekp(pos_end); + } else { + if (kVerbose) { + kStdOut << "Assembler64x0: Write raw binary...\n"; + } + } + + // byte from byte, we write this. + for (auto& byte : kBytes) { + file_ptr_out.write(reinterpret_cast<const char*>(&byte), sizeof(byte)); + } + + if (kVerbose) kStdOut << "Assembler64x0: Wrote file with program in it.\n"; + + file_ptr_out.flush(); + file_ptr_out.close(); + + if (kVerbose) kStdOut << "Assembler64x0: Exit succeeded.\n"; + + return 0; + } + +asm_fail_exit: + + if (kVerbose) kStdOut << "Assembler64x0: Exit failed.\n"; + + return 1; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Check for attributes +// returns true if any was found. + +///////////////////////////////////////////////////////////////////////////////////////// + +static bool asm_read_attributes(std::string line) { + // extern_segment is the opposite of public_segment, it signals to the ld + // that we need this symbol. + if (CompilerKit::find_word(line, "extern_segment")) { + if (kOutputAsBinary) { + Detail::print_error("Invalid extern_segment directive in flat binary mode.", "CompilerKit"); + throw std::runtime_error("invalid_extern_segment_bin"); + } + + auto name = line.substr(line.find("extern_segment") + strlen("extern_segment")); + + /// sanity check to avoid stupid linker errors. + if (name.size() == 0) { + Detail::print_error("Invalid extern_segment", "power-as"); + throw std::runtime_error("invalid_extern_segment"); + } + + std::string result = std::to_string(name.size()); + result += kUndefinedSymbol; + + // mangle this + for (char& j : name) { + if (j == ' ' || j == ',') j = '$'; + } + + result += name; + + if (name.find(".code64") != std::string::npos) { + // data is treated as code. + kCurrentRecord.fKind = CompilerKit::kPefCode; + } else if (name.find(".data64") != std::string::npos) { + // no code will be executed from here. + kCurrentRecord.fKind = CompilerKit::kPefData; + } else if (name.find(".zero64") != std::string::npos) { + // this is a bss section. + kCurrentRecord.fKind = CompilerKit::kPefZero; + } + + // this is a special case for the start stub. + // we want this so that ld can find it. + + if (name == kPefStart) { + kCurrentRecord.fKind = CompilerKit::kPefCode; + } + + // now we can tell the code size of the previous kCurrentRecord. + + if (!kRecords.empty()) kRecords[kRecords.size() - 1].fSize = kBytes.size(); + + memset(kCurrentRecord.fName, 0, kAESymbolLen); + memcpy(kCurrentRecord.fName, result.c_str(), result.size()); + + ++kCounter; + + memset(kCurrentRecord.fPad, kAENullType, kAEPad); + + kRecords.emplace_back(kCurrentRecord); + + return true; + } + // public_segment is a special keyword used by Assembler64x0 to tell the AE output stage to + // mark this section as a header. it currently supports .code64, .data64., + // .zero64 + else if (CompilerKit::find_word(line, "public_segment")) { + if (kOutputAsBinary) { + Detail::print_error("Invalid public_segment directive in flat binary mode.", "CompilerKit"); + throw std::runtime_error("invalid_public_segment_bin"); + } + + auto name = line.substr(line.find("public_segment") + strlen("public_segment")); + + std::string name_copy = name; + + for (char& j : name) { + if (j == ' ') j = '$'; + } + + if (name.find(".code64") != std::string::npos) { + // data is treated as code. + + name_copy.erase(name_copy.find(".code64"), strlen(".code64")); + kCurrentRecord.fKind = CompilerKit::kPefCode; + } else if (name.find(".data64") != std::string::npos) { + // no code will be executed from here. + + name_copy.erase(name_copy.find(".data64"), strlen(".data64")); + kCurrentRecord.fKind = CompilerKit::kPefData; + } else if (name.find(".zero64") != std::string::npos) { + // this is a bss section. + + name_copy.erase(name_copy.find(".zero64"), strlen(".zero64")); + kCurrentRecord.fKind = CompilerKit::kPefZero; + } + + // this is a special case for the start stub. + // we want this so that ld can find it. + + if (name == kPefStart) { + kCurrentRecord.fKind = CompilerKit::kPefCode; + } + + while (name_copy.find(" ") != std::string::npos) name_copy.erase(name_copy.find(" "), 1); + + kOriginLabel.push_back(std::make_pair(name_copy, kOrigin)); + ++kOrigin; + + // now we can tell the code size of the previous kCurrentRecord. + + if (!kRecords.empty()) kRecords[kRecords.size() - 1].fSize = kBytes.size(); + + memset(kCurrentRecord.fName, 0, kAESymbolLen); + memcpy(kCurrentRecord.fName, name.c_str(), name.size()); + + ++kCounter; + + memset(kCurrentRecord.fPad, kAENullType, kAEPad); + + kRecords.emplace_back(kCurrentRecord); + + return true; + } + + return false; +} + +// \brief algorithms and helpers. + +namespace Detail::algorithm { +// \brief authorize a brief set of characters. +static inline bool is_not_alnum_space(char c) { + return !(isalpha(c) || isdigit(c) || (c == ' ') || (c == '\t') || (c == ',') || (c == '(') || + (c == ')') || (c == '"') || (c == '\'') || (c == '[') || (c == ']') || (c == '+') || + (c == '_') || (c == ':') || (c == '@') || (c == '.')); +} + +bool is_valid_64x0(std::string str) { + return std::find_if(str.begin(), str.end(), is_not_alnum_space) == str.end(); +} +} // namespace Detail::algorithm + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Check for line (syntax check) + +///////////////////////////////////////////////////////////////////////////////////////// + +std::string CompilerKit::Encoder64x0::CheckLine(std::string line, std::string file) { + std::string err_str; + + if (line.empty() || CompilerKit::find_word(line, "extern_segment") || + CompilerKit::find_word(line, "public_segment") || line.find('#') != std::string::npos || + CompilerKit::find_word(line, ";")) { + if (line.find('#') != std::string::npos) { + line.erase(line.find('#')); + } else if (line.find(';') != std::string::npos) { + line.erase(line.find(';')); + } else { + // now check the line for validity + if (!Detail::algorithm::is_valid_64x0(line)) { + err_str = "Line contains non alphanumeric characters.\nhere -> "; + err_str += line; + } + } + + return err_str; + } + + if (!Detail::algorithm::is_valid_64x0(line)) { + err_str = "Line contains non alphanumeric characters.\nhere -> "; + err_str += line; + + return err_str; + } + + // check for a valid instruction format. + + if (line.find(',') != std::string::npos) { + if (line.find(',') + 1 == line.size()) { + err_str += "\nInstruction lacks right register, here -> "; + err_str += line.substr(line.find(',')); + + return err_str; + } else { + bool nothing_on_right = true; + + if (line.find(',') + 1 > line.size()) { + err_str += "\nInstruction not complete, here -> "; + err_str += line; + + return err_str; + } + + auto substr = line.substr(line.find(',') + 1); + + for (auto& ch : substr) { + if (ch != ' ' && ch != '\t') { + nothing_on_right = false; + } + } + + // this means we found nothing after that ',' . + if (nothing_on_right) { + err_str += "\nInstruction not complete, here -> "; + err_str += line; + + return err_str; + } + } + } + + // these do take an argument. + std::vector<std::string> operands_inst = {"stw", "ldw", "lda", "sta"}; + + // these don't. + std::vector<std::string> filter_inst = {"jlr", "jrl", "int"}; + + for (auto& opcode64x0 : kOpcodes64x0) { + if (line.find(opcode64x0.fName) != std::string::npos) { + if (opcode64x0.fFunct7 == kAsmNoArgs) return err_str; + + for (auto& op : operands_inst) { + // if only the instruction was found. + if (line == op) { + err_str += "\nMalformed "; + err_str += op; + err_str += " instruction, here -> "; + err_str += line; + } + } + + // if it is like that -> addr1, 0x0 + if (auto it = std::find(filter_inst.begin(), filter_inst.end(), opcode64x0.fName); + it == filter_inst.cend()) { + if (CompilerKit::find_word(line, opcode64x0.fName)) { + if (!isspace(line[line.find(opcode64x0.fName) + strlen(opcode64x0.fName)])) { + err_str += "\nMissing space between "; + err_str += opcode64x0.fName; + err_str += " and operands.\nhere -> "; + err_str += line; + } + } + } + + return err_str; + } + } + + err_str += "Unrecognized instruction: " + line; + + return err_str; +} + +bool CompilerKit::Encoder64x0::WriteNumber(const std::size_t& pos, std::string& jump_label) { + if (!isdigit(jump_label[pos])) return false; + + switch (jump_label[pos + 1]) { + case 'x': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16); !res) { + if (errno != 0) { + Detail::print_error("invalid hex number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_hex_number"); + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16)); + + for (char& i : num.number) { + kBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "Assembler64x0: found a base 16 number here: " << jump_label.substr(pos) << "\n"; + } + + return true; + } + case 'b': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2); !res) { + if (errno != 0) { + Detail::print_error("invalid binary number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_bin"); + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2)); + + if (kVerbose) { + kStdOut << "Assembler64x0: found a base 2 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + kBytes.push_back(i); + } + + return true; + } + case 'o': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7); !res) { + if (errno != 0) { + Detail::print_error("invalid octal number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_octal"); + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7)); + + if (kVerbose) { + kStdOut << "Assembler64x0: found a base 8 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + kBytes.push_back(i); + } + + return true; + } + default: { + break; + } + } + + /* check for errno and stuff like that */ + if (auto res = strtol(jump_label.substr(pos).c_str(), nullptr, 10); !res) { + if (errno != 0) { + return false; + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos).c_str(), nullptr, 10)); + + for (char& i : num.number) { + kBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "Assembler64x0: found a base 10 number here: " << jump_label.substr(pos) << "\n"; + } + + return true; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Read and write an instruction to the output array. + +///////////////////////////////////////////////////////////////////////////////////////// + +bool CompilerKit::Encoder64x0::WriteLine(std::string line, std::string file) { + if (CompilerKit::find_word(line, "public_segment ")) return true; + + for (auto& opcode64x0 : kOpcodes64x0) { + // strict check here + if (CompilerKit::find_word(line, opcode64x0.fName) && Detail::algorithm::is_valid_64x0(line)) { + std::string name(opcode64x0.fName); + std::string jump_label, cpy_jump_label; + + kBytes.emplace_back(opcode64x0.fOpcode); + kBytes.emplace_back(opcode64x0.fFunct3); + kBytes.emplace_back(opcode64x0.fFunct7); + + // check funct7 type. + switch (opcode64x0.fFunct7) { + // reg to reg means register to register transfer operation. + case kAsmRegToReg: + case kAsmImmediate: { + // \brief how many registers we found. + std::size_t found_some = 0UL; + + for (size_t line_index = 0UL; line_index < line.size(); line_index++) { + if (line[line_index] == kAsmRegisterPrefix[0] && isdigit(line[line_index + 1])) { + std::string register_syntax = kAsmRegisterPrefix; + register_syntax += line[line_index + 1]; + + if (isdigit(line[line_index + 2])) register_syntax += line[line_index + 2]; + + std::string reg_str; + reg_str += line[line_index + 1]; + + if (isdigit(line[line_index + 2])) reg_str += line[line_index + 2]; + + // it ranges from r0 to r19 + // something like r190 doesn't exist in the instruction set. + if (kOutputArch == CompilerKit::kPefArch64000) { + if (isdigit(line[line_index + 3]) && isdigit(line[line_index + 2])) { + reg_str += line[line_index + 3]; + Detail::print_error("invalid register index, r" + reg_str + + "\nnote: The 64x0 accepts registers from r0 to r20.", + file); + throw std::runtime_error("invalid_register_index"); + } + } + + // finally cast to a size_t + std::size_t reg_index = strtol(reg_str.c_str(), nullptr, 10); + + if (reg_index > kAsmRegisterLimit) { + Detail::print_error("invalid register index, r" + reg_str, file); + throw std::runtime_error("invalid_register_index"); + } + + kBytes.emplace_back(reg_index); + ++found_some; + + if (kVerbose) { + kStdOut << "Assembler64x0: Register found: " << register_syntax << "\n"; + kStdOut << "Assembler64x0: Register amount in instruction: " << found_some << "\n"; + } + } + } + + // we're not in immediate addressing, reg to reg. + if (opcode64x0.fFunct7 != kAsmImmediate) { + // remember! register to register! + if (found_some == 1) { + Detail::print_error( + "Too few registers.\ntip: each Assembler64x0 register " + "starts with 'r'.\nline: " + + line, + file); + throw std::runtime_error("not_a_register"); + } + } + + if (found_some < 1 && name != "ldw" && name != "lda" && name != "stw") { + Detail::print_error("invalid combination of opcode and registers.\nline: " + line, + file); + throw std::runtime_error("invalid_comb_op_reg"); + } else if (found_some == 1 && name == "add") { + Detail::print_error("invalid combination of opcode and registers.\nline: " + line, + file); + throw std::runtime_error("invalid_comb_op_reg"); + } else if (found_some == 1 && name == "sub") { + Detail::print_error("invalid combination of opcode and registers.\nline: " + line, + file); + throw std::runtime_error("invalid_comb_op_reg"); + } + + if (found_some > 0 && name == "pop") { + Detail::print_error( + "invalid combination for opcode 'pop'.\ntip: it expects " + "nothing.\nline: " + + line, + file); + throw std::runtime_error("invalid_comb_op_pop"); + } + } + default: + break; + } + + // try to fetch a number from the name + if (name == "stw" || name == "ldw" || name == "lda" || name == "sta") { + auto where_string = name; + + // if we load something, we'd need it's symbol/literal + if (name == "stw" || name == "sta" || name == "ldw" || name == "lda" || name == "sta") + where_string = ","; + + jump_label = line; + + auto found_sym = false; + + while (jump_label.find(where_string) != std::string::npos) { + jump_label = jump_label.substr(jump_label.find(where_string) + where_string.size()); + + while (jump_label.find(" ") != std::string::npos) { + jump_label.erase(jump_label.find(" "), 1); + } + + if (jump_label[0] != kAsmRegisterPrefix[0] && !isdigit(jump_label[1])) { + if (found_sym) { + Detail::print_error( + "invalid combination of opcode and operands.\nhere -> " + jump_label, file); + throw std::runtime_error("invalid_comb_op_ops"); + } else { + // death trap installed. + found_sym = true; + } + } + } + + cpy_jump_label = jump_label; + + // replace any spaces with $ + if (jump_label[0] == ' ') { + while (jump_label.find(' ') != std::string::npos) { + if (isalnum(jump_label[0]) || isdigit(jump_label[0])) break; + + jump_label.erase(jump_label.find(' '), 1); + } + } + + if (!this->WriteNumber(0, jump_label)) { + // sta expects this: sta 0x000000, r0 + if (name == "sta") { + Detail::print_error("invalid combination of opcode and operands.\nHere ->" + line, + file); + throw std::runtime_error("invalid_comb_op_ops"); + } + } else { + if (name == "sta" && cpy_jump_label.find("extern_segment ") != std::string::npos) { + Detail::print_error("invalid usage extern_segment on 'sta', here: " + line, file); + throw std::runtime_error("invalid_sta_usage"); + } + } + + goto asm_write_label; + } + + // This is the case where we jump to a label, it is also used as a goto. + if (name == "lda" || name == "sta") { + asm_write_label: + if (cpy_jump_label.find('\n') != std::string::npos) + cpy_jump_label.erase(cpy_jump_label.find('\n'), 1); + + if (cpy_jump_label.find("extern_segment") != std::string::npos) { + cpy_jump_label.erase(cpy_jump_label.find("extern_segment"), strlen("extern_segment")); + + if (name == "sta") { + Detail::print_error("extern_segment is not allowed on a sta operation.", file); + throw std::runtime_error("extern_segment_sta_op"); + } else { + goto asm_end_label_cpy; + } + } + + if (name == "lda" || name == "sta") { + for (auto& label : kOriginLabel) { + if (cpy_jump_label == label.first) { + if (kVerbose) { + kStdOut << "Assembler64x0: Replace label " << cpy_jump_label + << " to address: " << label.second << std::endl; + } + + CompilerKit::NumberCast64 num(label.second); + + for (auto& num : num.number) { + kBytes.push_back(num); + } + + goto asm_end_label_cpy; + } + } + + if (cpy_jump_label[0] == '0') { + switch (cpy_jump_label[1]) { + case 'x': + case 'o': + case 'b': + if (this->WriteNumber(0, cpy_jump_label)) goto asm_end_label_cpy; + + break; + default: + break; + } + + if (isdigit(cpy_jump_label[0])) { + if (this->WriteNumber(0, cpy_jump_label)) goto asm_end_label_cpy; + + break; + } + } + } + + if (cpy_jump_label.size() < 1) { + Detail::print_error("label is empty, can't jump on it.", file); + throw std::runtime_error("label_empty"); + } + + /// don't go any further if: + /// load word (ldw) or store word. (stw) + + if (name == "ldw" || name == "stw") break; + + auto mld_reloc_str = std::to_string(cpy_jump_label.size()); + mld_reloc_str += kUndefinedSymbol; + mld_reloc_str += cpy_jump_label; + + bool ignore_back_slash = false; + + for (auto& reloc_chr : mld_reloc_str) { + if (reloc_chr == '\\') { + ignore_back_slash = true; + continue; + } + + if (ignore_back_slash) { + ignore_back_slash = false; + continue; + } + + kBytes.push_back(reloc_chr); + } + + kBytes.push_back('\0'); + goto asm_end_label_cpy; + } + + asm_end_label_cpy: + kOrigin += k64x0IPAlignment; + + break; + } + } + + return true; +} + +// Last rev 13-1-24 diff --git a/dev/CompilerKit/src/Backend/AssemblerAMD64.cc b/dev/CompilerKit/src/Backend/AssemblerAMD64.cc new file mode 100644 index 0000000..0446a10 --- /dev/null +++ b/dev/CompilerKit/src/Backend/AssemblerAMD64.cc @@ -0,0 +1,1206 @@ +/* ------------------------------------------- + + Copyright (C) 2024-2025 Amlal EL Mahrouss, all rights reserved + +------------------------------------------- */ + +///////////////////////////////////////////////////////////////////////////////////////// + +/// @file AssemblerAMD64.cc +/// @author EL Mahrouss Amlal +/// @brief AMD64 Assembler. +/// REMINDER: when dealing with an undefined symbol use (string +/// size):LinkerFindSymbol:(string) so that ld will look for it. + +///////////////////////////////////////////////////////////////////////////////////////// + +/// bugs: 0 + +/// feature request: 1 +/// Encode registers in mov, add, xor... + +///////////////////////////////////////////////////////////////////////////////////////// + +#include "CompilerKit/Defines.h" +#ifndef __ASM_NEED_AMD64__ +#define __ASM_NEED_AMD64__ 1 +#endif + +#define kAssemblerPragmaSymStr "#" +#define kAssemblerPragmaSym '#' + +#include <CompilerKit/AE.h> +#include <CompilerKit/detail/X64.h> +#include <CompilerKit/Frontend.h> +#include <CompilerKit/PEF.h> +#include <algorithm> +#include <cstdlib> +#include <filesystem> +#include <fstream> +#include <iostream> +#include <vector> + +///////////////////// + +// ANSI ESCAPE CODES + +///////////////////// + +#define kBlank "\e[0;30m" +#define kRed "\e[0;31m" +#define kWhite "\e[0;97m" +#define kYellow "\e[0;33m" + +static char kOutputArch = CompilerKit::kPefArchAMD64; + +constexpr auto kIPAlignement = 0x1U; + +static std::size_t kCounter = 1UL; + +static std::uintptr_t kOrigin = kPefBaseOrigin; +static std::vector<std::pair<std::string, std::uintptr_t>> kOriginLabel; + +/// @brief keep it simple by default. +static std::int32_t kRegisterBitWidth = 16U; + +static std::vector<i64_byte_t> kAppBytes; + +static CompilerKit::AERecordHeader kCurrentRecord{ + .fName = "", .fKind = CompilerKit::kPefCode, .fSize = 0, .fOffset = 0}; + +static std::vector<CompilerKit::AERecordHeader> kRecords; +static std::vector<std::string> kDefinedSymbols; +static std::vector<std::string> kUndefinedSymbols; + +static const std::string kUndefinedSymbol = ":UndefinedSymbol:"; + +// \brief forward decl. +static bool asm_read_attributes(std::string line); + +#include <CompilerKit/utils/AsmUtils.h> + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief AMD64 assembler entrypoint, the program/module starts here. + +///////////////////////////////////////////////////////////////////////////////////////// + +LIBCOMPILER_MODULE(AssemblerMainAMD64) { + //////////////// CPU OPCODES BEGIN //////////////// + + CompilerKit::install_signal(SIGSEGV, Detail::drvi_crash_handler); + + std::string opcodes_jump[kJumpLimit] = {"ja", "jae", "jb", "jbe", "jc", "je", "jg", "jge", + "jl", "jle", "jna", "jnae", "jnb", "jnbe", "jnc", "jne", + "jng", "jnge", "jnl", "jnle", "jno", "jnp", "jns", "jnz", + "jo", "jp", "jpe", "jpo", "js", "jz"}; + + for (i64_hword_t i = 0; i < kJumpLimit; i++) { + CpuOpcodeAMD64 code{.fName = opcodes_jump[i], + .fOpcode = static_cast<i64_hword_t>(kAsmJumpOpcode + i)}; + kOpcodesAMD64.push_back(code); + } + + CpuOpcodeAMD64 code{.fName = "jcxz", .fOpcode = 0xE3}; + kOpcodesAMD64.push_back(code); + + for (i64_hword_t i = kJumpLimitStandard; i < kJumpLimitStandardLimit; i++) { + CpuOpcodeAMD64 code{.fName = "jmp", .fOpcode = i}; + kOpcodesAMD64.push_back(code); + } + + CpuOpcodeAMD64 lahf{.fName = "lahf", .fOpcode = 0x9F}; + kOpcodesAMD64.push_back(lahf); + + CpuOpcodeAMD64 lds{.fName = "lds", .fOpcode = 0xC5}; + kOpcodesAMD64.push_back(lds); + + CpuOpcodeAMD64 lea{.fName = "lea", .fOpcode = 0x8D}; + kOpcodesAMD64.push_back(lea); + + CpuOpcodeAMD64 nop{.fName = "nop", .fOpcode = 0x90}; + kOpcodesAMD64.push_back(nop); + + //////////////// CPU OPCODES END //////////////// + + for (size_t i = 1; i < argc; ++i) { + if (argv[i][0] == '-') { + if (strcmp(argv[i], "--amd64:ver") == 0 || strcmp(argv[i], "--amd64:v") == 0) { + kStdOut << "AssemblerAMD64: AMD64 Assembler Driver.\nAssemblerAMD64: " + "v1.10\nAssemblerAMD64: Copyright " + "(c) Amlal El Mahrouss\n"; + return 0; + } else if (strcmp(argv[i], "--amd64:h") == 0) { + kStdOut << "AssemblerAMD64: AMD64 Assembler Driver.\nAssemblerAMD64: Copyright (c) 2024 " + "Amlal El Mahrouss\n"; + kStdOut << "--version: Print program version.\n"; + kStdOut << "--verbose: Print verbose output.\n"; + kStdOut << "--binary: Output as flat binary.\n"; + + return 0; + } else if (strcmp(argv[i], "--amd64:binary") == 0) { + kOutputAsBinary = true; + continue; + } else if (strcmp(argv[i], "--amd64:verbose") == 0) { + kVerbose = true; + continue; + } + + kStdOut << "AssemblerAMD64: ignore " << argv[i] << "\n"; + continue; + } + + if (!std::filesystem::exists(argv[i])) { + kStdOut << "AssemblerAMD64: can't open: " << argv[i] << std::endl; + goto asm_fail_exit; + } + + std::string object_output(argv[i]); + std::string asm_input(argv[i]); + + for (auto& ext : kAsmFileExts) { + if (object_output.ends_with(ext)) { + object_output.erase(object_output.find(ext), std::strlen(ext)); + break; + } + } + + object_output += kOutputAsBinary ? kBinaryFileExt : kObjectFileExt; + + std::ifstream file_ptr(argv[i]); + std::ofstream file_ptr_out(object_output, std::ofstream::binary); + + kStdOut << "AssemblerAMD64: Assembling: " << argv[i] << "\n"; + + if (file_ptr_out.bad()) { + if (kVerbose) { + kStdOut << "AssemblerAMD64: error: " << strerror(errno) << "\n"; + } + + return 1; + } + + std::string line; + + CompilerKit::AEHeader hdr{0}; + + memset(hdr.fPad, kAENullType, kAEPad); + + hdr.fMagic[0] = kAEMag0; + hdr.fMagic[1] = kAEMag1; + hdr.fSize = sizeof(CompilerKit::AEHeader); + hdr.fArch = kOutputArch; + + ///////////////////////////////////////////////////////////////////////////////////////// + + // COMPILATION LOOP + + ///////////////////////////////////////////////////////////////////////////////////////// + + CompilerKit::EncoderAMD64 asm64; + + if (kVerbose) { + kStdOut << "Compiling: " + asm_input << "\n"; + kStdOut << "From: " + line << "\n"; + } + + while (std::getline(file_ptr, line)) { + if (auto ln = asm64.CheckLine(line, argv[i]); !ln.empty()) { + Detail::print_error(ln, argv[i]); + continue; + } + + try { + asm_read_attributes(line); + asm64.WriteLine(line, argv[i]); + } catch (const std::exception& e) { + if (kVerbose) { + std::string what = e.what(); + Detail::print_warning("exit because of: " + what, "CompilerKit"); + } + + try { + std::filesystem::remove(object_output); + } catch (...) { + } + + goto asm_fail_exit; + } + } + + if (!kOutputAsBinary) { + if (kVerbose) { + kStdOut << "AssemblerAMD64: Writing object file...\n"; + } + + // this is the final step, write everything to the file. + + auto pos = file_ptr_out.tellp(); + + hdr.fCount = kRecords.size() + kUndefinedSymbols.size(); + + file_ptr_out << hdr; + + if (kRecords.empty()) { + kStdErr << "AssemblerAMD64: At least one record is needed to write an object " + "file.\nAssemblerAMD64: Make one using `public_segment .code64 foo_bar`.\n"; + + std::filesystem::remove(object_output); + return 1; + } + + kRecords[kRecords.size() - 1].fSize = kAppBytes.size(); + + std::size_t record_count = 0UL; + + for (auto& rec : kRecords) { + if (kVerbose) kStdOut << "AssemblerAMD64: Wrote record " << rec.fName << " to file...\n"; + + rec.fFlags |= CompilerKit::kKindRelocationAtRuntime; + rec.fOffset = record_count; + ++record_count; + + file_ptr_out << rec; + } + + // increment once again, so that we won't lie about the kUndefinedSymbols. + ++record_count; + + for (auto& sym : kUndefinedSymbols) { + CompilerKit::AERecordHeader _record_hdr{0}; + + if (kVerbose) kStdOut << "AssemblerAMD64: Wrote symbol " << sym << " to file...\n"; + + _record_hdr.fKind = kAENullType; + _record_hdr.fSize = sym.size(); + _record_hdr.fOffset = record_count; + + ++record_count; + + memset(_record_hdr.fPad, kAENullType, kAEPad); + memcpy(_record_hdr.fName, sym.c_str(), sym.size()); + + file_ptr_out << _record_hdr; + + ++kCounter; + } + + auto pos_end = file_ptr_out.tellp(); + + file_ptr_out.seekp(pos); + + hdr.fStartCode = pos_end; + hdr.fCodeSize = kAppBytes.size(); + + file_ptr_out << hdr; + + file_ptr_out.seekp(pos_end); + } else { + if (kVerbose) { + kStdOut << "AssemblerAMD64: Write raw binary...\n"; + } + } + + // byte from byte, we write this. + for (auto& byte : kAppBytes) { + if (byte == 0) continue; + + if (byte == 0xFF) { + byte = 0; + } + + file_ptr_out << reinterpret_cast<const char*>(&byte)[0]; + } + + if (kVerbose) kStdOut << "AssemblerAMD64: Wrote file with program in it.\n"; + + file_ptr_out.flush(); + file_ptr_out.close(); + + if (kVerbose) kStdOut << "AssemblerAMD64: Exit succeeded.\n"; + + return 0; + } + +asm_fail_exit: + + if (kVerbose) kStdOut << "AssemblerAMD64: Exit failed.\n"; + + return 1; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Check for attributes +// returns true if any was found. + +///////////////////////////////////////////////////////////////////////////////////////// + +static bool asm_read_attributes(std::string line) { + // extern_segment is the opposite of public_segment, it signals to the ld + // that we need this symbol. + if (CompilerKit::find_word(line, "extern_segment")) { + if (kOutputAsBinary) { + Detail::print_error("Invalid directive in flat binary mode.", "CompilerKit"); + throw std::runtime_error("invalid_extern_segment_bin"); + } + + auto name = line.substr(line.find("extern_segment") + strlen("extern_segment") + 1); + + if (name.size() == 0) { + Detail::print_error("Invalid extern_segment", "power-as"); + throw std::runtime_error("invalid_extern_segment"); + } + + std::string result = std::to_string(name.size()); + result += kUndefinedSymbol; + + // mangle this + for (char& j : name) { + if (j == ' ' || j == ',') j = '$'; + } + + result += name; + + if (name.find(kPefCode64) != std::string::npos) { + // data is treated as code. + kCurrentRecord.fKind = CompilerKit::kPefCode; + } else if (name.find(kPefData64) != std::string::npos) { + // no code will be executed from here. + kCurrentRecord.fKind = CompilerKit::kPefData; + } else if (name.find(kPefZero64) != std::string::npos) { + // this is a bss section. + kCurrentRecord.fKind = CompilerKit::kPefZero; + } + + // this is a special case for the start stub. + // we want this so that ld can find it. + + if (name == kPefStart) { + kCurrentRecord.fKind = CompilerKit::kPefCode; + } + + // now we can tell the code size of the previous kCurrentRecord. + + if (!kRecords.empty()) kRecords[kRecords.size() - 1].fSize = kAppBytes.size(); + + memset(kCurrentRecord.fName, 0, kAESymbolLen); + memcpy(kCurrentRecord.fName, result.c_str(), result.size()); + + ++kCounter; + + memset(kCurrentRecord.fPad, kAENullType, kAEPad); + + kRecords.emplace_back(kCurrentRecord); + + return true; + } + // public_segment is a special keyword used by AssemblerAMD64 to tell the AE output stage to + // mark this section as a header. it currently supports .code64, .data64 and + // .zero64. + else if (CompilerKit::find_word(line, "public_segment")) { + if (kOutputAsBinary) { + Detail::print_error("Invalid directive in flat binary mode.", "CompilerKit"); + throw std::runtime_error("invalid_public_segment_bin"); + } + + auto name = line.substr(line.find("public_segment") + strlen("public_segment") + 1); + + std::string name_copy = name; + + for (char& j : name) { + if (j == ' ') j = '$'; + } + + if (std::find(kDefinedSymbols.begin(), kDefinedSymbols.end(), name) != kDefinedSymbols.end()) { + Detail::print_error("Symbol already defined.", "CompilerKit"); + throw std::runtime_error("invalid_public_segment_bin"); + } + + kDefinedSymbols.push_back(name); + + if (name.find(".code64") != std::string::npos) { + // data is treated as code. + + name_copy.erase(name_copy.find(".code64"), strlen(".code64")); + kCurrentRecord.fKind = CompilerKit::kPefCode; + } else if (name.find(".data64") != std::string::npos) { + // no code will be executed from here. + + name_copy.erase(name_copy.find(".data64"), strlen(".data64")); + kCurrentRecord.fKind = CompilerKit::kPefData; + } else if (name.find(".zero64") != std::string::npos) { + // this is a bss section. + + name_copy.erase(name_copy.find(".zero64"), strlen(".zero64")); + kCurrentRecord.fKind = CompilerKit::kPefZero; + } + + // this is a special case for the start stub. + // we want this so that ld can find it. + + if (name == kPefStart) { + kCurrentRecord.fKind = CompilerKit::kPefCode; + } + + while (name_copy.find(" ") != std::string::npos) name_copy.erase(name_copy.find(" "), 1); + + kOriginLabel.push_back(std::make_pair(name_copy, kOrigin)); + ++kOrigin; + + // now we can tell the code size of the previous kCurrentRecord. + + if (!kRecords.empty()) kRecords[kRecords.size() - 1].fSize = kAppBytes.size(); + + memset(kCurrentRecord.fName, 0, kAESymbolLen); + memcpy(kCurrentRecord.fName, name.c_str(), name.size()); + + ++kCounter; + + memset(kCurrentRecord.fPad, kAENullType, kAEPad); + + kRecords.emplace_back(kCurrentRecord); + + return true; + } + + return false; +} + +// \brief algorithms and helpers. + +namespace Detail::algorithm { +// \brief authorize a brief set of characters. +static inline bool is_not_valid(char c) { + if ((isalpha(c) || isdigit(c)) || + ((c == ' ') || (c == '\t') || (c == ',') || (c == '(') || (c == ')') || (c == '"') || + (c == '*') || (c == '\'') || (c == '[') || (c == ']') || (c == '+') || (c == '_') || + (c == ':') || (c == '@') || (c == '.') || (c == '#') || (c == ';'))) + return false; + + return true; +} + +bool is_valid_amd64(std::string str) { + return std::find_if(str.begin(), str.end(), is_not_valid) == str.end(); +} +} // namespace Detail::algorithm + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Check for line (syntax check) + +///////////////////////////////////////////////////////////////////////////////////////// + +std::string CompilerKit::EncoderAMD64::CheckLine(std::string line, std::string file) { + std::string err_str; + + if (line.empty() || CompilerKit::find_word(line, "extern_segment") || + CompilerKit::find_word(line, "public_segment") || + CompilerKit::find_word(line, kAssemblerPragmaSymStr) || CompilerKit::find_word(line, ";") || + line[0] == kAssemblerPragmaSym) { + if (line.find(';') != std::string::npos) { + line.erase(line.find(';')); + } else { + // now check the line for validity + if (!Detail::algorithm::is_valid_amd64(line)) { + err_str = "Line contains non valid characters.\nhere -> "; + err_str += line; + } + } + + return err_str; + } + + // check for a valid instruction format. + + if (line.find(',') != std::string::npos) { + if (line.find(',') + 1 == line.size()) { + err_str += "\nInstruction lacks right register, here -> "; + err_str += line.substr(line.find(',')); + + return err_str; + } else { + bool nothing_on_right = true; + + if (line.find(',') + 1 > line.size()) { + err_str += "\nInstruction not complete, here -> "; + err_str += line; + + return err_str; + } + + auto substr = line.substr(line.find(',') + 1); + + for (auto& ch : substr) { + if (ch != ' ' && ch != '\t') { + nothing_on_right = false; + } + } + + // this means we found nothing after that ',' . + if (nothing_on_right) { + err_str += "\nInstruction not complete, here -> "; + err_str += line; + + return err_str; + } + } + } + for (auto& opcodeAMD64 : kOpcodesAMD64) { + if (CompilerKit::find_word(line, opcodeAMD64.fName)) { + return err_str; + } + } + + err_str += "\nUnrecognized instruction -> " + line; + + return err_str; +} + +bool CompilerKit::EncoderAMD64::WriteNumber(const std::size_t& pos, std::string& jump_label) { + if (!isdigit(jump_label[pos])) return false; + + switch (jump_label[pos + 1]) { + case 'x': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16); !res) { + if (errno != 0) { + Detail::print_error("invalid hex number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_hex"); + } + } + + CompilerKit::NumberCast64 num = + CompilerKit::NumberCast64(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16)); + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 16 number here: " << jump_label.substr(pos) + << "\n"; + } + + return true; + } + case 'b': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2); !res) { + if (errno != 0) { + Detail::print_error("invalid binary number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_bin"); + } + } + + CompilerKit::NumberCast64 num = + CompilerKit::NumberCast64(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2)); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 2 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + return true; + } + case 'o': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7); !res) { + if (errno != 0) { + Detail::print_error("invalid octal number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_octal"); + } + } + + CompilerKit::NumberCast64 num = + CompilerKit::NumberCast64(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7)); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 8 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + return true; + } + default: { + break; + } + } + + /* check for errno and stuff like that */ + if (auto res = strtol(jump_label.substr(pos).c_str(), nullptr, 10); !res) { + if (errno != 0) { + return false; + } + } + + CompilerKit::NumberCast64 num = + CompilerKit::NumberCast64(strtol(jump_label.substr(pos).c_str(), nullptr, 10)); + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 10 number here: " << jump_label.substr(pos) << "\n"; + } + + return true; +} + +bool CompilerKit::EncoderAMD64::WriteNumber32(const std::size_t& pos, std::string& jump_label) { + if (!isdigit(jump_label[pos])) return false; + + switch (jump_label[pos + 1]) { + case 'x': { + auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16); + res += kOrigin; + + if (errno != 0) { + return false; + } + + CompilerKit::NumberCast32 num = CompilerKit::NumberCast32(res); + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 16 number here: " << jump_label.substr(pos) + << "\n"; + } + + return true; + } + case 'b': { + auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2); + res += kOrigin; + + if (errno != 0) { + return false; + } + + CompilerKit::NumberCast32 num = CompilerKit::NumberCast32(res); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 2 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + return true; + } + case 'o': { + auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7); + res += kOrigin; + + if (errno != 0) { + return false; + } + + CompilerKit::NumberCast32 num = CompilerKit::NumberCast32(res); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 8 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + return true; + } + default: { + break; + } + } + + auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 10); + res += kOrigin; + + if (errno != 0) { + return false; + } + + CompilerKit::NumberCast32 num = CompilerKit::NumberCast32(res); + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 10 number here: " << jump_label.substr(pos) << "\n"; + } + + return true; +} + +bool CompilerKit::EncoderAMD64::WriteNumber16(const std::size_t& pos, std::string& jump_label) { + if (!isdigit(jump_label[pos])) return false; + + switch (jump_label[pos + 1]) { + case 'x': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16); !res) { + if (errno != 0) { + Detail::print_error("invalid hex number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_hex"); + } + } + + CompilerKit::NumberCast16 num = + CompilerKit::NumberCast16(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16)); + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 16 number here: " << jump_label.substr(pos) + << "\n"; + } + + return true; + } + case 'b': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2); !res) { + if (errno != 0) { + Detail::print_error("invalid binary number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_bin"); + } + } + + CompilerKit::NumberCast16 num = + CompilerKit::NumberCast16(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2)); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 2 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + return true; + } + case 'o': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7); !res) { + if (errno != 0) { + Detail::print_error("invalid octal number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_octal"); + } + } + + CompilerKit::NumberCast16 num = + CompilerKit::NumberCast16(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7)); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 8 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + return true; + } + default: { + break; + } + } + + /* check for errno and stuff like that */ + if (auto res = strtol(jump_label.substr(pos).c_str(), nullptr, 10); !res) { + if (errno != 0) { + return false; + } + } + + CompilerKit::NumberCast16 num = + CompilerKit::NumberCast16(strtol(jump_label.substr(pos).c_str(), nullptr, 10)); + + for (char& i : num.number) { + if (i == 0) i = 0xFF; + + kAppBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 10 number here: " << jump_label.substr(pos) << "\n"; + } + + return true; +} + +bool CompilerKit::EncoderAMD64::WriteNumber8(const std::size_t& pos, std::string& jump_label) { + if (!isdigit(jump_label[pos])) return false; + + switch (jump_label[pos + 1]) { + case 'x': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16); !res) { + if (errno != 0) { + Detail::print_error("invalid hex number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_hex"); + } + } + + CompilerKit::NumberCast8 num = + CompilerKit::NumberCast8(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16)); + + kAppBytes.push_back(num.number); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 16 number here: " << jump_label.substr(pos) + << "\n"; + } + + return true; + } + case 'b': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2); !res) { + if (errno != 0) { + Detail::print_error("invalid binary number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_bin"); + } + } + + CompilerKit::NumberCast8 num = + CompilerKit::NumberCast8(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2)); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 2 number here: " << jump_label.substr(pos) << "\n"; + } + + kAppBytes.push_back(num.number); + + return true; + } + case 'o': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7); !res) { + if (errno != 0) { + Detail::print_error("invalid octal number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_octal"); + } + } + + CompilerKit::NumberCast8 num = + CompilerKit::NumberCast8(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7)); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 8 number here: " << jump_label.substr(pos) << "\n"; + } + + kAppBytes.push_back(num.number); + + return true; + } + default: { + break; + } + } + + /* check for errno and stuff like that */ + if (auto res = strtol(jump_label.substr(pos).c_str(), nullptr, 10); !res) { + if (errno != 0) { + return false; + } + } + + CompilerKit::NumberCast8 num = + CompilerKit::NumberCast8(strtol(jump_label.substr(pos).c_str(), nullptr, 10)); + + kAppBytes.push_back(num.number); + + if (kVerbose) { + kStdOut << "AssemblerAMD64: Found a base 10 number here: " << jump_label.substr(pos) << "\n"; + } + + return true; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Read and write an instruction to the output array. + +///////////////////////////////////////////////////////////////////////////////////////// + +bool CompilerKit::EncoderAMD64::WriteLine(std::string line, std::string file) { + if (CompilerKit::find_word(line, "public_segment ")) return true; + + struct RegMapAMD64 { + CompilerKit::STLString fName; + i64_byte_t fModRM; + }; + + std::vector<RegMapAMD64> kRegisterList{ + {.fName = "ax", .fModRM = 0x0}, {.fName = "cx", .fModRM = 1}, + {.fName = "dx", .fModRM = 0x2}, {.fName = "bx", .fModRM = 3}, + {.fName = "sp", .fModRM = 0x4}, {.fName = "bp", .fModRM = 5}, + {.fName = "si", .fModRM = 0x6}, {.fName = "di", .fModRM = 7}, + }; + + BOOL foundInstruction = false; + + for (auto& opcodeAMD64 : kOpcodesAMD64) { + // strict check here + if (CompilerKit::find_word(line, opcodeAMD64.fName) && + Detail::algorithm::is_valid_amd64(line)) { + foundInstruction = true; + std::string name(opcodeAMD64.fName); + + /// Move instruction handler. + if (line.find(name) != std::string::npos) { + if (name == "mov" || name == "xor") { + std::string substr = line.substr(line.find(name) + name.size()); + + uint64_t bits = kRegisterBitWidth; + + if (substr.find(",") == std::string::npos) { + Detail::print_error("Syntax error: missing right operand.", "CompilerKit"); + throw std::runtime_error("syntax_err"); + } + + bool onlyOneReg = true; + + std::vector<RegMapAMD64> currentRegList; + + for (auto& reg : kRegisterList) { + std::vector<char> regExt = {'e', 'r'}; + + for (auto& ext : regExt) { + std::string registerName; + + if (bits > 16) registerName.push_back(ext); + + registerName += reg.fName; + + while (line.find(registerName) != std::string::npos) { + line.erase(line.find(registerName), registerName.size()); + + if (bits == 16) { + if (registerName[0] == 'r') { + Detail::print_error("invalid size for register, current bit width is: " + + std::to_string(kRegisterBitWidth), + file); + throw std::runtime_error("invalid_reg_size"); + } + } + + currentRegList.push_back({.fName = registerName, .fModRM = reg.fModRM}); + } + } + } + + if (currentRegList.size() > 1) onlyOneReg = false; + + bool hasRBasedRegs = false; + + if (!onlyOneReg) { + /// very tricky to understand. + /// but this checks for a r8 through r15 register. + if (currentRegList[0].fName[0] == 'r' || currentRegList[1].fName[0] == 'r') { + if (isdigit(currentRegList[0].fName[1]) && isdigit(currentRegList[1].fName[1])) { + kAppBytes.emplace_back(0x4d); + hasRBasedRegs = true; + } else if (isdigit(currentRegList[0].fName[1]) || + isdigit(currentRegList[1].fName[1])) { + kAppBytes.emplace_back(0x4c); + hasRBasedRegs = true; + } + } + } + + if (name == "mov") { + if (bits == 64 || bits == 32) { + if (!hasRBasedRegs && bits >= 32) { + kAppBytes.emplace_back(opcodeAMD64.fOpcode); + } + + if (!onlyOneReg) kAppBytes.emplace_back(0x89); + } else if (bits == 16) { + if (hasRBasedRegs) { + Detail::print_error("Invalid combination of operands and registers.", + "CompilerKit"); + throw std::runtime_error("comb_op_reg"); + } else { + kAppBytes.emplace_back(0x66); + kAppBytes.emplace_back(0x89); + } + } + } else { + if (!hasRBasedRegs && bits >= 32) { + kAppBytes.emplace_back(opcodeAMD64.fOpcode); + } + + kAppBytes.emplace_back(0x31); + } + + if (onlyOneReg) { + auto num = GetNumber32(line, ","); + + for (auto& num_idx : num.number) { + if (num_idx == 0) num_idx = 0xFF; + } + + auto modrm = (0x3 << 6 | currentRegList[0].fModRM); + + kAppBytes.emplace_back(0xC7); // prefixed before placing the modrm and then the number. + kAppBytes.emplace_back(modrm); + + if (name != "xor") { + kAppBytes.emplace_back(num.number[0]); + kAppBytes.emplace_back(num.number[1]); + kAppBytes.emplace_back(num.number[2]); + kAppBytes.emplace_back(num.number[3]); + } + + break; + } + + if (currentRegList[1].fName[0] == 'r' && currentRegList[0].fName[0] == 'e') { + Detail::print_error("Invalid combination of operands and registers.", "CompilerKit"); + throw std::runtime_error("comb_op_reg"); + } + + if (currentRegList[0].fName[0] == 'r' && currentRegList[1].fName[0] == 'e') { + Detail::print_error("Invalid combination of operands and registers.", "CompilerKit"); + throw std::runtime_error("comb_op_reg"); + } + + if (bits == 16) { + if (currentRegList[0].fName[0] == 'r' || currentRegList[0].fName[0] == 'e') { + Detail::print_error("Invalid combination of operands and registers.", "CompilerKit"); + throw std::runtime_error("comb_op_reg"); + } + + if (currentRegList[1].fName[0] == 'r' || currentRegList[1].fName[0] == 'e') { + Detail::print_error("Invalid combination of operands and registers.", "CompilerKit"); + throw std::runtime_error("comb_op_reg"); + } + } else { + if (currentRegList[0].fName[0] != 'r' || currentRegList[0].fName[0] == 'e') { + Detail::print_error("Invalid combination of operands and registers.", "CompilerKit"); + throw std::runtime_error("comb_op_reg"); + } + + if (currentRegList[1].fName[0] != 'r' || currentRegList[1].fName[0] == 'e') { + Detail::print_error("Invalid combination of operands and registers.", "CompilerKit"); + throw std::runtime_error("comb_op_reg"); + } + } + + /// encode register using the modrm encoding. + + auto modrm = (0x3 << 6 | currentRegList[1].fModRM << 3 | currentRegList[0].fModRM); + + kAppBytes.emplace_back(modrm); + + break; + } + } + + if (name == "int" || name == "into" || name == "intd") { + kAppBytes.emplace_back(opcodeAMD64.fOpcode); + this->WriteNumber8(line.find(name) + name.size() + 1, line); + + break; + } else if (name == "jmp" || name == "call") { + kAppBytes.emplace_back(opcodeAMD64.fOpcode); + + if (!this->WriteNumber32(line.find(name) + name.size() + 1, line)) { + throw std::runtime_error("BUG: WriteNumber32"); + } + + break; + } else if (name == "syscall") { + kAppBytes.emplace_back(opcodeAMD64.fOpcode); + kAppBytes.emplace_back(0x05); + + break; + } else { + kAppBytes.emplace_back(opcodeAMD64.fOpcode); + + break; + } + } + } + + if (line[0] == kAssemblerPragmaSym) { + if (foundInstruction) { + Detail::print_error("Syntax error: " + line, file); + throw std::runtime_error("syntax_err"); + } + + if (line.find("bits 64") != std::string::npos) { + kRegisterBitWidth = 64U; + } else if (line.find("bits 32") != std::string::npos) { + kRegisterBitWidth = 32U; + } else if (line.find("bits 16") != std::string::npos) { + kRegisterBitWidth = 16U; + } else if (line.find("org") != std::string::npos) { + size_t base[] = {10, 16, 2, 7}; + + for (size_t i = 0; i < 4; i++) { + if (kOrigin = strtol((line.substr(line.find("org") + strlen("org") + 1)).c_str(), nullptr, + base[i]); + kOrigin) { + if (errno != 0) { + continue; + } else { + if (kVerbose) { + kStdOut << "AssemblerAMD64: Origin Set: " << kOrigin << std::endl; + } + + break; + } + } + } + } + } + /// write a dword + else if (line.find(".dword") != std::string::npos) { + this->WriteNumber32(line.find(".dword") + strlen(".dword") + 1, line); + } + /// write a long + else if (line.find(".long") != std::string::npos) { + this->WriteNumber(line.find(".long") + strlen(".long") + 1, line); + } + /// write a 16-bit number + else if (line.find(".word") != std::string::npos) { + this->WriteNumber16(line.find(".word") + strlen(".word") + 1, line); + } + + kOrigin += kIPAlignement; + + return true; +} + +// Last rev 13-1-24 diff --git a/dev/CompilerKit/src/Backend/AssemblerARM64.cc b/dev/CompilerKit/src/Backend/AssemblerARM64.cc new file mode 100644 index 0000000..d290e24 --- /dev/null +++ b/dev/CompilerKit/src/Backend/AssemblerARM64.cc @@ -0,0 +1,587 @@ +/* ------------------------------------------- + + Copyright (C) 2024-2025 Amlal EL Mahrouss, all rights reserved + +------------------------------------------- */ + +///////////////////////////////////////////////////////////////////////////////////////// + +/// @file AssemblerARM64.cxx +/// @author EL Mahrouss Amlal +/// @brief 'ACORN' Assembler. + +/// REMINDER: when dealing with an undefined symbol use (string +/// size):LinkerFindSymbol:(string) so that li will look for it. + +///////////////////////////////////////////////////////////////////////////////////////// + +#ifndef __ASM_NEED_ARM64__ +#define __ASM_NEED_ARM64__ 1 +#endif + +#include <CompilerKit/AE.h> +#include <CompilerKit/detail/Aarch64.h> +#include <CompilerKit/ErrorID.h> +#include <CompilerKit/Frontend.h> +#include <CompilerKit/PEF.h> +#include <CompilerKit/utils/AsmUtils.h> +#include <CompilerKit/Version.h> +#include <algorithm> +#include <filesystem> +#include <fstream> +#include <iostream> +#include <vector> + +///////////////////// + +// ANSI ESCAPE CODES + +///////////////////// + +#define kBlank "\e[0;30m" +#define kRed "\e[0;31m" +#define kWhite "\e[0;97m" +#define kYellow "\e[0;33m" + +constexpr auto cPowerIPAlignment = 0x1U; + +static Char kOutputArch = CompilerKit::kPefArchARM64; + +static std::size_t kCounter = 1UL; + +static std::uintptr_t kOrigin = kPefBaseOrigin; +static std::vector<std::pair<std::string, std::uintptr_t>> kOriginLabel; + +static std::vector<uint8_t> kBytes; + +static CompilerKit::AERecordHeader kCurrentRecord{ + .fName = "", .fKind = CompilerKit::kPefCode, .fSize = 0, .fOffset = 0}; + +static std::vector<CompilerKit::AERecordHeader> kRecords; +static std::vector<std::string> kUndefinedSymbols; + +static const std::string kUndefinedSymbol = ":UndefinedSymbol:"; +static const std::string kRelocSymbol = ":RuntimeSymbol:"; + +// \brief forward decl. +static bool asm_read_attributes(std::string line); + +///////////////////////////////////////////////////////////////////////////////////////// + +/// @brief POWER assembler entrypoint, the program/module starts here. + +///////////////////////////////////////////////////////////////////////////////////////// + +LIBCOMPILER_MODULE(AssemblerMainARM64) { + CompilerKit::install_signal(SIGSEGV, Detail::drvi_crash_handler); + + for (size_t i = 1; i < argc; ++i) { + if (argv[i][0] == '-') { + if (strcmp(argv[i], "--ver") == 0 || strcmp(argv[i], "--v") == 0) { + kStdOut << "AssemblerPower: AARCH64 Assembler Driver.\nAssemblerPower: " << kDistVersion + << "\nAssemblerPower: " + "Copyright (c) " + "Amlal El Mahrouss\n"; + return 0; + } else if (strcmp(argv[i], "--h") == 0) { + kStdOut << "AssemblerPower: AARCH64 Assembler Driver.\nAssemblerPower: Copyright (c) 2024 " + "Amlal El Mahrouss\n"; + kStdOut << "--version,/v: print program version.\n"; + kStdOut << "--verbose: print verbose output.\n"; + kStdOut << "--binary: output as flat binary.\n"; + + return 0; + } else if (strcmp(argv[i], "--binary") == 0) { + kOutputAsBinary = true; + continue; + } else if (strcmp(argv[i], "--verbose") == 0) { + kVerbose = true; + continue; + } + + kStdOut << "AssemblerPower: ignore " << argv[i] << "\n"; + continue; + } + + if (!std::filesystem::exists(argv[i])) { + kStdOut << "AssemblerPower: can't open: " << argv[i] << std::endl; + goto asm_fail_exit; + } + + std::string object_output(argv[i]); + + for (auto& ext : kAsmFileExts) { + if (object_output.find(ext) != std::string::npos) { + object_output.erase(object_output.find(ext), std::strlen(ext)); + } + } + + object_output += kOutputAsBinary ? kBinaryFileExt : kObjectFileExt; + + std::ifstream file_ptr(argv[i]); + std::ofstream file_ptr_out(object_output, std::ofstream::binary); + + if (file_ptr_out.bad()) { + if (kVerbose) { + kStdOut << "AssemblerPower: error: " << strerror(errno) << "\n"; + } + } + + std::string line; + + CompilerKit::AEHeader hdr{0}; + + memset(hdr.fPad, kAENullType, kAEPad); + + hdr.fMagic[0] = kAEMag0; + hdr.fMagic[1] = kAEMag1; + hdr.fSize = sizeof(CompilerKit::AEHeader); + hdr.fArch = kOutputArch; + + ///////////////////////////////////////////////////////////////////////////////////////// + + // COMPILATION LOOP + + ///////////////////////////////////////////////////////////////////////////////////////// + + CompilerKit::EncoderARM64 asm64; + + while (std::getline(file_ptr, line)) { + if (auto ln = asm64.CheckLine(line, argv[i]); !ln.empty()) { + Detail::print_error(ln, argv[i]); + continue; + } + + try { + asm_read_attributes(line); + asm64.WriteLine(line, argv[i]); + } catch (const std::exception& e) { + if (kVerbose) { + std::string what = e.what(); + Detail::print_warning("exit because of: " + what, "CompilerKit"); + } + + std::filesystem::remove(object_output); + goto asm_fail_exit; + } + } + + if (!kOutputAsBinary) { + if (kVerbose) { + kStdOut << "AssemblerARM64: Writing object file...\n"; + } + + // this is the final step, write everything to the file. + + auto pos = file_ptr_out.tellp(); + + hdr.fCount = kRecords.size() + kUndefinedSymbols.size(); + + file_ptr_out << hdr; + + if (kRecords.empty()) { + kStdErr << "AssemblerARM64: At least one record is needed to write an object " + "file.\nAssemblerARM64: Make one using `public_segment .code64 foo_bar`.\n"; + + std::filesystem::remove(object_output); + return 1; + } + + kRecords[kRecords.size() - 1].fSize = kBytes.size(); + + std::size_t record_count = 0UL; + + for (auto& record_hdr : kRecords) { + record_hdr.fFlags |= CompilerKit::kKindRelocationAtRuntime; + record_hdr.fOffset = record_count; + ++record_count; + + file_ptr_out << record_hdr; + + if (kVerbose) kStdOut << "AssemblerARM64: Wrote record " << record_hdr.fName << "...\n"; + } + + // increment once again, so that we won't lie about the kUndefinedSymbols. + ++record_count; + + for (auto& sym : kUndefinedSymbols) { + CompilerKit::AERecordHeader undefined_sym{0}; + + if (kVerbose) kStdOut << "AssemblerARM64: Wrote symbol " << sym << " to file...\n"; + + undefined_sym.fKind = kAENullType; + undefined_sym.fSize = sym.size(); + undefined_sym.fOffset = record_count; + + ++record_count; + + memset(undefined_sym.fPad, kAENullType, kAEPad); + memcpy(undefined_sym.fName, sym.c_str(), sym.size()); + + file_ptr_out << undefined_sym; + + ++kCounter; + } + + auto pos_end = file_ptr_out.tellp(); + + file_ptr_out.seekp(pos); + + hdr.fStartCode = pos_end; + hdr.fCodeSize = kBytes.size(); + + file_ptr_out << hdr; + + file_ptr_out.seekp(pos_end); + } else { + if (kVerbose) { + kStdOut << "AssemblerARM64: Write raw binary...\n"; + } + } + + // byte from byte, we write this. + for (auto& byte : kBytes) { + file_ptr_out.write(reinterpret_cast<const char*>(&byte), sizeof(byte)); + } + + if (kVerbose) kStdOut << "AssemblerARM64: Wrote file with program in it.\n"; + + file_ptr_out.flush(); + file_ptr_out.close(); + + if (kVerbose) kStdOut << "AssemblerARM64: Exit succeeded.\n"; + + return 0; + } + +asm_fail_exit: + + if (kVerbose) kStdOut << "AssemblerARM64: Exit failed.\n"; + + return LIBCOMPILER_EXEC_ERROR; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Check for attributes +// returns true if any was found. + +///////////////////////////////////////////////////////////////////////////////////////// + +static bool asm_read_attributes(std::string line) { + // extern_segment is the opposite of public_segment, it signals to the li + // that we need this symbol. + if (CompilerKit::find_word(line, "extern_segment")) { + if (kOutputAsBinary) { + Detail::print_error("Invalid extern_segment directive in flat binary mode.", "CompilerKit"); + throw std::runtime_error("invalid_extern_segment_bin"); + } + + auto name = line.substr(line.find("extern_segment") + strlen("extern_segment") + 1); + + if (name.size() == 0) { + Detail::print_error("Invalid extern_segment", "CompilerKit"); + throw std::runtime_error("invalid_extern_segment"); + } + + std::string result = std::to_string(name.size()); + result += kUndefinedSymbol; + + // mangle this + for (char& j : name) { + if (j == ' ' || j == ',') j = '$'; + } + + result += name; + + if (name.find(".code64") != std::string::npos) { + // data is treated as code. + kCurrentRecord.fKind = CompilerKit::kPefCode; + } else if (name.find(".data64") != std::string::npos) { + // no code will be executed from here. + kCurrentRecord.fKind = CompilerKit::kPefData; + } else if (name.find(".zero64") != std::string::npos) { + // this is a bss section. + kCurrentRecord.fKind = CompilerKit::kPefZero; + } + + // this is a special case for the start stub. + // we want this so that li can find it. + + if (name == kPefStart) { + kCurrentRecord.fKind = CompilerKit::kPefCode; + } + + // now we can tell the code size of the previous kCurrentRecord. + + if (!kRecords.empty()) kRecords[kRecords.size() - 1].fSize = kBytes.size(); + + memset(kCurrentRecord.fName, 0, kAESymbolLen); + memcpy(kCurrentRecord.fName, result.c_str(), result.size()); + + ++kCounter; + + memset(kCurrentRecord.fPad, kAENullType, kAEPad); + + kRecords.emplace_back(kCurrentRecord); + + return true; + } + // public_segment is a special keyword used by Assembler to tell the AE output stage to + // mark this section as a header. it currently supports .code64, .data64., + // .zero64 + else if (CompilerKit::find_word(line, "public_segment")) { + if (kOutputAsBinary) { + Detail::print_error("Invalid public_segment directive in flat binary mode.", "CompilerKit"); + throw std::runtime_error("invalid_public_segment_bin"); + } + + auto name = line.substr(line.find("public_segment") + strlen("public_segment")); + + std::string name_copy = name; + + for (char& j : name) { + if (j == ' ') j = '$'; + } + + if (name.find(".code64") != std::string::npos) { + // data is treated as code. + + name_copy.erase(name_copy.find(".code64"), strlen(".code64")); + kCurrentRecord.fKind = CompilerKit::kPefCode; + } else if (name.find(".data64") != std::string::npos) { + // no code will be executed from here. + + name_copy.erase(name_copy.find(".data64"), strlen(".data64")); + kCurrentRecord.fKind = CompilerKit::kPefData; + } else if (name.find(".zero64") != std::string::npos) { + // this is a bss section. + + name_copy.erase(name_copy.find(".zero64"), strlen(".zero64")); + kCurrentRecord.fKind = CompilerKit::kPefZero; + } + + // this is a special case for the start stub. + // we want this so that li can find it. + + if (name == kPefStart) { + kCurrentRecord.fKind = CompilerKit::kPefCode; + } + + while (name_copy.find(" ") != std::string::npos) name_copy.erase(name_copy.find(" "), 1); + + kOriginLabel.push_back(std::make_pair(name_copy, kOrigin)); + ++kOrigin; + + // now we can tell the code size of the previous kCurrentRecord. + + if (!kRecords.empty()) kRecords[kRecords.size() - 1].fSize = kBytes.size(); + + memset(kCurrentRecord.fName, 0, kAESymbolLen); + memcpy(kCurrentRecord.fName, name.c_str(), name.size()); + + ++kCounter; + + memset(kCurrentRecord.fPad, kAENullType, kAEPad); + + kRecords.emplace_back(kCurrentRecord); + + return true; + } + + return false; +} + +// \brief algorithms and helpers. + +namespace Detail::algorithm { +// \brief authorize a brief set of characters. +static inline bool is_not_alnum_space(char c) { + return !(isalpha(c) || isdigit(c) || (c == ' ') || (c == '\t') || (c == ',') || (c == '(') || + (c == ')') || (c == '"') || (c == '\'') || (c == '[') || (c == ']') || (c == '+') || + (c == '_') || (c == ':') || (c == '@') || (c == '.')); +} + +bool is_valid_arm64(std::string str) { + return std::find_if(str.begin(), str.end(), is_not_alnum_space) == str.end(); +} +} // namespace Detail::algorithm + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Check for line (syntax check) + +///////////////////////////////////////////////////////////////////////////////////////// + +std::string CompilerKit::EncoderARM64::CheckLine(std::string line, std::string file) { + std::string err_str; + + if (line.empty() || CompilerKit::find_word(line, "extern_segment") || + CompilerKit::find_word(line, "public_segment") || line.find('#') != std::string::npos || + CompilerKit::find_word(line, ";")) { + if (line.find('#') != std::string::npos) { + line.erase(line.find('#')); + } else if (line.find(';') != std::string::npos) { + line.erase(line.find(';')); + } else { + /// does the line contains valid input? + if (!Detail::algorithm::is_valid_arm64(line)) { + err_str = "Line contains non alphanumeric characters.\nhere -> "; + err_str += line; + } + } + + return err_str; + } + + if (!Detail::algorithm::is_valid_arm64(line)) { + err_str = "Line contains non alphanumeric characters.\nhere -> "; + err_str += line; + + return err_str; + } + + // check for a valid instruction format. + + if (line.find(',') != std::string::npos) { + if (line.find(',') + 1 == line.size()) { + err_str += "\nInstruction lacks right register, here -> "; + err_str += line.substr(line.find(',')); + + return err_str; + } else { + bool nothing_on_right = true; + + if (line.find(',') + 1 > line.size()) { + err_str += "\nInstruction not complete, here -> "; + err_str += line; + + return err_str; + } + + auto substr = line.substr(line.find(',') + 1); + + for (auto& ch : substr) { + if (ch != ' ' && ch != '\t') { + nothing_on_right = false; + } + } + + // this means we found nothing after that ',' . + if (nothing_on_right) { + err_str += "\nInstruction not complete, here -> "; + err_str += line; + + return err_str; + } + } + } + + return err_str; +} + +bool CompilerKit::EncoderARM64::WriteNumber(const std::size_t& pos, std::string& jump_label) { + if (!isdigit(jump_label[pos])) return false; + + switch (jump_label[pos + 1]) { + case 'x': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16); !res) { + if (errno != 0) { + Detail::print_error("invalid hex number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_hex"); + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16)); + + for (char& i : num.number) { + kBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerARM64: found a base 16 number here: " << jump_label.substr(pos) + << "\n"; + } + + return true; + } + case 'b': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2); !res) { + if (errno != 0) { + Detail::print_error("invalid binary number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_bin"); + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2)); + + if (kVerbose) { + kStdOut << "AssemblerARM64: found a base 2 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + kBytes.push_back(i); + } + + return true; + } + case 'o': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7); !res) { + if (errno != 0) { + Detail::print_error("invalid octal number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_octal"); + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7)); + + if (kVerbose) { + kStdOut << "AssemblerARM64: found a base 8 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + kBytes.push_back(i); + } + + return true; + } + default: { + break; + } + } + + /* check for errno and stuff like that */ + if (auto res = strtol(jump_label.substr(pos).c_str(), nullptr, 10); !res) { + if (errno != 0) { + return false; + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos).c_str(), nullptr, 10)); + + for (char& i : num.number) { + kBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerARM64: found a base 10 number here: " << jump_label.substr(pos) << "\n"; + } + + return true; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +/// @brief Read and write an instruction to the output array. + +///////////////////////////////////////////////////////////////////////////////////////// + +bool CompilerKit::EncoderARM64::WriteLine(std::string line, std::string file) { + if (CompilerKit::find_word(line, "public_segment")) return false; + + if (!Detail::algorithm::is_valid_arm64(line)) return false; + + return true; +} + +// Last rev 13-1-24 diff --git a/dev/CompilerKit/src/Backend/AssemblerPowerPC.cc b/dev/CompilerKit/src/Backend/AssemblerPowerPC.cc new file mode 100644 index 0000000..a04a52b --- /dev/null +++ b/dev/CompilerKit/src/Backend/AssemblerPowerPC.cc @@ -0,0 +1,911 @@ +/* ------------------------------------------- + + Copyright (C) 2024-2025 Amlal EL Mahrouss, all rights reserved + +------------------------------------------- */ + +///////////////////////////////////////////////////////////////////////////////////////// + +/// @file AssemblerPower.cxx +/// @author EL Mahrouss Amlal +/// @brief POWER Assembler. + +/// REMINDER: when dealing with an undefined symbol use (string +/// size):LinkerFindSymbol:(string) so that li will look for it. + +///////////////////////////////////////////////////////////////////////////////////////// + +#ifndef __ASM_NEED_PPC__ +#define __ASM_NEED_PPC__ 1 +#endif + +#include <CompilerKit/AE.h> +#include <CompilerKit/detail/PowerPC.h> +#include <CompilerKit/ErrorID.h> +#include <CompilerKit/Frontend.h> +#include <CompilerKit/PEF.h> +#include <CompilerKit/utils/AsmUtils.h> +#include <CompilerKit/Version.h> +#include <algorithm> +#include <filesystem> +#include <fstream> +#include <iostream> +#include <vector> + +///////////////////// + +// ANSI ESCAPE CODES + +///////////////////// + +#define kBlank "\e[0;30m" +#define kRed "\e[0;31m" +#define kWhite "\e[0;97m" +#define kYellow "\e[0;33m" + +constexpr auto cPowerIPAlignment = 0x4U; + +static Char kOutputArch = CompilerKit::kPefArchPowerPC; + +static std::size_t kCounter = 1UL; + +static std::uintptr_t kOrigin = kPefBaseOrigin; +static std::vector<std::pair<std::string, std::uintptr_t>> kOriginLabel; + +static std::vector<uint8_t> kBytes; + +static CompilerKit::AERecordHeader kCurrentRecord{ + .fName = "", .fKind = CompilerKit::kPefCode, .fSize = 0, .fOffset = 0}; + +static std::vector<CompilerKit::AERecordHeader> kRecords; +static std::vector<std::string> kUndefinedSymbols; + +static const std::string kUndefinedSymbol = ":UndefinedSymbol:"; +static const std::string kRelocSymbol = ":RuntimeSymbol:"; + +// \brief forward decl. +static bool asm_read_attributes(std::string line); + +///////////////////////////////////////////////////////////////////////////////////////// + +/// @brief POWER assembler entrypoint, the program/module starts here. + +///////////////////////////////////////////////////////////////////////////////////////// + +LIBCOMPILER_MODULE(AssemblerMainPower64) { + CompilerKit::install_signal(SIGSEGV, Detail::drvi_crash_handler); + + for (size_t i = 1; i < argc; ++i) { + if (argv[i][0] == '-') { + if (strcmp(argv[i], "--ver") == 0 || strcmp(argv[i], "--v") == 0) { + kStdOut << "AssemblerPower: POWER64 Assembler Driver.\nAssemblerPower: " << kDistVersion + << "\nAssemblerPower: " + "Copyright (c) " + "Amlal El Mahrouss\n"; + return 0; + } else if (strcmp(argv[i], "--h") == 0) { + kStdOut << "AssemblerPower: POWER64 Assembler Driver.\nAssemblerPower: Copyright (c) 2024 " + "Amlal El Mahrouss\n"; + kStdOut << "--version,/v: print program version.\n"; + kStdOut << "--verbose: print verbose output.\n"; + kStdOut << "--binary: output as flat binary.\n"; + + return 0; + } else if (strcmp(argv[i], "--binary") == 0) { + kOutputAsBinary = true; + continue; + } else if (strcmp(argv[i], "--verbose") == 0) { + kVerbose = true; + continue; + } + + kStdOut << "AssemblerPower: ignore " << argv[i] << "\n"; + continue; + } + + if (!std::filesystem::exists(argv[i])) { + kStdOut << "AssemblerPower: can't open: " << argv[i] << std::endl; + goto asm_fail_exit; + } + + std::string object_output(argv[i]); + + for (auto& ext : kAsmFileExts) { + if (object_output.find(ext) != std::string::npos) { + object_output.erase(object_output.find(ext), std::strlen(ext)); + } + } + + object_output += kOutputAsBinary ? kBinaryFileExt : kObjectFileExt; + + std::ifstream file_ptr(argv[i]); + std::ofstream file_ptr_out(object_output, std::ofstream::binary); + + if (file_ptr_out.bad()) { + if (kVerbose) { + kStdOut << "AssemblerPower: error: " << strerror(errno) << "\n"; + } + } + + std::string line; + + CompilerKit::AEHeader hdr{0}; + + memset(hdr.fPad, kAENullType, kAEPad); + + hdr.fMagic[0] = kAEMag0; + hdr.fMagic[1] = kAEMag1; + hdr.fSize = sizeof(CompilerKit::AEHeader); + hdr.fArch = kOutputArch; + + ///////////////////////////////////////////////////////////////////////////////////////// + + // COMPILATION LOOP + + ///////////////////////////////////////////////////////////////////////////////////////// + + CompilerKit::EncoderPowerPC asm64; + + while (std::getline(file_ptr, line)) { + if (auto ln = asm64.CheckLine(line, argv[i]); !ln.empty()) { + Detail::print_error(ln, argv[i]); + continue; + } + + try { + asm_read_attributes(line); + asm64.WriteLine(line, argv[i]); + } catch (const std::exception& e) { + if (kVerbose) { + std::string what = e.what(); + Detail::print_warning("exit because of: " + what, "CompilerKit"); + } + + std::filesystem::remove(object_output); + goto asm_fail_exit; + } + } + + if (!kOutputAsBinary) { + if (kVerbose) { + kStdOut << "AssemblerPower: Writing object file...\n"; + } + + // this is the final step, write everything to the file. + + auto pos = file_ptr_out.tellp(); + + hdr.fCount = kRecords.size() + kUndefinedSymbols.size(); + + file_ptr_out << hdr; + + if (kRecords.empty()) { + kStdErr << "AssemblerPower: At least one record is needed to write an object " + "file.\nAssemblerPower: Make one using `public_segment .code64 foo_bar`.\n"; + + std::filesystem::remove(object_output); + return 1; + } + + kRecords[kRecords.size() - 1].fSize = kBytes.size(); + + std::size_t record_count = 0UL; + + for (auto& record_hdr : kRecords) { + record_hdr.fFlags |= CompilerKit::kKindRelocationAtRuntime; + record_hdr.fOffset = record_count; + ++record_count; + + file_ptr_out << record_hdr; + + if (kVerbose) kStdOut << "AssemblerPower: Wrote record " << record_hdr.fName << "...\n"; + } + + // increment once again, so that we won't lie about the kUndefinedSymbols. + ++record_count; + + for (auto& sym : kUndefinedSymbols) { + CompilerKit::AERecordHeader undefined_sym{0}; + + if (kVerbose) kStdOut << "AssemblerPower: Wrote symbol " << sym << " to file...\n"; + + undefined_sym.fKind = kAENullType; + undefined_sym.fSize = sym.size(); + undefined_sym.fOffset = record_count; + + ++record_count; + + memset(undefined_sym.fPad, kAENullType, kAEPad); + memcpy(undefined_sym.fName, sym.c_str(), sym.size()); + + file_ptr_out << undefined_sym; + + ++kCounter; + } + + auto pos_end = file_ptr_out.tellp(); + + file_ptr_out.seekp(pos); + + hdr.fStartCode = pos_end; + hdr.fCodeSize = kBytes.size(); + + file_ptr_out << hdr; + + file_ptr_out.seekp(pos_end); + } else { + if (kVerbose) { + kStdOut << "AssemblerPower: Write raw binary...\n"; + } + } + + // byte from byte, we write this. + for (auto& byte : kBytes) { + file_ptr_out.write(reinterpret_cast<const char*>(&byte), sizeof(byte)); + } + + if (kVerbose) kStdOut << "AssemblerPower: Wrote file with program in it.\n"; + + file_ptr_out.flush(); + file_ptr_out.close(); + + if (kVerbose) kStdOut << "AssemblerPower: Exit succeeded.\n"; + + return 0; + } + +asm_fail_exit: + + if (kVerbose) kStdOut << "AssemblerPower: Exit failed.\n"; + + return LIBCOMPILER_EXEC_ERROR; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Check for attributes +// returns true if any was found. + +///////////////////////////////////////////////////////////////////////////////////////// + +static bool asm_read_attributes(std::string line) { + // extern_segment is the opposite of public_segment, it signals to the li + // that we need this symbol. + if (CompilerKit::find_word(line, "extern_segment")) { + if (kOutputAsBinary) { + Detail::print_error("Invalid extern_segment directive in flat binary mode.", "CompilerKit"); + throw std::runtime_error("invalid_extern_segment_bin"); + } + + auto name = line.substr(line.find("extern_segment") + strlen("extern_segment") + 1); + + if (name.size() == 0) { + Detail::print_error("Invalid extern_segment", "CompilerKit"); + throw std::runtime_error("invalid_extern_segment"); + } + + std::string result = std::to_string(name.size()); + result += kUndefinedSymbol; + + // mangle this + for (char& j : name) { + if (j == ' ' || j == ',') j = '$'; + } + + result += name; + + if (name.find(".code64") != std::string::npos) { + // data is treated as code. + kCurrentRecord.fKind = CompilerKit::kPefCode; + } else if (name.find(".data64") != std::string::npos) { + // no code will be executed from here. + kCurrentRecord.fKind = CompilerKit::kPefData; + } else if (name.find(".zero64") != std::string::npos) { + // this is a bss section. + kCurrentRecord.fKind = CompilerKit::kPefZero; + } + + // this is a special case for the start stub. + // we want this so that li can find it. + + if (name == kPefStart) { + kCurrentRecord.fKind = CompilerKit::kPefCode; + } + + // now we can tell the code size of the previous kCurrentRecord. + + if (!kRecords.empty()) kRecords[kRecords.size() - 1].fSize = kBytes.size(); + + memset(kCurrentRecord.fName, 0, kAESymbolLen); + memcpy(kCurrentRecord.fName, result.c_str(), result.size()); + + ++kCounter; + + memset(kCurrentRecord.fPad, kAENullType, kAEPad); + + kRecords.emplace_back(kCurrentRecord); + + return true; + } + // public_segment is a special keyword used by AssemblerPower to tell the AE output stage to + // mark this section as a header. it currently supports .code64, .data64., + // .zero64 + else if (CompilerKit::find_word(line, "public_segment")) { + if (kOutputAsBinary) { + Detail::print_error("Invalid public_segment directive in flat binary mode.", "CompilerKit"); + throw std::runtime_error("invalid_public_segment_bin"); + } + + auto name = line.substr(line.find("public_segment") + strlen("public_segment")); + + std::string name_copy = name; + + for (char& j : name) { + if (j == ' ') j = '$'; + } + + if (name.find(".code64") != std::string::npos) { + // data is treated as code. + + name_copy.erase(name_copy.find(".code64"), strlen(".code64")); + kCurrentRecord.fKind = CompilerKit::kPefCode; + } else if (name.find(".data64") != std::string::npos) { + // no code will be executed from here. + + name_copy.erase(name_copy.find(".data64"), strlen(".data64")); + kCurrentRecord.fKind = CompilerKit::kPefData; + } else if (name.find(".zero64") != std::string::npos) { + // this is a bss section. + + name_copy.erase(name_copy.find(".zero64"), strlen(".zero64")); + kCurrentRecord.fKind = CompilerKit::kPefZero; + } + + // this is a special case for the start stub. + // we want this so that li can find it. + + if (name == kPefStart) { + kCurrentRecord.fKind = CompilerKit::kPefCode; + } + + while (name_copy.find(" ") != std::string::npos) name_copy.erase(name_copy.find(" "), 1); + + kOriginLabel.push_back(std::make_pair(name_copy, kOrigin)); + ++kOrigin; + + // now we can tell the code size of the previous kCurrentRecord. + + if (!kRecords.empty()) kRecords[kRecords.size() - 1].fSize = kBytes.size(); + + memset(kCurrentRecord.fName, 0, kAESymbolLen); + memcpy(kCurrentRecord.fName, name.c_str(), name.size()); + + ++kCounter; + + memset(kCurrentRecord.fPad, kAENullType, kAEPad); + + kRecords.emplace_back(kCurrentRecord); + + return true; + } + + return false; +} + +// \brief algorithms and helpers. + +namespace Detail::algorithm { +// \brief authorize a brief set of characters. +static inline bool is_not_alnum_space(char c) { + return !(isalpha(c) || isdigit(c) || (c == ' ') || (c == '\t') || (c == ',') || (c == '(') || + (c == ')') || (c == '"') || (c == '\'') || (c == '[') || (c == ']') || (c == '+') || + (c == '_') || (c == ':') || (c == '@') || (c == '.')); +} + +bool is_valid_power64(std::string str) { + return std::find_if(str.begin(), str.end(), is_not_alnum_space) == str.end(); +} +} // namespace Detail::algorithm + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Check for line (syntax check) + +///////////////////////////////////////////////////////////////////////////////////////// + +std::string CompilerKit::EncoderPowerPC::CheckLine(std::string line, std::string file) { + std::string err_str; + + if (line.empty() || CompilerKit::find_word(line, "extern_segment") || + CompilerKit::find_word(line, "public_segment") || line.find('#') != std::string::npos || + CompilerKit::find_word(line, ";")) { + if (line.find('#') != std::string::npos) { + line.erase(line.find('#')); + } else if (line.find(';') != std::string::npos) { + line.erase(line.find(';')); + } else { + /// does the line contains valid input? + if (!Detail::algorithm::is_valid_power64(line)) { + err_str = "Line contains non alphanumeric characters.\nhere -> "; + err_str += line; + } + } + + return err_str; + } + + if (!Detail::algorithm::is_valid_power64(line)) { + err_str = "Line contains non alphanumeric characters.\nhere -> "; + err_str += line; + + return err_str; + } + + // check for a valid instruction format. + + if (line.find(',') != std::string::npos) { + if (line.find(',') + 1 == line.size()) { + err_str += "\nInstruction lacks right register, here -> "; + err_str += line.substr(line.find(',')); + + return err_str; + } else { + bool nothing_on_right = true; + + if (line.find(',') + 1 > line.size()) { + err_str += "\nInstruction not complete, here -> "; + err_str += line; + + return err_str; + } + + auto substr = line.substr(line.find(',') + 1); + + for (auto& ch : substr) { + if (ch != ' ' && ch != '\t') { + nothing_on_right = false; + } + } + + // this means we found nothing after that ',' . + if (nothing_on_right) { + err_str += "\nInstruction not complete, here -> "; + err_str += line; + + return err_str; + } + } + } + + // these do take an argument. + std::vector<std::string> operands_inst = {"stw", "li"}; + + // these don't. + std::vector<std::string> filter_inst = {"blr", "bl", "sc"}; + + for (auto& opcode_risc : kOpcodesPowerPC) { + if (CompilerKit::find_word(line, opcode_risc.name)) { + for (auto& op : operands_inst) { + // if only the instruction was found. + if (line == op) { + err_str += "\nMalformed "; + err_str += op; + err_str += " instruction, here -> "; + err_str += line; + } + } + + // if it is like that -> addr1, 0x0 + if (auto it = std::find(filter_inst.begin(), filter_inst.end(), opcode_risc.name); + it == filter_inst.cend()) { + if (CompilerKit::find_word(line, opcode_risc.name)) { + if (!isspace(line[line.find(opcode_risc.name) + strlen(opcode_risc.name)])) { + err_str += "\nMissing space between "; + err_str += opcode_risc.name; + err_str += " and operands.\nhere -> "; + err_str += line; + } + } + } + + return err_str; + } + } + + err_str += "Unrecognized instruction: " + line; + + return err_str; +} + +bool CompilerKit::EncoderPowerPC::WriteNumber(const std::size_t& pos, std::string& jump_label) { + if (!isdigit(jump_label[pos])) return false; + + switch (jump_label[pos + 1]) { + case 'x': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16); !res) { + if (errno != 0) { + Detail::print_error("invalid hex number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_hex"); + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 16)); + + for (char& i : num.number) { + kBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerPower: found a base 16 number here: " << jump_label.substr(pos) + << "\n"; + } + + return true; + } + case 'b': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2); !res) { + if (errno != 0) { + Detail::print_error("invalid binary number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_bin"); + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 2)); + + if (kVerbose) { + kStdOut << "AssemblerPower: found a base 2 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + kBytes.push_back(i); + } + + return true; + } + case 'o': { + if (auto res = strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7); !res) { + if (errno != 0) { + Detail::print_error("invalid octal number: " + jump_label, "CompilerKit"); + throw std::runtime_error("invalid_octal"); + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos + 2).c_str(), nullptr, 7)); + + if (kVerbose) { + kStdOut << "AssemblerPower: found a base 8 number here: " << jump_label.substr(pos) << "\n"; + } + + for (char& i : num.number) { + kBytes.push_back(i); + } + + return true; + } + default: { + break; + } + } + + /* check for errno and stuff like that */ + if (auto res = strtol(jump_label.substr(pos).c_str(), nullptr, 10); !res) { + if (errno != 0) { + return false; + } + } + + CompilerKit::NumberCast64 num(strtol(jump_label.substr(pos).c_str(), nullptr, 10)); + + for (char& i : num.number) { + kBytes.push_back(i); + } + + if (kVerbose) { + kStdOut << "AssemblerPower: found a base 10 number here: " << jump_label.substr(pos) << "\n"; + } + + return true; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +/// @brief Read and write an instruction to the output array. + +///////////////////////////////////////////////////////////////////////////////////////// + +bool CompilerKit::EncoderPowerPC::WriteLine(std::string line, std::string file) { + if (CompilerKit::find_word(line, "public_segment")) return false; + if (!Detail::algorithm::is_valid_power64(line)) return false; + + for (auto& opcode_risc : kOpcodesPowerPC) { + // strict check here + if (CompilerKit::find_word(line, opcode_risc.name)) { + std::string name(opcode_risc.name); + std::string jump_label, cpy_jump_label; + std::vector<size_t> found_registers_index; + + // check funct7 type. + switch (opcode_risc.ops->type) { + default: { + NumberCast32 num(opcode_risc.opcode); + + for (auto ch : num.number) { + kBytes.emplace_back(ch); + } + break; + } + case BADDR: + case PCREL: { + auto num = GetNumber32(line, name); + + kBytes.emplace_back(num.number[0]); + kBytes.emplace_back(num.number[1]); + kBytes.emplace_back(num.number[2]); + kBytes.emplace_back(0x48); + + break; + } + /// General purpose, float, vector operations. Everything that involve + /// registers. + case G0REG: + case FREG: + case VREG: + case GREG: { + // \brief how many registers we found. + std::size_t found_some_count = 0UL; + std::size_t register_count = 0UL; + std::string opcodeName = opcode_risc.name; + std::size_t register_sum = 0; + + NumberCast64 num(opcode_risc.opcode); + + for (size_t line_index = 0UL; line_index < line.size(); line_index++) { + if (line[line_index] == kAsmRegisterPrefix[0] && isdigit(line[line_index + 1])) { + std::string register_syntax = kAsmRegisterPrefix; + register_syntax += line[line_index + 1]; + + if (isdigit(line[line_index + 2])) register_syntax += line[line_index + 2]; + + std::string reg_str; + reg_str += line[line_index + 1]; + + if (isdigit(line[line_index + 2])) reg_str += line[line_index + 2]; + + // it ranges from r0 to r19 + // something like r190 doesn't exist in the instruction set. + if (isdigit(line[line_index + 3]) && isdigit(line[line_index + 2])) { + reg_str += line[line_index + 3]; + Detail::print_error("invalid register index, r" + reg_str + + "\nnote: The POWER accepts registers from r0 to r32.", + file); + throw std::runtime_error("invalid_register_index"); + } + + // finally cast to a size_t + std::size_t reg_index = strtol(reg_str.c_str(), nullptr, 10); + + if (reg_index > kAsmRegisterLimit) { + Detail::print_error("invalid register index, r" + reg_str, file); + throw std::runtime_error("invalid_register_index"); + } + + if (opcodeName == "li") { + char numIndex = 0; + + for (size_t i = 0; i != reg_index; i++) { + numIndex += 0x20; + } + + auto num = GetNumber32(line, reg_str); + + kBytes.push_back(num.number[0]); + kBytes.push_back(num.number[1]); + kBytes.push_back(numIndex); + kBytes.push_back(0x38); + + // check if bigger than two. + for (size_t i = 2; i < 4; i++) { + if (num.number[i] > 0) { + Detail::print_warning("number overflow on li operation.", file); + break; + } + } + + break; + } + + if ((opcodeName[0] == 's' && opcodeName[1] == 't')) { + if (register_sum == 0) { + for (size_t indexReg = 0UL; indexReg < reg_index; ++indexReg) { + register_sum += 0x20; + } + } else { + register_sum += reg_index; + } + } + + if (opcodeName == "mr") { + switch (register_count) { + case 0: { + kBytes.push_back(0x78); + + char numIndex = 0x3; + + for (size_t i = 0; i != reg_index; i++) { + numIndex += 0x8; + } + + kBytes.push_back(numIndex); + + break; + } + case 1: { + char numIndex = 0x1; + + for (size_t i = 0; i != reg_index; i++) { + numIndex += 0x20; + } + + for (size_t i = 0; i != reg_index; i++) { + kBytes[kBytes.size() - 1] += 0x8; + } + + kBytes[kBytes.size() - 1] -= 0x8; + + kBytes.push_back(numIndex); + + if (reg_index >= 10 && reg_index < 20) + kBytes.push_back(0x7d); + else if (reg_index >= 20 && reg_index < 30) + kBytes.push_back(0x7e); + else if (reg_index >= 30) + kBytes.push_back(0x7f); + else + kBytes.push_back(0x7c); + + break; + } + default: + break; + } + + ++register_count; + ++found_some_count; + } + + if (opcodeName == "addi") { + if (found_some_count == 2 || found_some_count == 0) + kBytes.emplace_back(reg_index); + else if (found_some_count == 1) + kBytes.emplace_back(0x00); + + ++found_some_count; + + if (found_some_count > 3) { + Detail::print_error("Too much registers. -> " + line, file); + throw std::runtime_error("too_much_regs"); + } + } + + if (opcodeName.find("cmp") != std::string::npos) { + ++found_some_count; + + if (found_some_count > 3) { + Detail::print_error("Too much registers. -> " + line, file); + throw std::runtime_error("too_much_regs"); + } + } + + if (opcodeName.find("mf") != std::string::npos || + opcodeName.find("mt") != std::string::npos) { + char numIndex = 0; + + for (size_t i = 0; i != reg_index; i++) { + numIndex += 0x20; + } + + num.number[2] += numIndex; + + ++found_some_count; + + if (found_some_count > 1) { + Detail::print_error("Too much registers. -> " + line, file); + throw std::runtime_error("too_much_regs"); + } + + if (kVerbose) { + kStdOut << "AssemblerPower: Found register: " << register_syntax << "\n"; + kStdOut << "AssemblerPower: Amount of registers in instruction: " + << found_some_count << "\n"; + } + + if (reg_index >= 10 && reg_index < 20) + num.number[3] = 0x7d; + else if (reg_index >= 20 && reg_index < 30) + num.number[3] = 0x7e; + else if (reg_index >= 30) + num.number[3] = 0x7f; + else + num.number[3] = 0x7c; + + for (auto ch : num.number) { + kBytes.emplace_back(ch); + } + } + + found_registers_index.push_back(reg_index); + } + } + + if (opcodeName == "addi") { + kBytes.emplace_back(0x38); + } + + if (opcodeName.find("cmp") != std::string::npos) { + char rightReg = 0x0; + + for (size_t i = 0; i != found_registers_index[1]; i++) { + rightReg += 0x08; + } + + kBytes.emplace_back(0x00); + kBytes.emplace_back(rightReg); + kBytes.emplace_back(found_registers_index[0]); + kBytes.emplace_back(0x7c); + } + + if ((opcodeName[0] == 's' && opcodeName[1] == 't')) { + size_t offset = 0UL; + + if (line.find('+') != std::string::npos) { + auto number = GetNumber32(line.substr(line.find("+")), "+"); + offset = number.raw; + } + + kBytes.push_back(offset); + kBytes.push_back(0x00); + kBytes.push_back(register_sum); + + kBytes.emplace_back(0x90); + } + + if (opcodeName == "mr") { + if (register_count == 1) { + Detail::print_error("Too few registers. -> " + line, file); + throw std::runtime_error("too_few_registers"); + } + } + + // we're not in immediate addressing, reg to reg. + if (opcode_risc.ops->type != GREG) { + // remember! register to register! + if (found_some_count == 1) { + Detail::print_error( + "Unrecognized register found.\ntip: each AssemblerPower register " + "starts with 'r'.\nline: " + + line, + file); + + throw std::runtime_error("not_a_register"); + } + } + + if (found_some_count < 1 && name[0] != 'l' && name[0] != 's') { + Detail::print_error("invalid combination of opcode and registers.\nline: " + line, + file); + throw std::runtime_error("invalid_comb_op_reg"); + } + + break; + } + } + + kOrigin += cPowerIPAlignment; + break; + } + } + + return true; +} + +// Last rev 13-1-24 diff --git a/dev/CompilerKit/src/BasicString.cc b/dev/CompilerKit/src/BasicString.cc new file mode 100644 index 0000000..ca257c0 --- /dev/null +++ b/dev/CompilerKit/src/BasicString.cc @@ -0,0 +1,207 @@ +/* + * ======================================================== + * + * CompilerKit + * Copyright (C) 2024-2025 Amlal El Mahrouss, all rights reserved. + * + * ======================================================== + */ + +/** + * @file BasicString.cc + * @author Amlal (amlal@el-mahrouss-logic.com) + * @brief C++ string manipulation API. + * @version 0.2 + * @date 2024-01-23 + * + * @copyright Copyright (c) Amlal El Mahrouss + * + */ + +#include <CompilerKit/BasicString.h> + +namespace CompilerKit { +Char* BasicString::Data() { + return m_Data; +} + +const Char* BasicString::CData() const { + return m_Data; +} + +SizeType BasicString::Length() const { + return strlen(m_Data); +} + +bool BasicString::operator==(const BasicString& rhs) const { + if (rhs.Length() != Length()) return false; + + for (SizeType index = 0; index < Length(); ++index) { + if (rhs.m_Data[index] != m_Data[index]) return false; + } + + return true; +} + +bool BasicString::operator==(const Char* rhs) const { + if (string_length(rhs) != Length()) return false; + + for (SizeType index = 0; index < string_length(rhs); ++index) { + if (rhs[index] != m_Data[index]) return false; + } + + return true; +} + +bool BasicString::operator!=(const BasicString& rhs) const { + if (rhs.Length() != Length()) return false; + + for (SizeType index = 0; index < rhs.Length(); ++index) { + if (rhs.m_Data[index] == m_Data[index]) return false; + } + + return true; +} + +bool BasicString::operator!=(const Char* rhs) const { + if (string_length(rhs) != Length()) return false; + + for (SizeType index = 0; index < string_length(rhs); ++index) { + if (rhs[index] == m_Data[index]) return false; + } + + return true; +} + +BasicString StringBuilder::Construct(const Char* data) { + if (!data || *data == 0) return BasicString(0); + + BasicString view(strlen(data)); + view += data; + + return view; +} + +const char* StringBuilder::FromInt(const char* fmt, int i) { + if (!fmt) return ("-1"); + + auto ret_len = 8 + string_length(fmt); + char* ret = new char[ret_len]; + + if (!ret) return ("-1"); + + memset(ret, 0, ret_len); + + Char result[sizeof(int64_t)]; + + if (!to_str(result, sizeof(int64_t), i)) { + delete[] ret; + return ("-1"); + } + + const auto fmt_len = string_length(fmt); + const auto res_len = string_length(result); + + for (SizeType idx = 0; idx < fmt_len; ++idx) { + if (fmt[idx] == '%') { + SizeType result_cnt = idx; + + for (auto y_idx = 0; y_idx < res_len; ++y_idx) { + ret[y_idx] = result[result_cnt]; + ++result_cnt; + } + + break; + } + + ret[idx] = fmt[idx]; + } + + return ret; /* Copy that ret into a buffer, Alloca allocates to the stack */ +} + +const char* StringBuilder::FromBool(const char* fmt, bool i) { + if (!fmt) return ("?"); + + const char* boolean_expr = i ? "true" : "false"; + char* ret = new char[i ? 4 : 5 + string_length(fmt)]; + + if (!ret) return ("?"); + + const auto fmt_len = string_length(fmt); + const auto res_len = string_length(boolean_expr); + + for (SizeType idx = 0; idx < fmt_len; ++idx) { + if (fmt[idx] == '%') { + SizeType result_cnt = idx; + + for (auto y_idx = idx; y_idx < res_len; ++y_idx) { + ret[result_cnt] = boolean_expr[y_idx]; + ++result_cnt; + } + + break; + } + + ret[idx] = fmt[idx]; + } + + return ret; +} + +bool StringBuilder::Equals(const char* lhs, const char* rhs) { + if (string_length(rhs) != string_length(lhs)) return false; + + for (SizeType index = 0; index < string_length(rhs); ++index) { + if (rhs[index] != lhs[index]) return false; + } + + return true; +} + +const char* StringBuilder::Format(const char* fmt, const char* fmtRight) { + if (!fmt || !fmtRight) return ("?"); + + char* ret = new char[string_length(fmtRight) + string_length(fmtRight)]; + if (!ret) return ("?"); + + for (SizeType idx = 0; idx < string_length(fmt); ++idx) { + if (fmt[idx] == '%') { + SizeType result_cnt = idx; + + for (SizeType y_idx = 0; y_idx < string_length(fmtRight); ++y_idx) { + ret[result_cnt] = fmtRight[y_idx]; + ++result_cnt; + } + + break; + } + + ret[idx] = fmt[idx]; + } + + return ret; +} + +BasicString& BasicString::operator+=(const Char* rhs) { + if (strlen(rhs) > this->m_Sz) { + throw std::runtime_error("out_of_bounds: BasicString"); + } + + memcpy(this->m_Data + this->m_Cur, rhs, strlen(rhs)); + this->m_Cur += strlen(rhs); + + return *this; +} + +BasicString& BasicString::operator+=(const BasicString& rhs) { + if (rhs.m_Cur > this->m_Sz) { + throw std::runtime_error("out_of_bounds: BasicString"); + } + + memcpy(this->m_Data + this->m_Cur, rhs.CData(), strlen(rhs.CData())); + this->m_Cur += strlen(rhs.CData()); + + return *this; +} +} // namespace CompilerKit diff --git a/dev/CompilerKit/src/CodeGen.cc b/dev/CompilerKit/src/CodeGen.cc new file mode 100644 index 0000000..e59001d --- /dev/null +++ b/dev/CompilerKit/src/CodeGen.cc @@ -0,0 +1,52 @@ +/* ------------------------------------------- + + Copyright (C) 2024-2025 Amlal EL Mahrouss, all rights reserved + +------------------------------------------- */ + +#include <CompilerKit/CodeGen.h> +#include <CompilerKit/ErrorID.h> + +/** + * @file CodeGen.cc + * @author Amlal El Mahrouss (amlal@nekernel.org) + * @brief CodeGen API of NeCTI + * @version 0.0.2 + * + * @copyright Copyright (c) 2024-2025 Amlal El Mahrouss + * + */ + +namespace CompilerKit { +///! @brief Compile for specific format (ELF, PEF, ZBIN) +Int32 AssemblyFactory::Compile(STLString sourceFile, const Int32& arch) noexcept { + if (sourceFile.length() < 1) return LIBCOMPILER_UNIMPLEMENTED; + + if (!fMounted) return LIBCOMPILER_UNIMPLEMENTED; + if (arch != fMounted->Arch()) return LIBCOMPILER_INVALID_ARCH; + + try { + return this->fMounted->CompileToFormat(sourceFile, arch); + } catch (std::exception& e) { + return LIBCOMPILER_EXEC_ERROR; + } +} + +///! @brief mount assembly backend. +void AssemblyFactory::Mount(AssemblyInterface* mountPtr) noexcept { + if (mountPtr) { + fMounted = mountPtr; + } +} + +///! @brief Unmount assembler. +AssemblyInterface* AssemblyFactory::Unmount() noexcept { + auto mount_prev = fMounted; + + if (fMounted) { + fMounted = nullptr; + } + + return mount_prev; +} +} // namespace CompilerKit diff --git a/dev/CompilerKit/src/Frontend.cc b/dev/CompilerKit/src/Frontend.cc new file mode 100644 index 0000000..37b36f7 --- /dev/null +++ b/dev/CompilerKit/src/Frontend.cc @@ -0,0 +1,51 @@ +/* ------------------------------------------- + + Copyright (C) 2025 Amlal EL Mahrouss, all rights reserved + +------------------------------------------- */ + +#include <CompilerKit/Frontend.h> + +namespace CompilerKit { +/// find the perfect matching word in a haystack. +/// \param haystack base string +/// \param needle the string we search for. +/// \return if we found it or not. +BOOL find_word(STLString haystack, STLString needle) noexcept { + auto index = haystack.find(needle); + + // check for needle validity. + if (index == STLString::npos) return false; + + // declare lambda + auto not_part_of_word = [&](int index) { + if (std::isspace(haystack[index]) || std::ispunct(haystack[index])) return true; + + if (index <= 0 || index >= haystack.size()) return true; + + return false; + }; + + return not_part_of_word(index - 1) && not_part_of_word(index + needle.size()); +} + +/// find a word within strict conditions and returns a range of it. +/// \param haystack +/// \param needle +/// \return position of needle. +SizeType find_word_range(STLString haystack, STLString needle) noexcept { + auto index = haystack.find(needle); + + // check for needle validity. + if (index == STLString::npos) return false; + + if (!isalnum((haystack[index + needle.size() + 1])) && + !isdigit(haystack[index + needle.size() + 1]) && + !isalnum((haystack[index - needle.size() - 1])) && + !isdigit(haystack[index - needle.size() - 1])) { + return index; + } + + return STLString::npos; +} +} // namespace CompilerKit
\ No newline at end of file diff --git a/dev/CompilerKit/src/Frontend/CCompiler64x0.cc b/dev/CompilerKit/src/Frontend/CCompiler64x0.cc new file mode 100644 index 0000000..ae2939b --- /dev/null +++ b/dev/CompilerKit/src/Frontend/CCompiler64x0.cc @@ -0,0 +1,1286 @@ +/* + * ======================================================== + * + * cc + * Copyright (C) 2024-2025 Amlal El Mahrouss, all rights reserved. + * + * ======================================================== + */ + +/// BUGS: 0 +/// TODO: none + +#include <CompilerKit/detail/64x0.h> +#include <CompilerKit/Frontend.h> +#include <CompilerKit/UUID.h> +#include <CompilerKit/utils/CompilerUtils.h> +#include <cstdio> +#include <fstream> +#include <iostream> +#include <memory> +#include <random> +#include <string> +#include <utility> +#include <vector> + +/* C driver */ +/* This is part of the CompilerKit. */ +/* (c) Amlal El Mahrouss */ + +/// @author EL Mahrouss Amlal (amlel) +/// @file 64x0-cc.cxx +/// @brief 64x0 C Compiler. + +/// TODO: support structures, else if, else, . and -> + +///////////////////// + +// ANSI ESCAPE CODES + +///////////////////// + +#define kExitOK (0) + +#define kBlank "\e[0;30m" +#define kRed "\e[0;31m" +#define kWhite "\e[0;97m" + +///////////////////////////////////// + +// INTERNAL STUFF OF THE C COMPILER + +///////////////////////////////////// + +namespace Detail { +// \brief Register map structure, used to keep track of each variable's registers. +struct CompilerRegisterMap final { + std::string fName; + std::string fReg; +}; + +// \brief Map for C structs +// \author amlel +struct CompilerStructMap final { + // 'my_foo' + std::string fName; + + // if instance: stores a valid register. + std::string fReg; + + // offset count + std::size_t fOffsetsCnt; + + // offset array. + std::vector<std::pair<Int32, std::string>> fOffsets; +}; + +struct CompilerState final { + std::vector<CompilerKit::SyntaxLeafList> fSyntaxTreeList; + std::vector<CompilerRegisterMap> kStackFrame; + std::vector<CompilerStructMap> kStructMap; + CompilerKit::SyntaxLeafList* fSyntaxTree{nullptr}; + std::unique_ptr<std::ofstream> fOutputAssembly; + std::string fLastFile; + std::string fLastError; + bool fVerbose; +}; +} // namespace Detail + +static Detail::CompilerState kState; +static std::string kIfFunction = ""; + +namespace Detail { +/// @brief prints an error into stdout. +/// @param reason the reason of the error. +/// @param file where does it originate from? +void print_error(std::string reason, std::string file) noexcept; + +struct CompilerType final { + std::string fName; + std::string fValue; +}; +} // namespace Detail + +///////////////////////////////////////////////////////////////////////////////////////// + +// Target architecture. +static int kMachine = 0; + +///////////////////////////////////////// + +// REGISTERS ACCORDING TO USED ASSEMBLER + +///////////////////////////////////////// + +static size_t kRegisterCnt = kAsmRegisterLimit; +static size_t kStartUsable = 2; +static size_t kUsableLimit = 15; +static size_t kRegisterCounter = kStartUsable; +static std::string kRegisterPrefix = kAsmRegisterPrefix; + +///////////////////////////////////////// + +// COMPILER PARSING UTILITIES/STATES. + +///////////////////////////////////////// + +static std::vector<std::string> kFileList; +static CompilerKit::AssemblyFactory kFactory; +static bool kInStruct = false; +static bool kOnWhileLoop = false; +static bool kOnForLoop = false; +static bool kInBraces = false; +static bool kIfFound = false; +static size_t kBracesCount = 0UL; + +/* @brief C compiler backend for C */ +class CompilerFrontend64x0 final : public CompilerKit::CompilerFrontendInterface { + public: + explicit CompilerFrontend64x0() = default; + ~CompilerFrontend64x0() override = default; + + LIBCOMPILER_COPY_DEFAULT(CompilerFrontend64x0); + + std::string Check(const char* text, const char* file); + CompilerKit::SyntaxLeafList::SyntaxLeaf Compile(std::string text, std::string file) override; + + const char* Language() override { return "64k C"; } +}; + +static CompilerFrontend64x0* kCompilerFrontend = nullptr; +static std::vector<Detail::CompilerType> kCompilerVariables; +static std::vector<std::string> kCompilerFunctions; +static std::vector<Detail::CompilerType> kCompilerTypes; + +namespace Detail { +union number_cast final { + public: + number_cast(UInt64 _Raw) : _Raw(_Raw) {} + + public: + char _Num[8]; + UInt64 _Raw; +}; + +union double_cast final { + public: + double_cast(float _Raw) : _Raw(_Raw) {} + + public: + char _Sign; + char _Lh[8]; + char _Rh[23]; + + float _Raw; +}; +} // namespace Detail + +///////////////////////////////////////////////////////////////////////////////////////// + +// @name Compile +// @brief Generate MASM from a C assignement. + +///////////////////////////////////////////////////////////////////////////////////////// + +CompilerKit::SyntaxLeafList::SyntaxLeaf CompilerFrontend64x0::Compile(std::string text_, std::string file) { + std::string text = text_; + + bool typeFound = false; + bool fnFound = false; + + // setup generator. + std::random_device rd; + + auto seed_data = std::array<int, std::mt19937::state_size>{}; + std::generate(std::begin(seed_data), std::end(seed_data), std::ref(rd)); + std::seed_seq seq(std::begin(seed_data), std::end(seed_data)); + std::mt19937 generator(seq); + + // start parsing + for (size_t text_index = 0; text_index < text.size(); ++text_index) { + auto syntaxLeaf = CompilerKit::SyntaxLeafList::SyntaxLeaf(); + + auto gen = uuids::uuid_random_generator{generator}; + uuids::uuid out = gen(); + + Detail::number_cast time_off = (UInt64) out.as_bytes().data(); + + if (!typeFound) { + auto substr = text.substr(text_index); + std::string match_type; + + for (size_t y = 0; y < substr.size(); ++y) { + if (substr[y] == ' ') { + while (match_type.find(' ') != std::string::npos) { + match_type.erase(match_type.find(' ')); + } + + for (auto& clType : kCompilerTypes) { + if (clType.fName == match_type) { + match_type.clear(); + + std::string buf; + + buf += clType.fValue; + buf += ' '; + + if (substr.find('=') != std::string::npos) { + break; + } + + if (text.find('(') != std::string::npos) { + syntaxLeaf.fUserValue = buf; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + } + + typeFound = true; + break; + } + } + + break; + } + + match_type += substr[y]; + } + } + + if (text[text_index] == '{') { + if (kInStruct) { + continue; + } + + kInBraces = true; + ++kBracesCount; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + } + + // return keyword handler + if (text[text_index] == 'r') { + std::string return_keyword; + return_keyword += "return"; + + std::size_t index = 0UL; + + std::string value; + + for (size_t return_index = text_index; return_index < text.size(); ++return_index) { + if (text[return_index] != return_keyword[index]) { + for (size_t value_index = return_index; value_index < text.size(); ++value_index) { + if (text[value_index] == ';') break; + + value += text[value_index]; + } + + break; + } + + ++index; + } + + if (index == return_keyword.size()) { + if (!value.empty()) { + if (value.find('(') != std::string::npos) { + value.erase(value.find('(')); + } + + if (!isdigit(value[value.find('(') + 2])) { + std::string tmp = value; + bool reg_to_reg = false; + + value.clear(); + + value += " extern_segment"; + value += tmp; + } + + syntaxLeaf.fUserValue = "\tldw r19, "; + + // make it pretty. + if (value.find('\t') != std::string::npos) value.erase(value.find('\t'), 1); + + syntaxLeaf.fUserValue += value + "\n"; + } + + syntaxLeaf.fUserValue += "\tjlr"; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + break; + } + } + + if (text[text_index] == 'i' && text[text_index + 1] == 'f') { + auto expr = text.substr(text_index + 2); + text.erase(text_index, 2); + + if (expr.find("{") != std::string::npos) { + expr.erase(expr.find("{")); + } + + if (expr.find("(") != std::string::npos) expr.erase(expr.find("(")); + + if (expr.find(")") != std::string::npos) expr.erase(expr.find(")")); + + kIfFunction = "__LIBCOMPILER_IF_PROC_"; + kIfFunction += std::to_string(time_off._Raw); + + syntaxLeaf.fUserValue = "\tlda r12, extern_segment "; + syntaxLeaf.fUserValue += kIfFunction + + "\n\t#r12 = Code to jump on, r11 right cond, r10 left cond.\n\tbeq " + "r10, r11, r12\ndword public_segment .code64 " + + kIfFunction + "\n"; + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + kIfFound = true; + } + + // Parse expressions and instructions here. + // what does this mean? + // we encounter an assignment, or we reached the end of an expression. + if (text[text_index] == '=' || text[text_index] == ';') { + if (fnFound) continue; + if (kIfFound) continue; + + if (text[text_index] == ';' && kInStruct) continue; + + if (text.find("typedef ") != std::string::npos) continue; + + if (text[text_index] == '=' && kInStruct) { + Detail::print_error("assignement of value in struct " + text, file); + continue; + } + + if (text[text_index] == ';' && kInStruct) { + bool space_found_ = false; + std::string sym; + + for (auto& ch : text) { + if (ch == ' ') { + space_found_ = true; + } + + if (ch == ';') break; + + if (space_found_) sym.push_back(ch); + } + + kState.kStructMap[kState.kStructMap.size() - 1].fOffsets.push_back( + std::make_pair(kState.kStructMap[kState.kStructMap.size() - 1].fOffsetsCnt + 4, sym)); + + kState.kStructMap[kState.kStructMap.size() - 1].fOffsetsCnt = + kState.kStructMap[kState.kStructMap.size() - 1].fOffsetsCnt + 4; + + continue; + } + + if (text[text_index] == '=' && kInStruct) { + continue; + } + + if (text[text_index + 1] == '=' || text[text_index - 1] == '!' || + text[text_index - 1] == '<' || text[text_index - 1] == '>') { + continue; + } + + std::string substr; + + if (text.find('=') != std::string::npos && kInBraces && !kIfFound) { + if (text.find("*") != std::string::npos) { + if (text.find("=") > text.find("*")) + substr += "\tlda "; + else + substr += "\tldw "; + } else { + substr += "\tldw "; + } + } else if (text.find('=') != std::string::npos && !kInBraces) { + substr += "stw public_segment .data64 "; + } + + int first_encountered = 0; + + std::string str_name; + + for (size_t text_index_2 = 0; text_index_2 < text.size(); ++text_index_2) { + if (text[text_index_2] == '\"') { + ++text_index_2; + + // want to add this, so that the parser recognizes that this is a + // string. + substr += '"'; + + for (; text_index_2 < text.size(); ++text_index_2) { + if (text[text_index_2] == '\"') break; + + substr += text[text_index_2]; + } + } + + if (text[text_index_2] == '{' || text[text_index_2] == '}') continue; + + if (text[text_index_2] == ';') { + break; + } + + if (text[text_index_2] == ' ' || text[text_index_2] == '\t') { + if (first_encountered != 2) { + if (text[text_index] != '=' && + substr.find("public_segment .data64") == std::string::npos && !kInStruct) + substr += "public_segment .data64 "; + } + + ++first_encountered; + + continue; + } + + if (text[text_index_2] == '=') { + if (!kInBraces) { + substr.replace(substr.find("public_segment .data64"), strlen("public_segment .data64"), + "public_segment .zero64 "); + } + + substr += ","; + continue; + } + + substr += text[text_index_2]; + } + + for (auto& clType : kCompilerTypes) { + if (substr.find(clType.fName) != std::string::npos) { + if (substr.find(clType.fName) > substr.find('"')) continue; + + substr.erase(substr.find(clType.fName), clType.fName.size()); + } else if (substr.find(clType.fValue) != std::string::npos) { + if (substr.find(clType.fValue) > substr.find('"')) continue; + + if (clType.fName == "const") continue; + + substr.erase(substr.find(clType.fValue), clType.fValue.size()); + } + } + + if (substr.find("extern") != std::string::npos) { + substr.replace(substr.find("extern"), strlen("extern"), "extern_segment "); + + if (substr.find("public_segment .data64") != std::string::npos) + substr.erase(substr.find("public_segment .data64"), strlen("public_segment .data64")); + } + + auto var_to_find = std::find_if( + kCompilerVariables.cbegin(), kCompilerVariables.cend(), + [&](Detail::CompilerType type) { return type.fName.find(substr) != std::string::npos; }); + + if (kRegisterCounter == 5 || kRegisterCounter == 6) ++kRegisterCounter; + + std::string reg = kAsmRegisterPrefix; + reg += std::to_string(kRegisterCounter); + + if (var_to_find == kCompilerVariables.cend()) { + ++kRegisterCounter; + + kState.kStackFrame.push_back({.fName = substr, .fReg = reg}); + kCompilerVariables.push_back({.fName = substr}); + } + + syntaxLeaf.fUserValue += substr; + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + if (text[text_index] == '=') break; + } + + // function handler. + + if (text[text_index] == '(' && !fnFound && !kIfFound) { + std::string substr; + std::string args_buffer; + std::string args; + + bool type_crossed = false; + + for (size_t idx = text.find('(') + 1; idx < text.size(); ++idx) { + if (text[idx] == ',') continue; + + if (text[idx] == ' ') continue; + + if (text[idx] == ')') break; + } + + for (char substr_first_index : text) { + if (substr_first_index != ',') + args_buffer += substr_first_index; + else + args_buffer += '$'; + + if (substr_first_index == ';') { + args_buffer = args_buffer.erase(0, args_buffer.find('(')); + args_buffer = args_buffer.erase(args_buffer.find(';'), 1); + args_buffer = args_buffer.erase(args_buffer.find(')'), 1); + args_buffer = args_buffer.erase(args_buffer.find('('), 1); + + if (!args_buffer.empty()) args += "\tldw r6, "; + + std::string register_type; + std::size_t index = 7UL; + + while (args_buffer.find("$") != std::string::npos) { + register_type = kRegisterPrefix; + register_type += std::to_string(index); + + ++index; + + args_buffer.replace(args_buffer.find('$'), 1, "\n\tldw " + register_type + ","); + } + + args += args_buffer; + args += "\n\tlda r19, "; + } + } + + for (char _text_i : text) { + if (_text_i == '\t' || _text_i == ' ') { + if (!type_crossed) { + substr.clear(); + type_crossed = true; + } + + continue; + } + + if (_text_i == '(') break; + + substr += _text_i; + } + + if (kInBraces) { + syntaxLeaf.fUserValue = args; + syntaxLeaf.fUserValue += substr; + syntaxLeaf.fUserValue += "\n\tjrl\n"; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + fnFound = true; + } else { + syntaxLeaf.fUserValue.clear(); + + syntaxLeaf.fUserValue += "public_segment .code64 "; + + syntaxLeaf.fUserValue += substr; + syntaxLeaf.fUserValue += "\n"; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + fnFound = true; + } + + kCompilerFunctions.push_back(text); + } + + if (text[text_index] == '-' && text[text_index + 1] == '-') { + text = text.replace(text.find("--"), strlen("--"), ""); + + for (int _text_i = 0; _text_i < text.size(); ++_text_i) { + if (text[_text_i] == '\t' || text[_text_i] == ' ') text.erase(_text_i, 1); + } + + syntaxLeaf.fUserValue += "sub "; + syntaxLeaf.fUserValue += text; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + break; + } + + if (text[text_index] == '}') { + kRegisterCounter = kStartUsable; + + --kBracesCount; + + if (kBracesCount < 1) { + kInBraces = false; + kBracesCount = 0; + } + + if (kIfFound) kIfFound = false; + + if (kInStruct) kInStruct = false; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + } + + syntaxLeaf.fUserValue.clear(); + } + + auto syntaxLeaf = CompilerKit::SyntaxLeafList::SyntaxLeaf(); + syntaxLeaf.fUserValue = "\n"; + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + return syntaxLeaf; +} + +static bool kShouldHaveBraces = false; +static std::string kFnName; + +std::string CompilerFrontend64x0::Check(const char* text, const char* file) { + std::string err_str; + std::string ln = text; + + if (ln.empty()) { + return err_str; + } + + bool non_ascii_found = false; + + for (int i = 0; i < ln.size(); ++i) { + if (isalnum(ln[i])) { + non_ascii_found = true; + break; + } + } + + if (kShouldHaveBraces && ln.find('{') != std::string::npos) { + kShouldHaveBraces = false; + } + + if (!non_ascii_found) return err_str; + + size_t string_index = 1UL; + + if (ln.find('\'') != std::string::npos) { + string_index = ln.find('\'') + 1; + + for (; string_index < ln.size(); ++string_index) { + if (ln[string_index] == '\'') { + if (ln[string_index + 1] != ';') { + ln.erase(string_index, 1); + } + + return err_str; + } + } + } else if (ln.find('"') != std::string::npos) { + string_index = ln.find('"') + 1; + + for (; string_index < ln.size(); ++string_index) { + if (ln[string_index] == '"') { + if (ln[string_index + 1] != ';') { + ln.erase(string_index, 1); + } else { + break; + } + } + } + } else if (ln.find('"') == std::string::npos && ln.find('\'') == std::string::npos) { + std::vector<std::string> forbidden_words; + + forbidden_words.push_back("\\"); + forbidden_words.push_back("?"); + forbidden_words.push_back("@"); + forbidden_words.push_back("~"); + forbidden_words.push_back("::"); + forbidden_words.push_back("--*"); + forbidden_words.push_back("*/"); + + // add them to avoid stupid mistakes. + forbidden_words.push_back("namespace"); + forbidden_words.push_back("class"); + forbidden_words.push_back("extern \"C\""); + + for (auto& forbidden : forbidden_words) { + if (ln.find(forbidden) != std::string::npos) { + err_str += "\nForbidden character detected: "; + err_str += forbidden; + + return err_str; + } + } + } + + struct CompilerVariableRange final { + std::string fBegin; + std::string fEnd; + }; + + const std::vector<CompilerVariableRange> variables_list = { + {.fBegin = "static ", .fEnd = "="}, {.fBegin = "=", .fEnd = ";"}, + {.fBegin = "if(", .fEnd = "="}, {.fBegin = "if (", .fEnd = "="}, + {.fBegin = "if(", .fEnd = "<"}, {.fBegin = "if (", .fEnd = "<"}, + {.fBegin = "if(", .fEnd = ">"}, {.fBegin = "if (", .fEnd = ">"}, + {.fBegin = "if(", .fEnd = ")"}, {.fBegin = "if (", .fEnd = ")"}, + + {.fBegin = "else(", .fEnd = "="}, {.fBegin = "else (", .fEnd = "="}, + {.fBegin = "else(", .fEnd = "<"}, {.fBegin = "else (", .fEnd = "<"}, + {.fBegin = "else(", .fEnd = ">"}, {.fBegin = "else (", .fEnd = ">"}, + {.fBegin = "else(", .fEnd = ")"}, {.fBegin = "else (", .fEnd = ")"}, + }; + + for (auto& variable : variables_list) { + if (ln.find(variable.fBegin) != std::string::npos) { + string_index = ln.find(variable.fBegin) + variable.fBegin.size(); + + while (ln[string_index] == ' ') ++string_index; + + std::string keyword; + + for (; string_index < ln.size(); ++string_index) { + if (ln[string_index] == variable.fEnd[0]) { + std::string varname = ""; + + for (size_t index_keyword = ln.find(' '); ln[index_keyword] != variable.fBegin[0]; + ++index_keyword) { + if (ln[index_keyword] == ' ') { + continue; + } + + if (isdigit(ln[index_keyword])) { + goto cc_next_loop; + } + + varname += ln[index_keyword]; + } + + if (varname.find(' ') != std::string::npos) { + varname.erase(0, varname.find(' ')); + + if (variable.fBegin == "extern") { + varname.erase(0, varname.find(' ')); + } + } + + if (kRegisterCounter == 5 || kRegisterCounter == 6) ++kRegisterCounter; + + std::string reg = kAsmRegisterPrefix; + reg += std::to_string(kRegisterCounter); + + kCompilerVariables.push_back({.fValue = varname}); + goto cc_check_done; + } + + keyword.push_back(ln[string_index]); + } + + goto cc_next_loop; + + cc_check_done: + + // skip digit value. + if (isdigit(keyword[0]) || keyword[0] == '"') { + goto cc_next_loop; + } + + while (keyword.find(' ') != std::string::npos) keyword.erase(keyword.find(' '), 1); + + for (auto& var : kCompilerVariables) { + if (var.fValue.find(keyword) != std::string::npos) { + err_str.clear(); + goto cc_next; + } + } + + for (auto& fn : kCompilerFunctions) { + if (fn.find(keyword[0]) != std::string::npos) { + auto where_begin = fn.find(keyword[0]); + auto keyword_begin = 0UL; + auto failed = false; + + for (; where_begin < keyword.size(); ++where_begin) { + if (fn[where_begin] == '(' && keyword[keyword_begin] == '(') break; + + if (fn[where_begin] != keyword[keyword_begin]) { + failed = true; + break; + } + + ++keyword_begin; + } + + if (!failed) { + err_str.clear(); + goto cc_next; + } else { + continue; + } + } + } + + cc_error_value: + if (keyword.find("->") != std::string::npos) return err_str; + + if (keyword.find(".") != std::string::npos) return err_str; + + if (isalnum(keyword[0])) err_str += "\nUndefined value: " + keyword; + + return err_str; + } + + cc_next_loop: + continue; + } + +cc_next: + + // extern does not declare anything, it extern_segments a variable. + // so that's why it's not declare upper. + if (CompilerKit::find_word(ln, "extern")) { + auto substr = ln.substr(ln.find("extern") + strlen("extern")); + kCompilerVariables.push_back({.fValue = substr}); + } + + if (kShouldHaveBraces && ln.find('{') == std::string::npos) { + err_str += "Missing '{' for function "; + err_str += kFnName; + err_str += "\n"; + + kShouldHaveBraces = false; + kFnName.clear(); + } else if (kShouldHaveBraces && ln.find('{') != std::string::npos) { + kShouldHaveBraces = false; + kFnName.clear(); + } + + bool type_not_found = true; + + if (ln.find('\'') != std::string::npos) { + ln.replace(ln.find('\''), 3, "0"); + } + + auto first = ln.find('"'); + if (first != std::string::npos) { + auto second = 0UL; + bool found_second_quote = false; + + for (size_t i = first + 1; i < ln.size(); ++i) { + if (ln[i] == '\"') { + found_second_quote = true; + second = i; + + break; + } + } + + if (!found_second_quote) { + err_str += "Missing terminating \"."; + err_str += " here -> " + ln.substr(ln.find('"'), second); + } + } + + if (ln.find(')') != std::string::npos && ln.find(';') == std::string::npos) { + if (ln.find('{') == std::string::npos) { + kFnName = ln; + kShouldHaveBraces = true; + + goto skip_braces_check; + } else if (ln.find('{') != std::string::npos) { + kShouldHaveBraces = false; + } + } + +skip_braces_check: + + for (auto& key : kCompilerTypes) { + if (CompilerKit::find_word(ln, key.fName)) { + if (isdigit(ln[ln.find(key.fName) + key.fName.size() + 1])) { + err_str += "\nNumber cannot be set for "; + err_str += key.fName; + err_str += "'s name. here -> "; + err_str += ln; + } + + if (ln.find(key.fName) == 0 || ln[ln.find(key.fName) - 1] == ' ' || + ln[ln.find(key.fName) - 1] == '\t') { + type_not_found = false; + + if (ln[ln.find(key.fName) + key.fName.size()] != ' ') { + type_not_found = true; + + if (ln[ln.find(key.fName) + key.fName.size()] == '\t') type_not_found = false; + + goto next; + } else if (ln[ln.find(key.fName) + key.fName.size()] != '\t') { + type_not_found = true; + + if (ln[ln.find(key.fName) + key.fName.size()] == ' ') type_not_found = false; + } + } + + next: + + if (ln.find(';') == std::string::npos) { + if (ln.find('(') != std::string::npos) { + if (ln.find('=') == std::string::npos) continue; + } + + err_str += "\nMissing ';', here -> "; + err_str += ln; + } else { + continue; + } + + if (ln.find('=') != std::string::npos) { + if (ln.find('(') != std::string::npos) { + if (ln.find(')') == std::string::npos) { + err_str += "\nMissing ')', after '(' here -> "; + err_str += ln.substr(ln.find('(')); + } + } + } + } + } + + if (kInBraces && ln.find("struct") != std::string::npos && + ln.find("union") != std::string::npos && ln.find("enum") != std::string::npos && + ln.find('=') != std::string::npos) { + if (ln.find(';') == std::string::npos) { + err_str += "\nMissing ';' after struct/union/enum declaration, here -> "; + err_str += ln; + } + } + + if (ln.find(';') != std::string::npos && ln.find("for") == std::string::npos) { + if (ln.find(';') + 1 != ln.size()) { + for (int i = 0; i < ln.substr(ln.find(';') + 1).size(); ++i) { + if ((ln.substr(ln.find(';') + 1)[i] != ' ') || (ln.substr(ln.find(';') + 1)[i] != '\t')) { + if (auto err = this->Check(ln.substr(ln.find(';') + 1).c_str(), file); !err.empty()) { + err_str += "\nUnexpected text after ';' -> "; + err_str += ln.substr(ln.find(';')); + err_str += err; + } + } + } + } + } + + if (ln.find('(') != std::string::npos) { + if (ln.find(';') == std::string::npos && !CompilerKit::find_word(ln, "|") && + !CompilerKit::find_word(ln, "||") && !CompilerKit::find_word(ln, "&") && + !CompilerKit::find_word(ln, "&&") && !CompilerKit::find_word(ln, "~")) { + bool found_func = false; + size_t i = ln.find('('); + std::vector<char> opens; + std::vector<char> closes; + + for (; i < ln.size(); ++i) { + if (ln[i] == ')') { + closes.push_back(1); + } + + if (ln[i] == '(') { + opens.push_back(1); + } + } + + if (closes.size() != opens.size()) err_str += "Unterminated (), here -> " + ln; + + bool space_found = false; + + for (int i = 0; i < ln.size(); ++i) { + if (ln[i] == ')' && !space_found) { + space_found = true; + continue; + } + + if (space_found) { + if (ln[i] == ' ' && isalnum(ln[i + 1])) { + err_str += "\nBad function format here -> "; + err_str += ln; + } + } + } + } + + if (ln.find('(') < 1) { + err_str += "\nMissing identifier before '(' here -> "; + err_str += ln; + } else { + if (type_not_found && ln.find(';') == std::string::npos && + ln.find("if") == std::string::npos && ln.find("|") == std::string::npos && + ln.find("&") == std::string::npos && ln.find("(") == std::string::npos && + ln.find(")") == std::string::npos) { + err_str += "\n Missing ';' or type, here -> "; + err_str += ln; + } + } + + if (ln.find(')') == std::string::npos) { + err_str += "\nMissing ')', after '(' here -> "; + err_str += ln.substr(ln.find('(')); + } + } else { + if (ln.find("for") != std::string::npos || ln.find("while") != std::string::npos) { + err_str += "\nMissing '(', after \"for\", here -> "; + err_str += ln; + } + } + + if (ln.find('}') != std::string::npos && !kInBraces) { + if (!kInStruct && ln.find(';') == std::string::npos) { + err_str += "\nMismatched '}', here -> "; + err_str += ln; + } + } + + if (!ln.empty()) { + if (ln.find(';') == std::string::npos && ln.find('{') == std::string::npos && + ln.find('}') == std::string::npos && ln.find(')') == std::string::npos && + ln.find('(') == std::string::npos && ln.find(',') == std::string::npos) { + if (ln.size() <= 2) return err_str; + + err_str += "\nMissing ';', here -> "; + err_str += ln; + } + } + + return err_str; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +/** + * @brief C To Assembly mount-point. + */ + +///////////////////////////////////////////////////////////////////////////////////////// + +class AssemblyCCInterface final LC_ASSEMBLY_INTERFACE { + public: + explicit AssemblyCCInterface() = default; + ~AssemblyCCInterface() override = default; + + LIBCOMPILER_COPY_DEFAULT(AssemblyCCInterface); + + UInt32 Arch() noexcept override { return CompilerKit::AssemblyFactory::kArch64x0; } + + Int32 CompileToFormat(std::string src, Int32 arch) override { + if (kCompilerFrontend == nullptr) return 1; + + /* @brief copy contents wihtout extension */ + std::string src_file = src.data(); + std::ifstream src_fp = std::ifstream(src_file, std::ios::in); + std::string dest; + + for (auto& ch : src_file) { + if (ch == '.') { + break; + } + + dest += ch; + } + + /* According to PEF ABI. */ + std::vector<const char*> exts = kAsmFileExts; + dest += exts[4]; + + kState.fOutputAssembly = std::make_unique<std::ofstream>(dest); + + auto fmt = CompilerKit::current_date(); + + (*kState.fOutputAssembly) << "# Path: " << src_file << "\n"; + (*kState.fOutputAssembly) << "# Language: 64x0 Assembly (Generated from ANSI C)\n"; + (*kState.fOutputAssembly) << "# Date: " << fmt << "\n\n"; + + CompilerKit::SyntaxLeafList syntax; + + kState.fSyntaxTreeList.push_back(syntax); + kState.fSyntaxTree = &kState.fSyntaxTreeList[kState.fSyntaxTreeList.size() - 1]; + + std::string line_src; + + while (std::getline(src_fp, line_src)) { + if (auto err = kCompilerFrontend->Check(line_src.c_str(), src.data()); err.empty()) { + kCompilerFrontend->Compile(line_src, src.data()); + } else { + Detail::print_error(err, src.data()); + } + } + + if (kAcceptableErrors > 0) return 1; + + std::vector<std::string> keywords = {"ldw", "stw", "lda", "sta", "add", "sub", "mv"}; + + /// + /// Replace, optimize, fix assembly output. + /// + + for (auto& leaf : kState.fSyntaxTree->fLeafList) { + std::vector<std::string> access_keywords = {"->", "."}; + + for (auto& access_ident : access_keywords) { + if (CompilerKit::find_word(leaf.fUserValue, access_ident)) { + for (auto& struc : kState.kStructMap) { + /// TODO: + } + } + } + + for (auto& keyword : keywords) { + if (CompilerKit::find_word(leaf.fUserValue, keyword)) { + std::size_t cnt = 0UL; + + for (auto& reg : kState.kStackFrame) { + std::string needle; + + for (size_t i = 0; i < reg.fName.size(); i++) { + if (reg.fName[i] == ' ') { + ++i; + + for (; i < reg.fName.size(); i++) { + if (reg.fName[i] == ',') { + break; + } + + if (reg.fName[i] == ' ') continue; + + needle += reg.fName[i]; + } + + break; + } + } + + if (CompilerKit::find_word(leaf.fUserValue, needle)) { + if (leaf.fUserValue.find("extern_segment " + needle) != std::string::npos) { + std::string range = "extern_segment " + needle; + leaf.fUserValue.replace(leaf.fUserValue.find("extern_segment " + needle), + range.size(), needle); + } + + if (leaf.fUserValue.find("ldw r6") != std::string::npos) { + std::string::difference_type countComma = + std::count(leaf.fUserValue.begin(), leaf.fUserValue.end(), ','); + + if (countComma == 1) { + leaf.fUserValue.replace(leaf.fUserValue.find("ldw"), strlen("ldw"), "mv"); + } + } + + leaf.fUserValue.replace(leaf.fUserValue.find(needle), needle.size(), reg.fReg); + + ++cnt; + } + } + + if (cnt > 1 && keyword != "mv" && keyword != "add" && keyword != "sub") { + leaf.fUserValue.replace(leaf.fUserValue.find(keyword), keyword.size(), "mv"); + } + } + } + } + + for (auto& leaf : kState.fSyntaxTree->fLeafList) { + (*kState.fOutputAssembly) << leaf.fUserValue; + } + + kState.fSyntaxTree = nullptr; + + kState.fOutputAssembly->flush(); + kState.fOutputAssembly.reset(); + + return kExitOK; + } +}; + +///////////////////////////////////////////////////////////////////////////////////////// + +#include <CompilerKit/Version.h> + +#define kPrintF printf +#define kSplashCxx() kPrintF(kWhite "NE C Driver, %s, (c) Amlal El Mahrouss\n", kDistVersion) + +static void cc_print_help() { + kSplashCxx(); +} + +///////////////////////////////////////////////////////////////////////////////////////// + +#define kExt ".c" + +LIBCOMPILER_MODULE(CompilerCLang64x0) { + ::signal(SIGSEGV, Detail::drvi_crash_handler); + + kCompilerTypes.push_back({.fName = "void", .fValue = "void"}); + kCompilerTypes.push_back({.fName = "char", .fValue = "byte"}); + kCompilerTypes.push_back({.fName = "short", .fValue = "hword"}); + kCompilerTypes.push_back({.fName = "int", .fValue = "dword"}); + kCompilerTypes.push_back({.fName = "long", .fValue = "qword"}); + kCompilerTypes.push_back({.fName = "*", .fValue = "offset"}); + + bool skip = false; + + kFactory.Mount(new AssemblyCCInterface()); + kMachine = CompilerKit::AssemblyFactory::kArch64x0; + kCompilerFrontend = new CompilerFrontend64x0(); + + for (auto index = 1UL; index < argc; ++index) { + if (skip) { + skip = false; + continue; + } + + if (argv[index][0] == '-') { + if (strcmp(argv[index], "--v") == 0 || strcmp(argv[index], "--version") == 0) { + kSplashCxx(); + return kExitOK; + } + + if (strcmp(argv[index], "--verbose") == 0) { + kState.fVerbose = true; + + continue; + } + + if (strcmp(argv[index], "--h") == 0 || strcmp(argv[index], "--help") == 0) { + cc_print_help(); + + return kExitOK; + } + + if (strcmp(argv[index], "--dialect") == 0) { + if (kCompilerFrontend) std::cout << kCompilerFrontend->Language() << "\n"; + + return kExitOK; + } + + if (strcmp(argv[index], "--fmax-exceptions") == 0) { + try { + kErrorLimit = std::strtol(argv[index + 1], nullptr, 10); + } + // catch anything here + catch (...) { + kErrorLimit = 0; + } + + skip = true; + + continue; + } + + std::string err = "Unknown command: "; + err += argv[index]; + + Detail::print_error(err, "cc"); + + continue; + } + + kFileList.emplace_back(argv[index]); + + std::string srcFile = argv[index]; + + if (strstr(argv[index], kExt) == nullptr) { + if (kState.fVerbose) { + Detail::print_error(srcFile + " is not a valid C source.\n", "cc"); + } + + return 1; + } + + if (kFactory.Compile(srcFile, kMachine) != kExitOK) return 1; + } + + return kExitOK; +} + +// Last rev 8-1-24 diff --git a/dev/CompilerKit/src/Frontend/CCompilerARM64.cc b/dev/CompilerKit/src/Frontend/CCompilerARM64.cc new file mode 100644 index 0000000..764bec2 --- /dev/null +++ b/dev/CompilerKit/src/Frontend/CCompilerARM64.cc @@ -0,0 +1,1284 @@ +/* + * ======================================================== + * + * cc + * Copyright (C) 2024-2025 Amlal El Mahrouss, all rights reserved. + * + * ======================================================== + */ + +/// BUGS: 0 +/// TODO: none + +#include <CompilerKit/detail/Aarch64.h> +#include <CompilerKit/Frontend.h> +#include <CompilerKit/UUID.h> +#include <CompilerKit/utils/CompilerUtils.h> +#include <cstdio> +#include <fstream> +#include <iostream> +#include <memory> +#include <random> +#include <string> +#include <utility> +#include <vector> + +/* C driver */ +/* This is part of the CompilerKit. */ +/* (c) Amlal El Mahrouss */ + +/// @author EL Mahrouss Amlal (amlel) +/// @file ARM64-cc.cxx +/// @brief ARM64 C Compiler. + +/// TODO: support structures, else if, else, . and -> + +///////////////////// + +// ANSI ESCAPE CODES + +///////////////////// + +#define kExitOK (0) + +#define kBlank "\e[0;30m" +#define kRed "\e[0;31m" +#define kWhite "\e[0;97m" + +///////////////////////////////////// + +// INTERNAL STUFF OF THE C COMPILER + +///////////////////////////////////// + +namespace Detail { +// \brief Register map structure, used to keep track of each variable's registers. +struct CompilerRegisterMap final { + std::string fName; + std::string fReg; +}; + +// \brief Map for C structs +// \author amlel +struct CompilerStructMap final { + // 'my_foo' + std::string fName; + + // if instance: stores a valid register. + std::string fReg; + + // offset count + std::size_t fOffsetsCnt; + + // offset array. + std::vector<std::pair<Int32, std::string>> fOffsets; +}; + +struct CompilerState final { + std::vector<CompilerKit::SyntaxLeafList> fSyntaxTreeList; + std::vector<CompilerRegisterMap> kStackFrame; + std::vector<CompilerStructMap> kStructMap; + CompilerKit::SyntaxLeafList* fSyntaxTree{nullptr}; + std::unique_ptr<std::ofstream> fOutputAssembly; + std::string fLastFile; + std::string fLastError; + bool fVerbose; +}; +} // namespace Detail + +static Detail::CompilerState kState; +static std::string kIfFunction = ""; + +namespace Detail { +/// @brief prints an error into stdout. +/// @param reason the reason of the error. +/// @param file where does it originate from? +void print_error(std::string reason, std::string file) noexcept; + +struct CompilerType final { + std::string fName; + std::string fValue; +}; +} // namespace Detail + +///////////////////////////////////////////////////////////////////////////////////////// + +// Target architecture. +static int kMachine = 0; + +///////////////////////////////////////// + +// REGISTERS ACCORDING TO USED ASSEMBLER + +///////////////////////////////////////// + +static size_t kRegisterCnt = kAsmRegisterLimit; +static size_t kStartUsable = 8; +static size_t kUsableLimit = 15; +static size_t kRegisterCounter = kStartUsable; +static std::string kRegisterPrefix = kAsmRegisterPrefix; + +///////////////////////////////////////// + +// COMPILER PARSING UTILITIES/STATES. + +///////////////////////////////////////// + +static std::vector<std::string> kFileList; +static CompilerKit::AssemblyFactory kFactory; +static bool kInStruct = false; +static bool kOnWhileLoop = false; +static bool kOnForLoop = false; +static bool kInBraces = false; +static bool kIfFound = false; +static size_t kBracesCount = 0UL; + +/* @brief C compiler backend for C */ +class CompilerFrontendARM64 final : public CompilerKit::CompilerFrontendInterface { + public: + explicit CompilerFrontendARM64() = default; + ~CompilerFrontendARM64() override = default; + + LIBCOMPILER_COPY_DEFAULT(CompilerFrontendARM64); + + std::string Check(const char* text, const char* file); + CompilerKit::SyntaxLeafList::SyntaxLeaf Compile(std::string text, std::string file) override; + + const char* Language() override { return "64k C"; } +}; + +static CompilerFrontendARM64* kCompilerFrontend = nullptr; +static std::vector<Detail::CompilerType> kCompilerVariables; +static std::vector<std::string> kCompilerFunctions; +static std::vector<Detail::CompilerType> kCompilerTypes; + +namespace Detail { +union number_cast final { + public: + number_cast(UInt64 _Raw) : _Raw(_Raw) {} + + public: + char _Num[8]; + UInt64 _Raw; +}; + +union double_cast final { + public: + double_cast(float _Raw) : _Raw(_Raw) {} + + public: + char _Sign; + char _Lh[8]; + char _Rh[23]; + + float _Raw; +}; +} // namespace Detail + +///////////////////////////////////////////////////////////////////////////////////////// + +// @name Compile +// @brief Generate MASM from a C assignement. + +///////////////////////////////////////////////////////////////////////////////////////// + +CompilerKit::SyntaxLeafList::SyntaxLeaf CompilerFrontendARM64::Compile(std::string text, std::string file) { + bool typeFound = false; + bool fnFound = false; + + // setup generator. + std::random_device rd; + + auto seed_data = std::array<int, std::mt19937::state_size>{}; + std::generate(std::begin(seed_data), std::end(seed_data), std::ref(rd)); + std::seed_seq seq(std::begin(seed_data), std::end(seed_data)); + std::mt19937 generator(seq); + + // start parsing + for (size_t text_index = 0; text_index < text.size(); ++text_index) { + auto syntaxLeaf = CompilerKit::SyntaxLeafList::SyntaxLeaf(); + + auto gen = uuids::uuid_random_generator{generator}; + uuids::uuid out = gen(); + + Detail::number_cast time_off = (UInt64) out.as_bytes().data(); + + if (!typeFound) { + auto substr = text.substr(text_index); + std::string match_type; + + for (size_t y = 0; y < substr.size(); ++y) { + if (substr[y] == ' ') { + while (match_type.find(' ') != std::string::npos) { + match_type.erase(match_type.find(' ')); + } + + for (auto& clType : kCompilerTypes) { + if (clType.fName == match_type) { + match_type.clear(); + + std::string buf; + + buf += clType.fValue; + buf += ' '; + + if (substr.find('=') != std::string::npos) { + break; + } + + if (text.find('(') != std::string::npos) { + syntaxLeaf.fUserValue = buf; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + } + + typeFound = true; + break; + } + } + + break; + } + + match_type += substr[y]; + } + } + + if (text[text_index] == '{') { + if (kInStruct) { + continue; + } + + kInBraces = true; + ++kBracesCount; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + } + + // return keyword handler + if (text[text_index] == 'r') { + std::string return_keyword; + return_keyword += "return"; + + std::size_t index = 0UL; + + std::string value; + + for (size_t return_index = text_index; return_index < text.size(); ++return_index) { + if (text[return_index] != return_keyword[index]) { + for (size_t value_index = return_index; value_index < text.size(); ++value_index) { + if (text[value_index] == ';') break; + + value += text[value_index]; + } + + break; + } + + ++index; + } + + if (index == return_keyword.size()) { + if (!value.empty()) { + if (value.find('(') != std::string::npos) { + value.erase(value.find('(')); + } + + if (!isdigit(value[value.find('(') + 2])) { + std::string tmp = value; + bool reg_to_reg = false; + + value.clear(); + + value += " extern_segment"; + value += tmp; + } + + syntaxLeaf.fUserValue = "\tldw r19, "; + + // make it pretty. + if (value.find('\t') != std::string::npos) value.erase(value.find('\t'), 1); + + syntaxLeaf.fUserValue += value + "\n"; + } + + syntaxLeaf.fUserValue += "\tjlr"; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + break; + } + } + + if (text[text_index] == 'i' && text[text_index + 1] == 'f') { + auto expr = text.substr(text_index + 2); + text.erase(text_index, 2); + + if (expr.find("{") != std::string::npos) { + expr.erase(expr.find("{")); + } + + if (expr.find("(") != std::string::npos) expr.erase(expr.find("(")); + + if (expr.find(")") != std::string::npos) expr.erase(expr.find(")")); + + kIfFunction = "__LIBCOMPILER_IF_PROC_"; + kIfFunction += std::to_string(time_off._Raw); + + syntaxLeaf.fUserValue = "\tlda r12, extern_segment "; + syntaxLeaf.fUserValue += kIfFunction + + "\n\t#r12 = Code to jump on, r11 right cond, r10 left cond.\n\tbeq " + "r10, r11, r12\ndword public_segment .code64 " + + kIfFunction + "\n"; + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + kIfFound = true; + } + + // Parse expressions and instructions here. + // what does this mean? + // we encounter an assignment, or we reached the end of an expression. + if (text[text_index] == '=' || text[text_index] == ';') { + if (fnFound) continue; + if (kIfFound) continue; + + if (text[text_index] == ';' && kInStruct) continue; + + if (text.find("typedef ") != std::string::npos) continue; + + if (text[text_index] == '=' && kInStruct) { + Detail::print_error("assignement of value in struct " + text, file); + continue; + } + + if (text[text_index] == ';' && kInStruct) { + bool space_found_ = false; + std::string sym; + + for (auto& ch : text) { + if (ch == ' ') { + space_found_ = true; + } + + if (ch == ';') break; + + if (space_found_) sym.push_back(ch); + } + + kState.kStructMap[kState.kStructMap.size() - 1].fOffsets.push_back( + std::make_pair(kState.kStructMap[kState.kStructMap.size() - 1].fOffsetsCnt + 4, sym)); + + kState.kStructMap[kState.kStructMap.size() - 1].fOffsetsCnt = + kState.kStructMap[kState.kStructMap.size() - 1].fOffsetsCnt + 4; + + continue; + } + + if (text[text_index] == '=' && kInStruct) { + continue; + } + + if (text[text_index + 1] == '=' || text[text_index - 1] == '!' || + text[text_index - 1] == '<' || text[text_index - 1] == '>') { + continue; + } + + std::string substr; + + if (text.find('=') != std::string::npos && kInBraces && !kIfFound) { + if (text.find("*") != std::string::npos) { + if (text.find("=") > text.find("*")) + substr += "\tlda "; + else + substr += "\tldw "; + } else { + substr += "\tldw "; + } + } else if (text.find('=') != std::string::npos && !kInBraces) { + substr += "stw public_segment .data64 "; + } + + int first_encountered = 0; + + std::string str_name; + + for (size_t text_index_2 = 0; text_index_2 < text.size(); ++text_index_2) { + if (text[text_index_2] == '\"') { + ++text_index_2; + + // want to add this, so that the parser recognizes that this is a + // string. + substr += '"'; + + for (; text_index_2 < text.size(); ++text_index_2) { + if (text[text_index_2] == '\"') break; + + substr += text[text_index_2]; + } + } + + if (text[text_index_2] == '{' || text[text_index_2] == '}') continue; + + if (text[text_index_2] == ';') { + break; + } + + if (text[text_index_2] == ' ' || text[text_index_2] == '\t') { + if (first_encountered != 2) { + if (text[text_index] != '=' && + substr.find("public_segment .data64") == std::string::npos && !kInStruct) + substr += "public_segment .data64 "; + } + + ++first_encountered; + + continue; + } + + if (text[text_index_2] == '=') { + if (!kInBraces) { + substr.replace(substr.find("public_segment .data64"), strlen("public_segment .data64"), + "public_segment .zero64 "); + } + + substr += ","; + continue; + } + + substr += text[text_index_2]; + } + + for (auto& clType : kCompilerTypes) { + if (substr.find(clType.fName) != std::string::npos) { + if (substr.find(clType.fName) > substr.find('"')) continue; + + substr.erase(substr.find(clType.fName), clType.fName.size()); + } else if (substr.find(clType.fValue) != std::string::npos) { + if (substr.find(clType.fValue) > substr.find('"')) continue; + + if (clType.fName == "const") continue; + + substr.erase(substr.find(clType.fValue), clType.fValue.size()); + } + } + + if (substr.find("extern") != std::string::npos) { + substr.replace(substr.find("extern"), strlen("extern"), "extern_segment "); + + if (substr.find("public_segment .data64") != std::string::npos) + substr.erase(substr.find("public_segment .data64"), strlen("public_segment .data64")); + } + + auto var_to_find = std::find_if( + kCompilerVariables.cbegin(), kCompilerVariables.cend(), + [&](Detail::CompilerType type) { return type.fName.find(substr) != std::string::npos; }); + + if (kRegisterCounter == 5 || kRegisterCounter == 6) ++kRegisterCounter; + + std::string reg = kAsmRegisterPrefix; + reg += std::to_string(kRegisterCounter); + + if (var_to_find == kCompilerVariables.cend()) { + ++kRegisterCounter; + + kState.kStackFrame.push_back({.fName = substr, .fReg = reg}); + kCompilerVariables.push_back({.fName = substr}); + } + + syntaxLeaf.fUserValue += substr; + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + if (text[text_index] == '=') break; + } + + // function handler. + + if (text[text_index] == '(' && !fnFound && !kIfFound) { + std::string substr; + std::string args_buffer; + std::string args; + + bool type_crossed = false; + + for (size_t idx = text.find('(') + 1; idx < text.size(); ++idx) { + if (text[idx] == ',') continue; + + if (text[idx] == ' ') continue; + + if (text[idx] == ')') break; + } + + for (char substr_first_index : text) { + if (substr_first_index != ',') + args_buffer += substr_first_index; + else + args_buffer += '$'; + + if (substr_first_index == ';') { + args_buffer = args_buffer.erase(0, args_buffer.find('(')); + args_buffer = args_buffer.erase(args_buffer.find(';'), 1); + args_buffer = args_buffer.erase(args_buffer.find(')'), 1); + args_buffer = args_buffer.erase(args_buffer.find('('), 1); + + if (!args_buffer.empty()) args += "\tldw r6, "; + + std::string register_type; + std::size_t index = 7UL; + + while (args_buffer.find("$") != std::string::npos) { + register_type = kRegisterPrefix; + register_type += std::to_string(index); + + ++index; + + args_buffer.replace(args_buffer.find('$'), 1, "\n\tldw " + register_type + ","); + } + + args += args_buffer; + args += "\n\tlda r19, "; + } + } + + for (char _text_i : text) { + if (_text_i == '\t' || _text_i == ' ') { + if (!type_crossed) { + substr.clear(); + type_crossed = true; + } + + continue; + } + + if (_text_i == '(') break; + + substr += _text_i; + } + + if (kInBraces) { + syntaxLeaf.fUserValue = args; + syntaxLeaf.fUserValue += substr; + syntaxLeaf.fUserValue += "\n\tjrl\n"; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + fnFound = true; + } else { + syntaxLeaf.fUserValue.clear(); + + syntaxLeaf.fUserValue += "public_segment .code64 "; + + syntaxLeaf.fUserValue += substr; + syntaxLeaf.fUserValue += "\n"; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + fnFound = true; + } + + kCompilerFunctions.push_back(text); + } + + if (text[text_index] == '-' && text[text_index + 1] == '-') { + text = text.replace(text.find("--"), strlen("--"), ""); + + for (int _text_i = 0; _text_i < text.size(); ++_text_i) { + if (text[_text_i] == '\t' || text[_text_i] == ' ') text.erase(_text_i, 1); + } + + syntaxLeaf.fUserValue += "sub "; + syntaxLeaf.fUserValue += text; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + break; + } + + if (text[text_index] == '}') { + kRegisterCounter = kStartUsable; + + --kBracesCount; + + if (kBracesCount < 1) { + kInBraces = false; + kBracesCount = 0; + } + + if (kIfFound) kIfFound = false; + + if (kInStruct) kInStruct = false; + + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + } + + syntaxLeaf.fUserValue.clear(); + } + + auto syntaxLeaf = CompilerKit::SyntaxLeafList::SyntaxLeaf(); + syntaxLeaf.fUserValue = "\n"; + kState.fSyntaxTree->fLeafList.push_back(syntaxLeaf); + + return syntaxLeaf; +} + +static bool kShouldHaveBraces = false; +static std::string kFnName; + +std::string CompilerFrontendARM64::Check(const char* text, const char* file) { + std::string err_str; + std::string ln = text; + + if (ln.empty()) { + return err_str; + } + + bool non_ascii_found = false; + + for (int i = 0; i < ln.size(); ++i) { + if (isalnum(ln[i])) { + non_ascii_found = true; + break; + } + } + + if (kShouldHaveBraces && ln.find('{') != std::string::npos) { + kShouldHaveBraces = false; + } + + if (!non_ascii_found) return err_str; + + size_t string_index = 1UL; + + if (ln.find('\'') != std::string::npos) { + string_index = ln.find('\'') + 1; + + for (; string_index < ln.size(); ++string_index) { + if (ln[string_index] == '\'') { + if (ln[string_index + 1] != ';') { + ln.erase(string_index, 1); + } + + return err_str; + } + } + } else if (ln.find('"') != std::string::npos) { + string_index = ln.find('"') + 1; + + for (; string_index < ln.size(); ++string_index) { + if (ln[string_index] == '"') { + if (ln[string_index + 1] != ';') { + ln.erase(string_index, 1); + } else { + break; + } + } + } + } else if (ln.find('"') == std::string::npos && ln.find('\'') == std::string::npos) { + std::vector<std::string> forbidden_words; + + forbidden_words.push_back("\\"); + forbidden_words.push_back("?"); + forbidden_words.push_back("@"); + forbidden_words.push_back("~"); + forbidden_words.push_back("::"); + forbidden_words.push_back("--*"); + forbidden_words.push_back("*/"); + + // add them to avoid stupid mistakes. + forbidden_words.push_back("namespace"); + forbidden_words.push_back("class"); + forbidden_words.push_back("extern \"C\""); + + for (auto& forbidden : forbidden_words) { + if (ln.find(forbidden) != std::string::npos) { + err_str += "\nForbidden character detected: "; + err_str += forbidden; + + return err_str; + } + } + } + + struct CompilerVariableRange final { + std::string fBegin; + std::string fEnd; + }; + + const std::vector<CompilerVariableRange> variables_list = { + {.fBegin = "static ", .fEnd = "="}, {.fBegin = "=", .fEnd = ";"}, + {.fBegin = "if(", .fEnd = "="}, {.fBegin = "if (", .fEnd = "="}, + {.fBegin = "if(", .fEnd = "<"}, {.fBegin = "if (", .fEnd = "<"}, + {.fBegin = "if(", .fEnd = ">"}, {.fBegin = "if (", .fEnd = ">"}, + {.fBegin = "if(", .fEnd = ")"}, {.fBegin = "if (", .fEnd = ")"}, + + {.fBegin = "else(", .fEnd = "="}, {.fBegin = "else (", .fEnd = "="}, + {.fBegin = "else(", .fEnd = "<"}, {.fBegin = "else (", .fEnd = "<"}, + {.fBegin = "else(", .fEnd = ">"}, {.fBegin = "else (", .fEnd = ">"}, + {.fBegin = "else(", .fEnd = ")"}, {.fBegin = "else (", .fEnd = ")"}, + }; + + for (auto& variable : variables_list) { + if (ln.find(variable.fBegin) != std::string::npos) { + string_index = ln.find(variable.fBegin) + variable.fBegin.size(); + + while (ln[string_index] == ' ') ++string_index; + + std::string keyword; + + for (; string_index < ln.size(); ++string_index) { + if (ln[string_index] == variable.fEnd[0]) { + std::string varname = ""; + + for (size_t index_keyword = ln.find(' '); ln[index_keyword] != variable.fBegin[0]; + ++index_keyword) { + if (ln[index_keyword] == ' ') { + continue; + } + + if (isdigit(ln[index_keyword])) { + goto cc_next_loop; + } + + varname += ln[index_keyword]; + } + + if (varname.find(' ') != std::string::npos) { + varname.erase(0, varname.find(' ')); + + if (variable.fBegin == "extern") { + varname.erase(0, varname.find(' ')); + } + } + + if (kRegisterCounter == 5 || kRegisterCounter == 6) ++kRegisterCounter; + + std::string reg = kAsmRegisterPrefix; + reg += std::to_string(kRegisterCounter); + + kCompilerVariables.push_back({.fValue = varname}); + goto cc_check_done; + } + + keyword.push_back(ln[string_index]); + } + + goto cc_next_loop; + + cc_check_done: + + // skip digit value. + if (isdigit(keyword[0]) || keyword[0] == '"') { + goto cc_next_loop; + } + + while (keyword.find(' ') != std::string::npos) keyword.erase(keyword.find(' '), 1); + + for (auto& var : kCompilerVariables) { + if (var.fValue.find(keyword) != std::string::npos) { + err_str.clear(); + goto cc_next; + } + } + + for (auto& fn : kCompilerFunctions) { + if (fn.find(keyword[0]) != std::string::npos) { + auto where_begin = fn.find(keyword[0]); + auto keyword_begin = 0UL; + auto failed = false; + + for (; where_begin < keyword.size(); ++where_begin) { + if (fn[where_begin] == '(' && keyword[keyword_begin] == '(') break; + + if (fn[where_begin] != keyword[keyword_begin]) { + failed = true; + break; + } + + ++keyword_begin; + } + + if (!failed) { + err_str.clear(); + goto cc_next; + } else { + continue; + } + } + } + + cc_error_value: + if (keyword.find("->") != std::string::npos) return err_str; + + if (keyword.find(".") != std::string::npos) return err_str; + + if (isalnum(keyword[0])) err_str += "\nUndefined value: " + keyword; + + return err_str; + } + + cc_next_loop: + continue; + } + +cc_next: + + // extern does not declare anything, it extern_segments a variable. + // so that's why it's not declare upper. + if (CompilerKit::find_word(ln, "extern")) { + auto substr = ln.substr(ln.find("extern") + strlen("extern")); + kCompilerVariables.push_back({.fValue = substr}); + } + + if (kShouldHaveBraces && ln.find('{') == std::string::npos) { + err_str += "Missing '{' for function "; + err_str += kFnName; + err_str += "\n"; + + kShouldHaveBraces = false; + kFnName.clear(); + } else if (kShouldHaveBraces && ln.find('{') != std::string::npos) { + kShouldHaveBraces = false; + kFnName.clear(); + } + + bool type_not_found = true; + + if (ln.find('\'') != std::string::npos) { + ln.replace(ln.find('\''), 3, "0"); + } + + auto first = ln.find('"'); + if (first != std::string::npos) { + auto second = 0UL; + bool found_second_quote = false; + + for (size_t i = first + 1; i < ln.size(); ++i) { + if (ln[i] == '\"') { + found_second_quote = true; + second = i; + + break; + } + } + + if (!found_second_quote) { + err_str += "Missing terminating \"."; + err_str += " here -> " + ln.substr(ln.find('"'), second); + } + } + + if (ln.find(')') != std::string::npos && ln.find(';') == std::string::npos) { + if (ln.find('{') == std::string::npos) { + kFnName = ln; + kShouldHaveBraces = true; + + goto skip_braces_check; + } else if (ln.find('{') != std::string::npos) { + kShouldHaveBraces = false; + } + } + +skip_braces_check: + + for (auto& key : kCompilerTypes) { + if (CompilerKit::find_word(ln, key.fName)) { + if (isdigit(ln[ln.find(key.fName) + key.fName.size() + 1])) { + err_str += "\nNumber cannot be set for "; + err_str += key.fName; + err_str += "'s name. here -> "; + err_str += ln; + } + + if (ln.find(key.fName) == 0 || ln[ln.find(key.fName) - 1] == ' ' || + ln[ln.find(key.fName) - 1] == '\t') { + type_not_found = false; + + if (ln[ln.find(key.fName) + key.fName.size()] != ' ') { + type_not_found = true; + + if (ln[ln.find(key.fName) + key.fName.size()] == '\t') type_not_found = false; + + goto next; + } else if (ln[ln.find(key.fName) + key.fName.size()] != '\t') { + type_not_found = true; + + if (ln[ln.find(key.fName) + key.fName.size()] == ' ') type_not_found = false; + } + } + + next: + + if (ln.find(';') == std::string::npos) { + if (ln.find('(') != std::string::npos) { + if (ln.find('=') == std::string::npos) continue; + } + + err_str += "\nMissing ';', here -> "; + err_str += ln; + } else { + continue; + } + + if (ln.find('=') != std::string::npos) { + if (ln.find('(') != std::string::npos) { + if (ln.find(')') == std::string::npos) { + err_str += "\nMissing ')', after '(' here -> "; + err_str += ln.substr(ln.find('(')); + } + } + } + } + } + + if (kInBraces && ln.find("struct") != std::string::npos && + ln.find("union") != std::string::npos && ln.find("enum") != std::string::npos && + ln.find('=') != std::string::npos) { + if (ln.find(';') == std::string::npos) { + err_str += "\nMissing ';' after struct/union/enum declaration, here -> "; + err_str += ln; + } + } + + if (ln.find(';') != std::string::npos && ln.find("for") == std::string::npos) { + if (ln.find(';') + 1 != ln.size()) { + for (int i = 0; i < ln.substr(ln.find(';') + 1).size(); ++i) { + if ((ln.substr(ln.find(';') + 1)[i] != ' ') || (ln.substr(ln.find(';') + 1)[i] != '\t')) { + if (auto err = this->Check(ln.substr(ln.find(';') + 1).c_str(), file); !err.empty()) { + err_str += "\nUnexpected text after ';' -> "; + err_str += ln.substr(ln.find(';')); + err_str += err; + } + } + } + } + } + + if (ln.find('(') != std::string::npos) { + if (ln.find(';') == std::string::npos && !CompilerKit::find_word(ln, "|") && + !CompilerKit::find_word(ln, "||") && !CompilerKit::find_word(ln, "&") && + !CompilerKit::find_word(ln, "&&") && !CompilerKit::find_word(ln, "~")) { + bool found_func = false; + size_t i = ln.find('('); + std::vector<char> opens; + std::vector<char> closes; + + for (; i < ln.size(); ++i) { + if (ln[i] == ')') { + closes.push_back(1); + } + + if (ln[i] == '(') { + opens.push_back(1); + } + } + + if (closes.size() != opens.size()) err_str += "Unterminated (), here -> " + ln; + + bool space_found = false; + + for (int i = 0; i < ln.size(); ++i) { + if (ln[i] == ')' && !space_found) { + space_found = true; + continue; + } + + if (space_found) { + if (ln[i] == ' ' && isalnum(ln[i + 1])) { + err_str += "\nBad function format here -> "; + err_str += ln; + } + } + } + } + + if (ln.find('(') < 1) { + err_str += "\nMissing identifier before '(' here -> "; + err_str += ln; + } else { + if (type_not_found && ln.find(';') == std::string::npos && + ln.find("if") == std::string::npos && ln.find("|") == std::string::npos && + ln.find("&") == std::string::npos && ln.find("(") == std::string::npos && + ln.find(")") == std::string::npos) { + err_str += "\n Missing ';' or type, here -> "; + err_str += ln; + } + } + + if (ln.find(')') == std::string::npos) { + err_str += "\nMissing ')', after '(' here -> "; + err_str += ln.substr(ln.find('(')); + } + } else { + if (ln.find("for") != std::string::npos || ln.find("while") != std::string::npos) { + err_str += "\nMissing '(', after \"for\", here -> "; + err_str += ln; + } + } + + if (ln.find('}') != std::string::npos && !kInBraces) { + if (!kInStruct && ln.find(';') == std::string::npos) { + err_str += "\nMismatched '}', here -> "; + err_str += ln; + } + } + + if (!ln.empty()) { + if (ln.find(';') == std::string::npos && ln.find('{') == std::string::npos && + ln.find('}') == std::string::npos && ln.find(')') == std::string::npos && + ln.find('(') == std::string::npos && ln.find(',') == std::string::npos) { + if (ln.size() <= 2) return err_str; + + err_str += "\nMissing ';', here -> "; + err_str += ln; + } + } + + return err_str; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +/** + * @brief C To Assembly mount-point. + */ + +///////////////////////////////////////////////////////////////////////////////////////// + +class AssemblyCCInterface final LC_ASSEMBLY_INTERFACE { + public: + explicit AssemblyCCInterface() = default; + ~AssemblyCCInterface() override = default; + + LIBCOMPILER_COPY_DEFAULT(AssemblyCCInterface); + + UInt32 Arch() noexcept override { return CompilerKit::AssemblyFactory::kArchAARCH64; } + + Int32 CompileToFormat(std::string src, Int32 arch) override { + if (kCompilerFrontend == nullptr) return 1; + + /* @brief copy contents wihtout extension */ + std::string src_file = src.data(); + std::ifstream src_fp = std::ifstream(src_file, std::ios::in); + std::string dest; + + for (auto& ch : src_file) { + if (ch == '.') { + break; + } + + dest += ch; + } + + /* According to PEF ABI. */ + std::vector<const char*> exts = kAsmFileExts; + dest += exts[4]; + + kState.fOutputAssembly = std::make_unique<std::ofstream>(dest); + + auto fmt = CompilerKit::current_date(); + + (*kState.fOutputAssembly) << "# Path: " << src_file << "\n"; + (*kState.fOutputAssembly) << "# Language: ARM64 Assembly (Generated from ANSI C)\n"; + (*kState.fOutputAssembly) << "# Date: " << fmt << "\n\n"; + + CompilerKit::SyntaxLeafList syntax; + + kState.fSyntaxTreeList.push_back(syntax); + kState.fSyntaxTree = &kState.fSyntaxTreeList[kState.fSyntaxTreeList.size() - 1]; + + std::string line_src; + + while (std::getline(src_fp, line_src)) { + if (auto err = kCompilerFrontend->Check(line_src.c_str(), src.data()); err.empty()) { + kCompilerFrontend->Compile(line_src, src.data()); + } else { + Detail::print_error(err, src.data()); + } + } + + if (kAcceptableErrors > 0) return 1; + + std::vector<std::string> keywords = {"ldw", "stw", "lda", "sta", "add", "sub", "mv"}; + + /// + /// Replace, optimize, fix assembly output. + /// + + for (auto& leaf : kState.fSyntaxTree->fLeafList) { + std::vector<std::string> access_keywords = {"->", "."}; + + for (auto& access_ident : access_keywords) { + if (CompilerKit::find_word(leaf.fUserValue, access_ident)) { + for (auto& struc : kState.kStructMap) { + /// TODO: + } + } + } + + for (auto& keyword : keywords) { + if (CompilerKit::find_word(leaf.fUserValue, keyword)) { + std::size_t cnt = 0UL; + + for (auto& reg : kState.kStackFrame) { + std::string needle; + + for (size_t i = 0; i < reg.fName.size(); i++) { + if (reg.fName[i] == ' ') { + ++i; + + for (; i < reg.fName.size(); i++) { + if (reg.fName[i] == ',') { + break; + } + + if (reg.fName[i] == ' ') continue; + + needle += reg.fName[i]; + } + + break; + } + } + + if (CompilerKit::find_word(leaf.fUserValue, needle)) { + if (leaf.fUserValue.find("extern_segment " + needle) != std::string::npos) { + std::string range = "extern_segment " + needle; + leaf.fUserValue.replace(leaf.fUserValue.find("extern_segment " + needle), + range.size(), needle); + } + + if (leaf.fUserValue.find("ldw r6") != std::string::npos) { + std::string::difference_type countComma = + std::count(leaf.fUserValue.begin(), leaf.fUserValue.end(), ','); + + if (countComma == 1) { + leaf.fUserValue.replace(leaf.fUserValue.find("ldw"), strlen("ldw"), "mv"); + } + } + + leaf.fUserValue.replace(leaf.fUserValue.find(needle), needle.size(), reg.fReg); + + ++cnt; + } + } + + if (cnt > 1 && keyword != "mv" && keyword != "add" && keyword != "sub") { + leaf.fUserValue.replace(leaf.fUserValue.find(keyword), keyword.size(), "mv"); + } + } + } + } + + for (auto& leaf : kState.fSyntaxTree->fLeafList) { + (*kState.fOutputAssembly) << leaf.fUserValue; + } + + kState.fSyntaxTree = nullptr; + + kState.fOutputAssembly->flush(); + kState.fOutputAssembly.reset(); + + return kExitOK; + } +}; + +///////////////////////////////////////////////////////////////////////////////////////// + +#include <CompilerKit/Version.h> + +#define kPrintF printf +#define kSplashCxx() kPrintF(kWhite "NE C Driver, %s, (c) Amlal El Mahrouss\n", kDistVersion) + +static void cc_print_help() { + kSplashCxx(); +} + +///////////////////////////////////////////////////////////////////////////////////////// + +#define kCExtension ".c" + +LIBCOMPILER_MODULE(CompilerCLangARM64) { + ::signal(SIGSEGV, Detail::drvi_crash_handler); + + kCompilerTypes.push_back({.fName = "void", .fValue = "void"}); + kCompilerTypes.push_back({.fName = "char", .fValue = "byte"}); + kCompilerTypes.push_back({.fName = "short", .fValue = "hword"}); + kCompilerTypes.push_back({.fName = "int", .fValue = "dword"}); + kCompilerTypes.push_back({.fName = "long", .fValue = "qword"}); + kCompilerTypes.push_back({.fName = "*", .fValue = "offset"}); + + bool skip = false; + + kFactory.Mount(new AssemblyCCInterface()); + kMachine = CompilerKit::AssemblyFactory::kArchAARCH64; + kCompilerFrontend = new CompilerFrontendARM64(); + + for (auto index = 1UL; index < argc; ++index) { + if (skip) { + skip = false; + continue; + } + + if (argv[index][0] == '-') { + if (strcmp(argv[index], "--v") == 0 || strcmp(argv[index], "--version") == 0) { + kSplashCxx(); + return kExitOK; + } + + if (strcmp(argv[index], "--verbose") == 0) { + kState.fVerbose = true; + + continue; + } + + if (strcmp(argv[index], "--h") == 0 || strcmp(argv[index], "--help") == 0) { + cc_print_help(); + + return kExitOK; + } + + if (strcmp(argv[index], "--dialect") == 0) { + if (kCompilerFrontend) std::cout << kCompilerFrontend->Language() << "\n"; + + return kExitOK; + } + + if (strcmp(argv[index], "--fmax-exceptions") == 0) { + try { + kErrorLimit = std::strtol(argv[index + 1], nullptr, 10); + } + // catch anything here + catch (...) { + kErrorLimit = 0; + } + + skip = true; + + continue; + } + + std::string err = "Unknown command: "; + err += argv[index]; + + Detail::print_error(err, "cc"); + + continue; + } + + kFileList.emplace_back(argv[index]); + + std::string srcFile = argv[index]; + + if (strstr(argv[index], kCExtension) == nullptr) { + if (kState.fVerbose) { + Detail::print_error(srcFile + " is not a valid C source.\n", "cc"); + } + + return 1; + } + + if (kFactory.Compile(srcFile, kMachine) != kExitOK) return 1; + } + + return kExitOK; +} + +// Last rev 8-1-24 diff --git a/dev/CompilerKit/src/Frontend/CCompilerPower64.cc b/dev/CompilerKit/src/Frontend/CCompilerPower64.cc new file mode 100644 index 0000000..aaa2308 --- /dev/null +++ b/dev/CompilerKit/src/Frontend/CCompilerPower64.cc @@ -0,0 +1,1303 @@ +/* + * ======================================================== + * + * CompilerPower64 + * Copyright (C) 2024-2025 Amlal El Mahrouss, all rights reserved. + * + * ======================================================== + */ + +#include <CompilerKit/detail/PowerPC.h> +#include <CompilerKit/Frontend.h> +#include <CompilerKit/UUID.h> +#include <CompilerKit/utils/CompilerUtils.h> +#include <cstdio> +#include <fstream> +#include <iostream> +#include <memory> +#include <random> +#include <string> +#include <utility> +#include <vector> + +#define kExitOK 0 + +/// @author EL Mahrouss Amlal (amlal@nekernel.org) +/// @file cc.cxx +/// @brief POWER64 C Compiler. + +///////////////////// + +/// ANSI ESCAPE CODES + +///////////////////// + +#define kBlank "\e[0;30m" +#define kRed "\e[0;31m" +#define kWhite "\e[0;97m" + +///////////////////////////////////// + +/// INTERNAL STRUCTURES OF THE C COMPILER + +///////////////////////////////////// + +namespace Detail { +// \brief name to register struct. +struct CompilerRegisterMap final { + std::string fName; + std::string fReg; +}; + +// \brief Map for C structs +// \author amlel +struct CompilerStructMap final { + /// 'struct::my_foo' + std::string fName; + + /// if instance: stores a valid register. + std::string fReg; + + /// offset count + std::size_t fOffsetsCnt; + + /// offset array. + std::vector<std::pair<Int32, std::string>> fOffsets; +}; + +struct CompilerState final { + std::vector<CompilerKit::SyntaxLeafList> fSyntaxTreeList; + std::vector<CompilerRegisterMap> kStackFrame; + std::vector<CompilerStructMap> kStructMap; + CompilerKit::SyntaxLeafList* fSyntaxTree{nullptr}; + std::unique_ptr<std::ofstream> fOutputAssembly; + std::string fLastFile; + std::string fLastError; + bool fVerbose; +}; +} // namespace Detail + +static Detail::CompilerState kState; +static std::string kIfFunction = ""; + +namespace Detail { +/// @brief prints an error into stdout. +/// @param reason the reason of the error. +/// @param file where does it originate from? +void print_error(std::string reason, std::string file) noexcept; + +struct CompilerType final { + std::string fName; + std::string fValue; +}; +} // namespace Detail + +///////////////////////////////////////////////////////////////////////////////////////// + +// Target architecture. +static int kMachine = 0; + +///////////////////////////////////////// + +// REGISTERS ACCORDING TO USED ASSEMBLER + +///////////////////////////////////////// + +static size_t kRegisterCnt = kAsmRegisterLimit; +static size_t kStartUsable = 2; +static size_t kUsableLimit = 15; +static size_t kRegisterCounter = kStartUsable; +static std::string kRegisterPrefix = kAsmRegisterPrefix; + +///////////////////////////////////////// + +// COMPILER PARSING UTILITIES/STATES. + +///////////////////////////////////////// + +static std::vector<std::string> kFileList; +static CompilerKit::AssemblyFactory kFactory; +static bool kInStruct = false; +static bool kOnWhileLoop = false; +static bool kOnForLoop = false; +static bool kInBraces = false; +static bool kIfFound = false; +static size_t kBracesCount = 0UL; + +/* @brief C compiler backend for C */ +class CompilerFrontendPower64 final : public CompilerKit::CompilerFrontendInterface { + public: + explicit CompilerFrontendPower64() = default; + ~CompilerFrontendPower64() override = default; + + LIBCOMPILER_COPY_DEFAULT(CompilerFrontendPower64); + + std::string Check(const char* text, const char* file); + CompilerKit::SyntaxLeafList::SyntaxLeaf Compile(std::string text, std::string file) override; + + const char* Language() override { return "POWER C"; } +}; + +static CompilerFrontendPower64* kCompilerFrontend = nullptr; +static std::vector<Detail::CompilerType> kCompilerVariables; +static std::vector<std::string> kCompilerFunctions; +static std::vector<Detail::CompilerType> kCompilerTypes; + +namespace Detail { +union number_cast final { + public: + number_cast(UInt64 _Raw) : _Raw(_Raw) {} + + public: + char _Num[8]; + UInt64 _Raw; +}; + +union double_cast final { + public: + double_cast(float _Raw) : _Raw(_Raw) {} + + public: + char _Sign; + char _Lh[8]; + char _Rh[23]; + + float _Raw; +}; +} // namespace Detail + +///////////////////////////////////////////////////////////////////////////////////////// + +// @name Compile +// @brief Generate MASM from a C assignement. + +///////////////////////////////////////////////////////////////////////////////////////// + +CompilerKit::SyntaxLeafList::SyntaxLeaf CompilerFrontendPower64::Compile(std::string text_, std::string file) { + std::string text = text_; + + bool typeFound = false; + bool fnFound = false; + + // setup generator. + std::random_device rd; + + auto seed_data = std::array<int, std::mt19937::state_size>{}; + std::generate(std::begin(seed_data), std::end(seed_data), std::ref(rd)); + std::seed_seq seq(std::begin(seed_data), std::end(seed_data)); + std::mt19937 generator(seq); + + // start parsing + for (size_t text_index = 0; text_index < text.size(); ++text_index) { + auto syntax_leaf = CompilerKit::SyntaxLeafList::SyntaxLeaf(); + + auto gen = uuids::uuid_random_generator{generator}; + uuids::uuid out = gen(); + + Detail::number_cast time_off = (UInt64) out.as_bytes().data(); + + if (!typeFound) { + auto substr = text.substr(text_index); + std::string match_type; + + for (size_t y = 0; y < substr.size(); ++y) { + if (substr[y] == ' ') { + while (match_type.find(' ') != std::string::npos) { + match_type.erase(match_type.find(' ')); + } + + for (auto& clType : kCompilerTypes) { + if (clType.fName == match_type) { + match_type.clear(); + + std::string buf; + + buf += clType.fValue; + buf += ' '; + + if (substr.find('=') != std::string::npos) { + break; + } + + if (text.find('(') != std::string::npos) { + syntax_leaf.fUserValue = buf; + + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + } + + typeFound = true; + break; + } + } + + break; + } + + match_type += substr[y]; + } + } + + if (text[text_index] == '{') { + if (kInStruct) { + continue; + } + + kInBraces = true; + ++kBracesCount; + + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + } + + // return keyword handler + if (text[text_index] == 'r') { + std::string return_keyword; + return_keyword += "return"; + + std::size_t index = 0UL; + + std::string value; + + for (size_t return_index = text_index; return_index < text.size(); ++return_index) { + if (text[return_index] != return_keyword[index]) { + for (size_t value_index = return_index; value_index < text.size(); ++value_index) { + if (text[value_index] == ';') break; + + value += text[value_index]; + } + + break; + } + + ++index; + } + + if (index == return_keyword.size()) { + if (!value.empty()) { + if (value.find('(') != std::string::npos) { + value.erase(value.find('(')); + } + + if (!isdigit(value[value.find('(') + 2])) { + std::string tmp = value; + bool reg_to_reg = false; + + value.clear(); + + value += " extern_segment"; + value += tmp; + } + + syntax_leaf.fUserValue = "\tmr r31, "; + + // make it pretty. + while (value.find('\t') != std::string::npos) value.erase(value.find('\t'), 1); + + while (value.find(' ') != std::string::npos) value.erase(value.find(' '), 1); + + while (value.find("extern_segment") != std::string::npos) + value.erase(value.find("extern_segment"), strlen("extern_segment")); + + bool found = false; + + for (auto& reg : kState.kStackFrame) { + if (value.find(reg.fName) != std::string::npos) { + found = true; + syntax_leaf.fUserValue += reg.fReg; + break; + } + } + + if (!found) syntax_leaf.fUserValue += "r0"; + } + + syntax_leaf.fUserValue += "\n\tblr"; + + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + + break; + } + } + + if (text[text_index] == 'i' && text[text_index + 1] == 'f') { + auto expr = text.substr(text_index + 2); + text.erase(text_index, 2); + + if (expr.find("{") != std::string::npos) { + expr.erase(expr.find("{")); + } + + if (expr.find("(") != std::string::npos) expr.erase(expr.find("(")); + + if (expr.find(")") != std::string::npos) expr.erase(expr.find(")")); + + kIfFunction = "__LIBCOMPILER_IF_PROC_"; + kIfFunction += std::to_string(time_off._Raw); + + syntax_leaf.fUserValue = + "\tcmpw " + "r10, r11"; + + syntax_leaf.fUserValue += "\n\tbeq extern_segment " + kIfFunction + + " \ndword public_segment .code64 " + kIfFunction + "\n"; + + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + + kIfFound = true; + } + + // Parse expressions and instructions here. + // what does this mean? + // we encounter an assignment, or we reached the end of an expression. + if (text[text_index] == '=' || text[text_index] == ';') { + if (fnFound) continue; + if (kIfFound) continue; + + if (text[text_index] == ';' && kInStruct) continue; + + if (text.find("typedef ") != std::string::npos) continue; + + if (text[text_index] == '=' && kInStruct) { + Detail::print_error("assignement of value inside a struct " + text, file); + continue; + } + + if (text[text_index] == ';' && kInStruct) { + bool space_found_ = false; + std::string sym; + + for (auto& ch : text) { + if (ch == ' ') { + space_found_ = true; + } + + if (ch == ';') break; + + if (space_found_) sym.push_back(ch); + } + + kState.kStructMap[kState.kStructMap.size() - 1].fOffsets.push_back( + std::make_pair(kState.kStructMap[kState.kStructMap.size() - 1].fOffsetsCnt + 4, sym)); + + kState.kStructMap[kState.kStructMap.size() - 1].fOffsetsCnt = + kState.kStructMap[kState.kStructMap.size() - 1].fOffsetsCnt + 4; + + continue; + } + + if (text[text_index] == '=' && kInStruct) { + continue; + } + + if (text[text_index + 1] == '=' || text[text_index - 1] == '!' || + text[text_index - 1] == '<' || text[text_index - 1] == '>') { + continue; + } + + std::string substr; + + if (text.find('=') != std::string::npos && kInBraces && !kIfFound) { + if (text.find("*") != std::string::npos) { + if (text.find("=") > text.find("*")) + substr += "\tli "; + else + substr += "\tli "; + } else { + substr += "\tli "; + } + } else if (text.find('=') != std::string::npos && !kInBraces) { + substr += "stw public_segment .data64 "; + } + + int first_encountered = 0; + + std::string str_name; + + for (size_t text_index_2 = 0; text_index_2 < text.size(); ++text_index_2) { + if (text[text_index_2] == '\"') { + ++text_index_2; + + // want to add this, so that the parser recognizes that this is a + // string. + substr += '"'; + + for (; text_index_2 < text.size(); ++text_index_2) { + if (text[text_index_2] == '\"') break; + + substr += text[text_index_2]; + } + } + + if (text[text_index_2] == '{' || text[text_index_2] == '}') continue; + + if (text[text_index_2] == ';') { + break; + } + + if (text[text_index_2] == ' ' || text[text_index_2] == '\t') { + if (first_encountered != 2) { + if (text[text_index] != '=' && + substr.find("public_segment .data64") == std::string::npos && !kInStruct) + substr += "public_segment .data64 "; + } + + ++first_encountered; + + continue; + } + + if (text[text_index_2] == '=') { + if (!kInBraces) { + substr.replace(substr.find("public_segment .data64"), strlen("public_segment .data64"), + "public_segment .zero64 "); + } + + substr += ","; + continue; + } + + substr += text[text_index_2]; + } + + for (auto& clType : kCompilerTypes) { + if (substr.find(clType.fName) != std::string::npos) { + if (substr.find(clType.fName) > substr.find('"')) continue; + + substr.erase(substr.find(clType.fName), clType.fName.size()); + } else if (substr.find(clType.fValue) != std::string::npos) { + if (substr.find(clType.fValue) > substr.find('"')) continue; + + if (clType.fName == "const") continue; + + substr.erase(substr.find(clType.fValue), clType.fValue.size()); + } + } + + if (substr.find("extern") != std::string::npos) { + substr.replace(substr.find("extern"), strlen("extern"), "extern_segment "); + + if (substr.find("public_segment .data64") != std::string::npos) + substr.erase(substr.find("public_segment .data64"), strlen("public_segment .data64")); + } + + auto var_to_find = std::find_if( + kCompilerVariables.cbegin(), kCompilerVariables.cend(), + [&](Detail::CompilerType type) { return type.fName.find(substr) != std::string::npos; }); + + kCompilerVariables.push_back({.fName = substr}); + + if (text[text_index] == ';') break; + + std::string reg = kAsmRegisterPrefix; + + ++kRegisterCounter; + reg += std::to_string(kRegisterCounter); + + auto newSubstr = substr.substr(substr.find(" ")); + + std::string symbol; + + for (size_t start = 0; start < newSubstr.size(); ++start) { + if (newSubstr[start] == ',') break; + + if (newSubstr[start] == ' ') continue; + + symbol += (newSubstr[start]); + } + + kState.kStackFrame.push_back({.fName = symbol, .fReg = reg}); + + syntax_leaf.fUserValue += "\n\tli " + reg + substr.substr(substr.find(',')); + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + } + + // function handler. + + if (text[text_index] == '(' && !fnFound && !kIfFound) { + std::string substr; + std::string args_buffer; + std::string args; + + bool type_crossed = false; + + for (size_t idx = text.find('(') + 1; idx < text.size(); ++idx) { + if (text[idx] == ',') continue; + + if (text[idx] == ' ') continue; + + if (text[idx] == ')') break; + } + + for (char substr_first_index : text) { + if (substr_first_index != ',') + args_buffer += substr_first_index; + else + args_buffer += '$'; + + if (substr_first_index == ';') { + args_buffer = args_buffer.erase(0, args_buffer.find('(')); + args_buffer = args_buffer.erase(args_buffer.find(';'), 1); + args_buffer = args_buffer.erase(args_buffer.find(')'), 1); + args_buffer = args_buffer.erase(args_buffer.find('('), 1); + + if (!args_buffer.empty()) args += "\tldw r6, "; + + std::string register_type; + std::size_t index = 7UL; + + while (args_buffer.find("$") != std::string::npos) { + register_type = kRegisterPrefix; + register_type += std::to_string(index); + + ++index; + + args_buffer.replace(args_buffer.find('$'), 1, "\n\tldw " + register_type + ","); + } + + args += args_buffer; + args += "\n\tli r31, "; + } + } + + for (char _text_i : text) { + if (_text_i == '\t' || _text_i == ' ') { + if (!type_crossed) { + substr.clear(); + type_crossed = true; + } + + continue; + } + + if (_text_i == '(') break; + + substr += _text_i; + } + + if (kInBraces) { + syntax_leaf.fUserValue = args; + syntax_leaf.fUserValue += substr; + + syntax_leaf.fUserValue += "\n\tblr\n"; + + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + + fnFound = true; + } else { + syntax_leaf.fUserValue.clear(); + + syntax_leaf.fUserValue += "public_segment .code64 "; + + syntax_leaf.fUserValue += substr; + syntax_leaf.fUserValue += "\n"; + + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + + fnFound = true; + } + + kCompilerFunctions.push_back(text); + } + + if (text[text_index] == '-' && text[text_index + 1] == '-') { + text = text.replace(text.find("--"), strlen("--"), ""); + + for (int _text_i = 0; _text_i < text.size(); ++_text_i) { + if (text[_text_i] == '\t' || text[_text_i] == ' ') text.erase(_text_i, 1); + } + + syntax_leaf.fUserValue += "dec "; + syntax_leaf.fUserValue += text; + + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + break; + } + + if (text[text_index] == '}') { + kRegisterCounter = kStartUsable; + + --kBracesCount; + + if (kBracesCount < 1) { + kInBraces = false; + kBracesCount = 0; + } + + if (kIfFound) kIfFound = false; + + if (kInStruct) kInStruct = false; + + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + } + + syntax_leaf.fUserValue.clear(); + } + + auto syntax_leaf = CompilerKit::SyntaxLeafList::SyntaxLeaf(); + syntax_leaf.fUserValue = "\n"; + kState.fSyntaxTree->fLeafList.push_back(syntax_leaf); + + return syntax_leaf; +} + +static bool kShouldHaveBraces = false; +static std::string kFnName; + +std::string CompilerFrontendPower64::Check(const char* text, const char* file) { + std::string err_str; + std::string ln = text; + + if (ln.empty()) { + return err_str; + } + + bool non_ascii_found = false; + + for (int i = 0; i < ln.size(); ++i) { + if (isalnum(ln[i])) { + non_ascii_found = true; + break; + } + } + + if (kShouldHaveBraces && ln.find('{') != std::string::npos) { + kShouldHaveBraces = false; + } + + if (!non_ascii_found) return err_str; + + size_t string_index = 1UL; + + if (ln.find('\'') != std::string::npos) { + string_index = ln.find('\'') + 1; + + for (; string_index < ln.size(); ++string_index) { + if (ln[string_index] == '\'') { + if (ln[string_index + 1] != ';') { + ln.erase(string_index, 1); + } + + return err_str; + } + } + } else if (ln.find('"') != std::string::npos) { + string_index = ln.find('"') + 1; + + for (; string_index < ln.size(); ++string_index) { + if (ln[string_index] == '"') { + if (ln[string_index + 1] != ';') { + ln.erase(string_index, 1); + } else { + break; + } + } + } + } else if (ln.find('"') == std::string::npos && ln.find('\'') == std::string::npos) { + std::vector<std::string> forbidden_words; + + forbidden_words.push_back("\\"); + forbidden_words.push_back("?"); + forbidden_words.push_back("@"); + forbidden_words.push_back("~"); + forbidden_words.push_back("::"); + forbidden_words.push_back("--*"); + forbidden_words.push_back("*/"); + + // add them to avoid stupid mistakes. + forbidden_words.push_back("namespace"); + forbidden_words.push_back("class"); + forbidden_words.push_back("extern \"C\""); + + for (auto& forbidden : forbidden_words) { + if (ln.find(forbidden) != std::string::npos) { + err_str += "\nForbidden character detected: "; + err_str += forbidden; + + return err_str; + } + } + } + + struct CompilerVariableRange final { + std::string fBegin; + std::string fEnd; + }; + + const std::vector<CompilerVariableRange> variables_list = { + {.fBegin = "static ", .fEnd = "="}, {.fBegin = "=", .fEnd = ";"}, + {.fBegin = "if(", .fEnd = "="}, {.fBegin = "if (", .fEnd = "="}, + {.fBegin = "if(", .fEnd = "<"}, {.fBegin = "if (", .fEnd = "<"}, + {.fBegin = "if(", .fEnd = ">"}, {.fBegin = "if (", .fEnd = ">"}, + {.fBegin = "if(", .fEnd = ")"}, {.fBegin = "if (", .fEnd = ")"}, + + {.fBegin = "else(", .fEnd = "="}, {.fBegin = "else (", .fEnd = "="}, + {.fBegin = "else(", .fEnd = "<"}, {.fBegin = "else (", .fEnd = "<"}, + {.fBegin = "else(", .fEnd = ">"}, {.fBegin = "else (", .fEnd = ">"}, + {.fBegin = "else(", .fEnd = ")"}, {.fBegin = "else (", .fEnd = ")"}, + }; + + for (auto& variable : variables_list) { + if (ln.find(variable.fBegin) != std::string::npos) { + string_index = ln.find(variable.fBegin) + variable.fBegin.size(); + + while (ln[string_index] == ' ') ++string_index; + + std::string keyword; + + for (; string_index < ln.size(); ++string_index) { + if (ln[string_index] == variable.fEnd[0]) { + std::string varname = ""; + + for (size_t index_keyword = ln.find(' '); ln[index_keyword] != variable.fBegin[0]; + ++index_keyword) { + if (ln[index_keyword] == ' ') { + continue; + } + + if (isdigit(ln[index_keyword])) { + goto cc_next_loop; + } + + varname += ln[index_keyword]; + } + + if (varname.find(' ') != std::string::npos) { + varname.erase(0, varname.find(' ')); + + if (variable.fBegin == "extern") { + varname.erase(0, varname.find(' ')); + } + } + + if (kRegisterCounter == 5 || kRegisterCounter == 6) ++kRegisterCounter; + + std::string reg = kAsmRegisterPrefix; + reg += std::to_string(kRegisterCounter); + + kCompilerVariables.push_back({.fValue = varname}); + goto cc_check_done; + } + + keyword.push_back(ln[string_index]); + } + + goto cc_next_loop; + + cc_check_done: + + // skip digit value. + if (isdigit(keyword[0]) || keyword[0] == '"') { + goto cc_next_loop; + } + + while (keyword.find(' ') != std::string::npos) keyword.erase(keyword.find(' '), 1); + + for (auto& var : kCompilerVariables) { + if (var.fValue.find(keyword) != std::string::npos) { + err_str.clear(); + goto cc_next; + } + } + + for (auto& fn : kCompilerFunctions) { + if (fn.find(keyword[0]) != std::string::npos) { + auto where_begin = fn.find(keyword[0]); + auto keyword_begin = 0UL; + auto failed = false; + + for (; where_begin < keyword.size(); ++where_begin) { + if (fn[where_begin] == '(' && keyword[keyword_begin] == '(') break; + + if (fn[where_begin] != keyword[keyword_begin]) { + failed = true; + break; + } + + ++keyword_begin; + } + + if (!failed) { + err_str.clear(); + goto cc_next; + } else { + continue; + } + } + } + + cc_error_value: + if (keyword.find("->") != std::string::npos) return err_str; + + if (keyword.find(".") != std::string::npos) return err_str; + + if (isalnum(keyword[0])) err_str += "\nUndefined value: " + keyword; + + return err_str; + } + + cc_next_loop: + continue; + } + +cc_next: + + // extern does not declare anything, it extern_segments a variable. + // so that's why it's not declare upper. + if (CompilerKit::find_word(ln, "extern")) { + auto substr = ln.substr(ln.find("extern") + strlen("extern")); + kCompilerVariables.push_back({.fValue = substr}); + } + + if (kShouldHaveBraces && ln.find('{') == std::string::npos) { + err_str += "Missing '{' for function "; + err_str += kFnName; + err_str += "\n"; + + kShouldHaveBraces = false; + kFnName.clear(); + } else if (kShouldHaveBraces && ln.find('{') != std::string::npos) { + kShouldHaveBraces = false; + kFnName.clear(); + } + + bool type_not_found = true; + + if (ln.find('\'') != std::string::npos) { + ln.replace(ln.find('\''), 3, "0"); + } + + auto first = ln.find('"'); + if (first != std::string::npos) { + auto second = 0UL; + bool found_second_quote = false; + + for (size_t i = first + 1; i < ln.size(); ++i) { + if (ln[i] == '\"') { + found_second_quote = true; + second = i; + + break; + } + } + + if (!found_second_quote) { + err_str += "Missing terminating \"."; + err_str += " here -> " + ln.substr(ln.find('"'), second); + } + } + + if (ln.find(')') != std::string::npos && ln.find(';') == std::string::npos) { + if (ln.find('{') == std::string::npos) { + kFnName = ln; + kShouldHaveBraces = true; + + goto skip_braces_check; + } else if (ln.find('{') != std::string::npos) { + kShouldHaveBraces = false; + } + } + +skip_braces_check: + + for (auto& key : kCompilerTypes) { + if (CompilerKit::find_word(ln, key.fName)) { + if (isdigit(ln[ln.find(key.fName) + key.fName.size() + 1])) { + err_str += "\nNumber cannot be set for "; + err_str += key.fName; + err_str += "'s name. here -> "; + err_str += ln; + } + + if (ln.find(key.fName) == 0 || ln[ln.find(key.fName) - 1] == ' ' || + ln[ln.find(key.fName) - 1] == '\t') { + type_not_found = false; + + if (ln[ln.find(key.fName) + key.fName.size()] != ' ') { + type_not_found = true; + + if (ln[ln.find(key.fName) + key.fName.size()] == '\t') type_not_found = false; + + goto next; + } else if (ln[ln.find(key.fName) + key.fName.size()] != '\t') { + type_not_found = true; + + if (ln[ln.find(key.fName) + key.fName.size()] == ' ') type_not_found = false; + } + } + + next: + + if (ln.find(';') == std::string::npos) { + if (ln.find('(') != std::string::npos) { + if (ln.find('=') == std::string::npos) continue; + } + + err_str += "\nMissing ';', here -> "; + err_str += ln; + } else { + continue; + } + + if (ln.find('=') != std::string::npos) { + if (ln.find('(') != std::string::npos) { + if (ln.find(')') == std::string::npos) { + err_str += "\nMissing ')', after '(' here -> "; + err_str += ln.substr(ln.find('(')); + } + } + } + } + } + + if (kInBraces && ln.find("struct") != std::string::npos && + ln.find("union") != std::string::npos && ln.find("enum") != std::string::npos && + ln.find('=') != std::string::npos) { + if (ln.find(';') == std::string::npos) { + err_str += "\nMissing ';' after struct/union/enum declaration, here -> "; + err_str += ln; + } + } + + if (ln.find(';') != std::string::npos && ln.find("for") == std::string::npos) { + if (ln.find(';') + 1 != ln.size()) { + for (int i = 0; i < ln.substr(ln.find(';') + 1).size(); ++i) { + if ((ln.substr(ln.find(';') + 1)[i] != ' ') || (ln.substr(ln.find(';') + 1)[i] != '\t')) { + if (auto err = this->Check(ln.substr(ln.find(';') + 1).c_str(), file); !err.empty()) { + err_str += "\nUnexpected text after ';' -> "; + err_str += ln.substr(ln.find(';')); + err_str += err; + } + } + } + } + } + + if (ln.find('(') != std::string::npos) { + if (ln.find(';') == std::string::npos && !CompilerKit::find_word(ln, "|") && + !CompilerKit::find_word(ln, "||") && !CompilerKit::find_word(ln, "&") && + !CompilerKit::find_word(ln, "&&") && !CompilerKit::find_word(ln, "~")) { + bool found_func = false; + size_t i = ln.find('('); + std::vector<char> opens; + std::vector<char> closes; + + for (; i < ln.size(); ++i) { + if (ln[i] == ')') { + closes.push_back(1); + } + + if (ln[i] == '(') { + opens.push_back(1); + } + } + + if (closes.size() != opens.size()) err_str += "Unterminated (), here -> " + ln; + + bool space_found = false; + + for (int i = 0; i < ln.size(); ++i) { + if (ln[i] == ')' && !space_found) { + space_found = true; + continue; + } + + if (space_found) { + if (ln[i] == ' ' && isalnum(ln[i + 1])) { + err_str += "\nBad function format here -> "; + err_str += ln; + } + } + } + } + + if (ln.find('(') < 1) { + err_str += "\nMissing identifier before '(' here -> "; + err_str += ln; + } else { + if (type_not_found && ln.find(';') == std::string::npos && + ln.find("if") == std::string::npos && ln.find("|") == std::string::npos && + ln.find("&") == std::string::npos && ln.find("(") == std::string::npos && + ln.find(")") == std::string::npos) { + err_str += "\n Missing ';' or type, here -> "; + err_str += ln; + } + } + + if (ln.find(')') == std::string::npos) { + err_str += "\nMissing ')', after '(' here -> "; + err_str += ln.substr(ln.find('(')); + } + } else { + if (ln.find("for") != std::string::npos || ln.find("while") != std::string::npos) { + err_str += "\nMissing '(', after \"for\", here -> "; + err_str += ln; + } + } + + if (ln.find('}') != std::string::npos && !kInBraces) { + if (!kInStruct && ln.find(';') == std::string::npos) { + err_str += "\nMismatched '}', here -> "; + err_str += ln; + } + } + + if (!ln.empty()) { + if (ln.find(';') == std::string::npos && ln.find('{') == std::string::npos && + ln.find('}') == std::string::npos && ln.find(')') == std::string::npos && + ln.find('(') == std::string::npos && ln.find(',') == std::string::npos) { + if (ln.size() <= 2) return err_str; + + err_str += "\nMissing ';', here -> "; + err_str += ln; + } + } + + return err_str; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +/** + * @brief C To Assembly mount-point. + */ + +///////////////////////////////////////////////////////////////////////////////////////// + +class AssemblyMountpointCLang final LC_ASSEMBLY_INTERFACE { + public: + explicit AssemblyMountpointCLang() = default; + ~AssemblyMountpointCLang() override = default; + + LIBCOMPILER_COPY_DEFAULT(AssemblyMountpointCLang); + + UInt32 Arch() noexcept override { return CompilerKit::AssemblyFactory::kArchPowerPC; } + + Int32 CompileToFormat(std::string src, Int32 arch) override { + if (kCompilerFrontend == nullptr) return 1; + + /* @brief copy contents wihtout extension */ + std::string src_file = src.data(); + std::ifstream src_fp = std::ifstream(src_file, std::ios::in); + std::string dest; + + for (auto& ch : src_file) { + if (ch == '.') { + break; + } + + dest += ch; + } + + /* According to PEF ABI. */ + std::vector<const char*> exts = kAsmFileExts; + dest += exts[4]; + + kState.fOutputAssembly = std::make_unique<std::ofstream>(dest); + + auto fmt = CompilerKit::current_date(); + + (*kState.fOutputAssembly) << "# Path: " << src_file << "\n"; + (*kState.fOutputAssembly) << "# Language: POWER Assembly (Generated from C)\n"; + (*kState.fOutputAssembly) << "# Date: " << fmt << "\n\n"; + + CompilerKit::SyntaxLeafList syntax; + + kState.fSyntaxTreeList.push_back(syntax); + kState.fSyntaxTree = &kState.fSyntaxTreeList[kState.fSyntaxTreeList.size() - 1]; + + std::string line_src; + + while (std::getline(src_fp, line_src)) { + if (auto err = kCompilerFrontend->Check(line_src.c_str(), src.data()); err.empty()) { + kCompilerFrontend->Compile(line_src, src.data()); + } else { + Detail::print_error(err, src.data()); + } + } + + if (kAcceptableErrors > 0) return 1; + + std::vector<std::string> keywords = {"ld", "stw", "add", "sub", "or"}; + + /// + /// Replace, optimize, fix assembly output. + /// + + for (auto& leaf : kState.fSyntaxTree->fLeafList) { + std::vector<std::string> access_keywords = {"->", "."}; + + for (auto& access_ident : access_keywords) { + if (CompilerKit::find_word(leaf.fUserValue, access_ident)) { + for (auto& struc : kState.kStructMap) { + /// TODO: + } + } + } + + for (auto& keyword : keywords) { + if (CompilerKit::find_word(leaf.fUserValue, keyword)) { + std::size_t cnt = 0UL; + + for (auto& reg : kState.kStackFrame) { + std::string needle; + + for (size_t i = 0; i < reg.fName.size(); i++) { + if (reg.fName[i] == ' ') { + ++i; + + for (; i < reg.fName.size(); i++) { + if (reg.fName[i] == ',') { + break; + } + + if (reg.fName[i] == ' ') continue; + + needle += reg.fName[i]; + } + + break; + } + } + + if (CompilerKit::find_word(leaf.fUserValue, needle)) { + if (leaf.fUserValue.find("extern_segment ") != std::string::npos) { + std::string range = "extern_segment "; + leaf.fUserValue.replace(leaf.fUserValue.find(range), range.size(), ""); + } + + if (leaf.fUserValue.find("ldw r6") != std::string::npos) { + std::string::difference_type countComma = + std::count(leaf.fUserValue.begin(), leaf.fUserValue.end(), ','); + + if (countComma == 1) { + leaf.fUserValue.replace(leaf.fUserValue.find("ldw"), strlen("ldw"), "mr"); + } + } + + leaf.fUserValue.replace(leaf.fUserValue.find(needle), needle.size(), reg.fReg); + + ++cnt; + } + } + + if (cnt > 1 && keyword != "mr" && keyword != "add" && keyword != "dec") { + leaf.fUserValue.replace(leaf.fUserValue.find(keyword), keyword.size(), "mr"); + } + } + } + } + + for (auto& leaf : kState.fSyntaxTree->fLeafList) { + (*kState.fOutputAssembly) << leaf.fUserValue; + } + + kState.fSyntaxTree = nullptr; + + kState.fOutputAssembly->flush(); + kState.fOutputAssembly.reset(); + + return kExitOK; + } +}; + +///////////////////////////////////////////////////////////////////////////////////////// + +#include <CompilerKit/Version.h> + +#define kPrintF printf +#define kSplashCxx() kPrintF(kWhite "cc, %s, (c) Amlal El Mahrouss\n", kDistVersion) + +static void cc_print_help() { + kSplashCxx(); +} + +///////////////////////////////////////////////////////////////////////////////////////// + +#define kExt ".c" + +LIBCOMPILER_MODULE(CompilerCLangPowerPC) { + ::signal(SIGSEGV, Detail::drvi_crash_handler); + + kCompilerTypes.push_back({.fName = "void", .fValue = "void"}); + kCompilerTypes.push_back({.fName = "char", .fValue = "byte"}); + kCompilerTypes.push_back({.fName = "short", .fValue = "hword"}); + kCompilerTypes.push_back({.fName = "int", .fValue = "dword"}); + kCompilerTypes.push_back({.fName = "long", .fValue = "qword"}); + kCompilerTypes.push_back({.fName = "*", .fValue = "offset"}); + + bool skip = false; + + kFactory.Mount(new AssemblyMountpointCLang()); + kMachine = CompilerKit::AssemblyFactory::kArchPowerPC; + kCompilerFrontend = new CompilerFrontendPower64(); + + for (auto index = 1UL; index < argc; ++index) { + if (skip) { + skip = false; + continue; + } + + if (argv[index][0] == '-') { + if (strcmp(argv[index], "-v") == 0 || strcmp(argv[index], "-version") == 0) { + kSplashCxx(); + return kExitOK; + } + + if (strcmp(argv[index], "-verbose") == 0) { + kState.fVerbose = true; + + continue; + } + + if (strcmp(argv[index], "-h") == 0 || strcmp(argv[index], "-help") == 0) { + cc_print_help(); + + return kExitOK; + } + + if (strcmp(argv[index], "-dialect") == 0) { + if (kCompilerFrontend) std::cout << kCompilerFrontend->Language() << "\n"; + + return kExitOK; + } + + if (strcmp(argv[index], "-fmax-exceptions") == 0) { + try { + kErrorLimit = std::strtol(argv[index + 1], nullptr, 10); + } + // catch anything here + catch (...) { + kErrorLimit = 0; + } + + skip = true; + + continue; + } + + std::string err = "Unknown command: "; + err += argv[index]; + + Detail::print_error(err, "cc"); + + continue; + } + + kFileList.emplace_back(argv[index]); + + std::string srcFile = argv[index]; + + if (strstr(argv[index], kExt) == nullptr) { + if (kState.fVerbose) { + Detail::print_error(srcFile + " is not a valid C source.\n", "cc"); + } + + return 1; + } + + if (kFactory.Compile(srcFile, kMachine) != kExitOK) return 1; + } + + return kExitOK; +} + +// Last rev 8-1-24 diff --git a/dev/CompilerKit/src/Frontend/CPlusPlusCompilerAMD64.cc b/dev/CompilerKit/src/Frontend/CPlusPlusCompilerAMD64.cc new file mode 100644 index 0000000..c38378a --- /dev/null +++ b/dev/CompilerKit/src/Frontend/CPlusPlusCompilerAMD64.cc @@ -0,0 +1,892 @@ +/* + * ======================================================== + * + * C++ Compiler Driver + * Copyright (C) 2024-2025 Amlal El Mahrouss, all rights reserved. + * + * ======================================================== + */ + +/// BUGS: 0 + +#define kPrintF printf + +#define kExitOK (EXIT_SUCCESS) +#define kExitNO (EXIT_FAILURE) + +#include <CompilerKit/detail/X64.h> +#include <CompilerKit/Frontend.h> +#include <CompilerKit/PEF.h> +#include <CompilerKit/UUID.h> +#include <CompilerKit/utils/CompilerUtils.h> + +/* NeKernel C++ Compiler Driver */ +/* This is part of the CompilerKit. */ +/* (c) Amlal El Mahrouss 2024-2025 */ + +/// @author EL Mahrouss Amlal (amlal@nekernel.org) +/// @file CPlusPlusCompilerAMD64.cxx +/// @brief Optimized C++ Compiler Driver. +/// @todo Throw error for scoped inside scoped variables when they get referenced outside. +/// @todo Add class/struct/enum support. + +/////////////////////// + +// ANSI ESCAPE CODES // + +/////////////////////// + +#define kBlank "\e[0;30m" +#define kRed "\e[0;31m" +#define kWhite "\e[0;97m" + +///////////////////////////////////// + +// INTERNALS OF THE C++ COMPILER + +///////////////////////////////////// + +/// @internal +namespace Detail { +std::filesystem::path expand_home(const std::filesystem::path& p) { + if (!p.empty() && p.string()[0] == '~') { + const char* home = std::getenv("HOME"); // For Unix-like systems + + if (!home) { + home = std::getenv("USERPROFILE"); // For Windows + } + + if (home) { + return std::filesystem::path(home) / p.relative_path().string().substr(1); + } else { + throw std::runtime_error("Home directory not found in environment variables"); + } + } + return p; +} + +struct CompilerRegisterMap final { + CompilerKit::STLString fName; + CompilerKit::STLString fReg; +}; + +/// \brief Offset based struct/class +struct CompilerStructMap final { + CompilerKit::STLString fName; + CompilerKit::STLString fReg; + std::vector<std::pair<UInt32, CompilerKit::STLString>> fOffsets; +}; + +/// \brief Compiler state structure. +struct CompilerState final { + std::vector<CompilerRegisterMap> fStackMapVector; + std::vector<CompilerStructMap> fStructMapVector; + CompilerKit::STLString fLastFile; + CompilerKit::STLString fLastError; +}; +} // namespace Detail + +static Detail::CompilerState kState; + +static Int32 kOnClassScope = 0; + +///////////////////////////////////////////////////////////////////////////////////////// + +// Target architecture. +static Int32 kMachine = CompilerKit::AssemblyFactory::kArchAMD64; + +///////////////////////////////////////// + +// ARGUMENTS REGISTERS (R8, R15) + +///////////////////////////////////////// + +static std::vector<CompilerKit::CompilerKeyword> kKeywords; + +///////////////////////////////////////// + +// COMPILER PARSING UTILITIES/STATES. + +///////////////////////////////////////// + +static CompilerKit::AssemblyFactory kFactory; +static Boolean kInStruct = false; +static Boolean kOnWhileLoop = false; +static Boolean kOnForLoop = false; +static Boolean kInBraces = false; +static size_t kBracesCount = 0UL; + +/* @brief C++ compiler backend for the NeKernel C++ driver */ +class CompilerFrontendCPlusPlusAMD64 final LC_COMPILER_FRONTEND { + public: + explicit CompilerFrontendCPlusPlusAMD64() = default; + ~CompilerFrontendCPlusPlusAMD64() override = default; + + LIBCOMPILER_COPY_DEFAULT(CompilerFrontendCPlusPlusAMD64); + + CompilerKit::SyntaxLeafList::SyntaxLeaf Compile(const CompilerKit::STLString text, + CompilerKit::STLString file) override; + + const char* Language() override; +}; + +/// @internal compiler variables + +static CompilerFrontendCPlusPlusAMD64* kCompilerFrontend = nullptr; + +static std::vector<CompilerKit::STLString> kRegisterMap; + +static std::vector<CompilerKit::STLString> kRegisterList = { + "rbx", "rsi", "r10", "r11", "r12", "r13", "r14", "r15", "xmm12", "xmm13", "xmm14", "xmm15", +}; + +/// @brief The PEF calling convention (caller must save rax, rbp) +/// @note callee must return via **rax**. +static std::vector<CompilerKit::STLString> kRegisterConventionCallList = { + "r8", "r9", "r10", "r11", "r12", "r13", "r14", "r15", +}; + +static std::size_t kFunctionEmbedLevel = 0UL; + +/// detail namespaces + +const char* CompilerFrontendCPlusPlusAMD64::Language() { + return "AMD64 C++"; +} + +static std::uintptr_t kOrigin = kPefBaseOrigin; +static std::vector<std::pair<CompilerKit::STLString, std::uintptr_t>> kOriginMap; + +///////////////////////////////////////////////////////////////////////////////////////// + +/// @name Compile +/// @brief Generate assembly from a C++ source. + +///////////////////////////////////////////////////////////////////////////////////////// + +CompilerKit::SyntaxLeafList::SyntaxLeaf CompilerFrontendCPlusPlusAMD64::Compile( + CompilerKit::STLString text, CompilerKit::STLString file) { + CompilerKit::SyntaxLeafList::SyntaxLeaf syntax_tree; + + if (text.length() < 1) return syntax_tree; + + std::size_t index = 0UL; + std::vector<std::pair<CompilerKit::CompilerKeyword, std::size_t>> keywords_list; + + for (auto& keyword : kKeywords) { + if (text.find(keyword.keyword_name) != std::string::npos) { + switch (keyword.keyword_kind) { + case CompilerKit::kKeywordKindCommentInline: { + break; + } + default: + break; + } + + std::size_t pos = text.find(keyword.keyword_name); + if (pos == std::string::npos) continue; + + // Safe guard: can't go before start of string + if (pos > 0 && text[pos - 1] == '+' && + keyword.keyword_kind == CompilerKit::kKeywordKindVariableAssign) + continue; + + if (pos > 0 && text[pos - 1] == '-' && + keyword.keyword_kind == CompilerKit::kKeywordKindVariableAssign) + continue; + + // Safe guard: don't go out of range + if ((pos + keyword.keyword_name.size()) < text.size() && + text[pos + keyword.keyword_name.size()] == '=' && + keyword.keyword_kind == CompilerKit::kKeywordKindVariableAssign) + continue; + + keywords_list.emplace_back(std::make_pair(keyword, index)); + ++index; + } + } + + for (auto& keyword : keywords_list) { + if (text.find(keyword.first.keyword_name) == CompilerKit::STLString::npos) continue; + + switch (keyword.first.keyword_kind) { + case CompilerKit::KeywordKind::kKeywordKindClass: { + ++kOnClassScope; + break; + } + case CompilerKit::KeywordKind::kKeywordKindIf: { + std::size_t keywordPos = text.find(keyword.first.keyword_name); + std::size_t openParen = text.find("(", keywordPos); + std::size_t closeParen = text.find(")", openParen); + + if (keywordPos == CompilerKit::STLString::npos || + openParen == CompilerKit::STLString::npos || + closeParen == CompilerKit::STLString::npos || closeParen <= openParen) { + Detail::print_error("Malformed if expression: " + text, file); + break; + } + + auto expr = text.substr(openParen + 1, closeParen - openParen - 1); + + if (expr.find(">=") != CompilerKit::STLString::npos) { + auto left = text.substr( + text.find(keyword.first.keyword_name) + keyword.first.keyword_name.size() + 2, + expr.find("<=") + strlen("<=")); + auto right = text.substr(expr.find(">=") + strlen(">="), text.find(")") - 1); + + size_t i = right.size() - 1; + + if (i < 1) break; + + try { + while (!std::isalnum(right[i])) { + right.erase(i, 1); + --i; + } + + right.erase(0, i); + } catch (...) { + right.erase(0, i); + } + + i = left.size() - 1; + try { + while (!std::isalnum(left[i])) { + left.erase(i, 1); + --i; + } + + left.erase(0, i); + } catch (...) { + left.erase(0, i); + } + + if (!isdigit(left[0]) || !isdigit(right[0])) { + auto indexRight = 0UL; + + auto& valueOfVar = !isdigit(left[0]) ? left : right; + + if (!valueOfVar.empty()) { + for (auto pairRight : kRegisterMap) { + ++indexRight; + + CompilerKit::STLString instr = "mov "; + + if (pairRight != valueOfVar) { + auto& valueOfVarOpposite = isdigit(left[0]) ? left : right; + + syntax_tree.fUserValue += + instr + kRegisterList[indexRight + 1] + ", " + valueOfVarOpposite + "\n"; + syntax_tree.fUserValue += "cmp " + kRegisterList[kRegisterMap.size() - 1] + "," + + kRegisterList[indexRight + 1] + "\n"; + + goto lc_done_iterarting_on_if; + } + + auto& valueOfVarOpposite = isdigit(left[0]) ? left : right; + + syntax_tree.fUserValue += + instr + kRegisterList[indexRight + 1] + ", " + valueOfVarOpposite + "\n"; + syntax_tree.fUserValue += "cmp " + kRegisterList[kRegisterMap.size() - 1] + ", " + + kRegisterList[indexRight + 1] + "\n"; + + break; + } + } + } + + lc_done_iterarting_on_if: + + CompilerKit::STLString symbol_name_fn = text; + + symbol_name_fn.erase(symbol_name_fn.find(keyword.first.keyword_name)); + + for (auto& ch : symbol_name_fn) { + if (ch == ' ') ch = '_'; + } + + syntax_tree.fUserValue += + "jge __OFFSET_ON_TRUE_LC\nsegment .code64 __OFFSET_ON_TRUE_LC:\n"; + } + + break; + } + case CompilerKit::KeywordKind::kKeywordKindFunctionStart: { + for (auto& ch : text) { + if (isdigit(ch)) { + goto dont_accept; + } + } + + goto accept; + + dont_accept: + break; + + accept: + CompilerKit::STLString symbol_name_fn = text; + size_t indexFnName = 0; + + // this one is for the type. + for (auto& ch : text) { + ++indexFnName; + + if (ch == '\t') break; + if (ch == ' ') break; + } + + symbol_name_fn = text.substr(indexFnName); + + if (text.find("return ") != CompilerKit::STLString::npos) { + text.erase(0, text.find("return ")); + break; + } + + if (text.ends_with(";") && text.find("return") == CompilerKit::STLString::npos) + goto lc_write_assembly; + else if (text.size() <= indexFnName) + Detail::print_error("Invalid function name: " + symbol_name_fn, file); + + indexFnName = 0; + + for (auto& ch : symbol_name_fn) { + if (ch == ' ' || ch == '\t') { + if (symbol_name_fn[indexFnName - 1] != ')') + Detail::print_error("Invalid function name: " + symbol_name_fn, file); + } + + ++indexFnName; + } + + if (symbol_name_fn.find("(") != CompilerKit::STLString::npos) { + symbol_name_fn.erase(symbol_name_fn.find("(")); + } + + syntax_tree.fUserValue = "public_segment .code64 __LIBCOMPILER_" + symbol_name_fn + "\n"; + ++kFunctionEmbedLevel; + + kOriginMap.push_back({"__LIBCOMPILER_" + symbol_name_fn, kOrigin}); + + break; + + lc_write_assembly: + auto it = + std::find_if(kOriginMap.begin(), kOriginMap.end(), + [&symbol_name_fn](std::pair<CompilerKit::STLString, std::uintptr_t> pair) -> bool { + return symbol_name_fn == pair.first; + }); + + if (it != kOriginMap.end()) { + std::stringstream ss; + ss << std::hex << it->second; + + syntax_tree.fUserValue = "jmp " + ss.str() + "\n"; + kOrigin += 1UL; + } + } + case CompilerKit::KeywordKind::kKeywordKindFunctionEnd: { + if (kOnClassScope) --kOnClassScope; + + if (text.ends_with(";")) break; + + --kFunctionEmbedLevel; + + if (kRegisterMap.size() > kRegisterList.size()) { + --kFunctionEmbedLevel; + } + + if (kFunctionEmbedLevel < 1) kRegisterMap.clear(); + + break; + } + case CompilerKit::KeywordKind::kKeywordKindEndInstr: + case CompilerKit::KeywordKind::kKeywordKindVariableInc: + case CompilerKit::KeywordKind::kKeywordKindVariableDec: + case CompilerKit::KeywordKind::kKeywordKindVariableAssign: { + CompilerKit::STLString valueOfVar = ""; + + if (keyword.first.keyword_kind == CompilerKit::KeywordKind::kKeywordKindVariableInc) { + valueOfVar = text.substr(text.find("+=") + 2); + } else if (keyword.first.keyword_kind == + CompilerKit::KeywordKind::kKeywordKindVariableDec) { + valueOfVar = text.substr(text.find("-=") + 2); + } else if (keyword.first.keyword_kind == + CompilerKit::KeywordKind::kKeywordKindVariableAssign) { + valueOfVar = text.substr(text.find("=") + 1); + } else if (keyword.first.keyword_kind == CompilerKit::KeywordKind::kKeywordKindEndInstr) { + break; + } + + while (valueOfVar.find(";") != CompilerKit::STLString::npos && + keyword.first.keyword_kind != CompilerKit::KeywordKind::kKeywordKindEndInstr) { + valueOfVar.erase(valueOfVar.find(";")); + } + + CompilerKit::STLString varName = text; + + if (keyword.first.keyword_kind == CompilerKit::KeywordKind::kKeywordKindVariableInc) { + varName.erase(varName.find("+=")); + } else if (keyword.first.keyword_kind == + CompilerKit::KeywordKind::kKeywordKindVariableDec) { + varName.erase(varName.find("-=")); + } else if (keyword.first.keyword_kind == + CompilerKit::KeywordKind::kKeywordKindVariableAssign) { + varName.erase(varName.find("=")); + } else if (keyword.first.keyword_kind == CompilerKit::KeywordKind::kKeywordKindEndInstr) { + varName.erase(varName.find(";")); + } + + static Boolean typeFound = false; + + for (auto& keyword : kKeywords) { + if (keyword.keyword_kind == CompilerKit::kKeywordKindType) { + if (text.find(keyword.keyword_name) != CompilerKit::STLString::npos) { + if (text[text.find(keyword.keyword_name)] == ' ') { + typeFound = false; + continue; + } + + typeFound = true; + } + } + } + + CompilerKit::STLString instr = "mov "; + + std::vector<CompilerKit::STLString> newVars; + + if (typeFound && + keyword.first.keyword_kind != CompilerKit::KeywordKind::kKeywordKindVariableInc && + keyword.first.keyword_kind != CompilerKit::KeywordKind::kKeywordKindVariableDec) { + if (kRegisterMap.size() > kRegisterList.size()) { + ++kFunctionEmbedLevel; + } + + while (varName.find(" ") != CompilerKit::STLString::npos) { + varName.erase(varName.find(" "), 1); + } + + while (varName.find("\t") != CompilerKit::STLString::npos) { + varName.erase(varName.find("\t"), 1); + } + + for (size_t i = 0; !isalnum(valueOfVar[i]); i++) { + if (i > valueOfVar.size()) break; + + valueOfVar.erase(i, 1); + } + + constexpr auto kTrueVal = "true"; + constexpr auto kFalseVal = "false"; + + if (valueOfVar == kTrueVal) { + valueOfVar = "1"; + } else if (valueOfVar == kFalseVal) { + valueOfVar = "0"; + } + + std::size_t indexRight = 0UL; + + for (auto pairRight : kRegisterMap) { + ++indexRight; + + if (pairRight != valueOfVar) { + if (valueOfVar[0] == '\"') { + syntax_tree.fUserValue = "segment .data64 __LIBCOMPILER_LOCAL_VAR_" + varName + + ": db " + valueOfVar + ", 0\n\n"; + syntax_tree.fUserValue += instr + kRegisterList[kRegisterMap.size() - 1] + ", " + + "__LIBCOMPILER_LOCAL_VAR_" + varName + "\n"; + kOrigin += 1UL; + } else { + syntax_tree.fUserValue = + instr + kRegisterList[kRegisterMap.size() - 1] + ", " + valueOfVar + "\n"; + kOrigin += 1UL; + } + + goto done; + } + } + + if (((int) indexRight - 1) < 0) { + if (valueOfVar[0] == '\"') { + syntax_tree.fUserValue = "segment .data64 __LIBCOMPILER_LOCAL_VAR_" + varName + + ": db " + valueOfVar + ", 0\n"; + syntax_tree.fUserValue += instr + kRegisterList[kRegisterMap.size()] + ", " + + "__LIBCOMPILER_LOCAL_VAR_" + varName + "\n"; + kOrigin += 1UL; + } else { + syntax_tree.fUserValue = + instr + kRegisterList[kRegisterMap.size()] + ", " + valueOfVar + "\n"; + kOrigin += 1UL; + } + + goto done; + } + + if (valueOfVar[0] != '\"' && valueOfVar[0] != '\'' && !isdigit(valueOfVar[0])) { + for (auto pair : kRegisterMap) { + if (pair == valueOfVar) goto done; + } + + Detail::print_error("Variable not declared: " + varName, file); + break; + } + + done: + for (auto& keyword : kKeywords) { + if (keyword.keyword_kind == CompilerKit::kKeywordKindType && + varName.find(keyword.keyword_name) != CompilerKit::STLString::npos) { + varName.erase(varName.find(keyword.keyword_name), keyword.keyword_name.size()); + break; + } + } + + newVars.push_back(varName); + + break; + } + + kRegisterMap.insert(kRegisterMap.end(), newVars.begin(), newVars.end()); + + if (keyword.second > 0 && + kKeywords[keyword.second - 1].keyword_kind == CompilerKit::kKeywordKindType || + kKeywords[keyword.second - 1].keyword_kind == CompilerKit::kKeywordKindTypePtr) { + syntax_tree.fUserValue = "\n"; + continue; + } + + if (keyword.first.keyword_kind == CompilerKit::KeywordKind::kKeywordKindEndInstr) { + syntax_tree.fUserValue = "\n"; + continue; + } + + if (keyword.first.keyword_kind == CompilerKit::KeywordKind::kKeywordKindVariableInc) { + instr = "add "; + } else if (keyword.first.keyword_kind == + CompilerKit::KeywordKind::kKeywordKindVariableDec) { + instr = "sub "; + } + + CompilerKit::STLString varErrCpy = varName; + + while (varName.find(" ") != CompilerKit::STLString::npos) { + varName.erase(varName.find(" "), 1); + } + + while (varName.find("\t") != CompilerKit::STLString::npos) { + varName.erase(varName.find("\t"), 1); + } + + std::size_t indxReg = 0UL; + + for (size_t i = 0; !isalnum(valueOfVar[i]); i++) { + if (i > valueOfVar.size()) break; + + valueOfVar.erase(i, 1); + } + + while (valueOfVar.find(" ") != CompilerKit::STLString::npos) { + valueOfVar.erase(valueOfVar.find(" "), 1); + } + + while (valueOfVar.find("\t") != CompilerKit::STLString::npos) { + valueOfVar.erase(valueOfVar.find("\t"), 1); + } + + constexpr auto kTrueVal = "true"; + constexpr auto kFalseVal = "false"; + + /// interpet boolean values, since we're on C++ + + if (valueOfVar == kTrueVal) { + valueOfVar = "1"; + } else if (valueOfVar == kFalseVal) { + valueOfVar = "0"; + } + + for (auto pair : kRegisterMap) { + ++indxReg; + + if (pair != varName) continue; + + std::size_t indexRight = 0ul; + + for (auto pairRight : kRegisterMap) { + ++indexRight; + + if (pairRight != varName) { + syntax_tree.fUserValue = + instr + kRegisterList[kRegisterMap.size()] + ", " + valueOfVar + "\n"; + kOrigin += 1UL; + continue; + } + + syntax_tree.fUserValue = + instr + kRegisterList[indexRight - 1] + ", " + valueOfVar + "\n"; + kOrigin += 1UL; + break; + } + + newVars.push_back(varName); + break; + } + + if (syntax_tree.fUserValue.empty()) { + Detail::print_error("Variable not declared: " + varName, file); + } + + kRegisterMap.insert(kRegisterMap.end(), newVars.begin(), newVars.end()); + + break; + } + case CompilerKit::KeywordKind::kKeywordKindReturn: { + try { + auto pos = text.find("return") + strlen("return") + 1; + CompilerKit::STLString subText = text.substr(pos); + subText = subText.erase(subText.find(";")); + size_t indxReg = 0UL; + + if (subText[0] != '\"' && subText[0] != '\'') { + if (!isdigit(subText[0])) { + for (auto pair : kRegisterMap) { + ++indxReg; + + if (pair != subText) continue; + + syntax_tree.fUserValue = "mov rax, " + kRegisterList[indxReg - 1] + "\nret\n"; + kOrigin += 1UL; + + break; + } + } else { + syntax_tree.fUserValue = "mov rax, " + subText + "\nret\n"; + kOrigin += 1UL; + + break; + } + } else { + syntax_tree.fUserValue = "__LIBCOMPILER_LOCAL_RETURN_STRING: db " + subText + + ", 0\nmov rcx, __LIBCOMPILER_LOCAL_RETURN_STRING\n"; + syntax_tree.fUserValue += "mov rax, rcx\nret\n"; + kOrigin += 1UL; + + break; + } + + if (syntax_tree.fUserValue.empty()) { + if (subText.find("(") != CompilerKit::STLString::npos) { + subText.erase(subText.find("(")); + + auto it = std::find_if( + kOriginMap.begin(), kOriginMap.end(), + [&subText](std::pair<CompilerKit::STLString, std::uintptr_t> pair) -> bool { + return pair.first.find(subText) != CompilerKit::STLString::npos; + }); + + if (it == kOriginMap.end()) + Detail::print_error("Invalid return value: " + subText, file); + + std::stringstream ss; + ss << it->second; + + syntax_tree.fUserValue = "jmp " + ss.str() + "\nret\n"; + kOrigin += 1UL; + break; + } + } + + break; + } catch (...) { + syntax_tree.fUserValue = "ret\n"; + kOrigin += 1UL; + } + } + default: { + continue; + } + } + } + + return syntax_tree; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +/** + * @brief C++ assembler class. + */ + +///////////////////////////////////////////////////////////////////////////////////////// + +class AssemblyCPlusPlusInterfaceAMD64 final LC_ASSEMBLY_INTERFACE { + public: + explicit AssemblyCPlusPlusInterfaceAMD64() = default; + ~AssemblyCPlusPlusInterfaceAMD64() override = default; + + LIBCOMPILER_COPY_DEFAULT(AssemblyCPlusPlusInterfaceAMD64); + + UInt32 Arch() noexcept override { return CompilerKit::AssemblyFactory::kArchAMD64; } + + Int32 CompileToFormat(CompilerKit::STLString src, Int32 arch) override { + if (kCompilerFrontend == nullptr) return kExitNO; + + CompilerKit::STLString dest = src; + dest += ".pp.masm"; + + std::ofstream out_fp(dest); + std::ifstream src_fp = std::ifstream(src + ".pp"); + + CompilerKit::STLString line_source; + + out_fp << "#bits 64\n"; + out_fp << "#org " << kOrigin << "\n\n"; + + while (std::getline(src_fp, line_source)) { + out_fp << kCompilerFrontend->Compile(line_source, src).fUserValue; + } + + return kExitOK; + } +}; + +///////////////////////////////////////////////////////////////////////////////////////// + +///////////////////////////////////////////////////////////////////////////////////////// + +#define kExtListCxx \ + { ".cpp", ".cxx", ".cc", ".c++", ".cp" } + +LIBCOMPILER_MODULE(CompilerCPlusPlusAMD64) { + Boolean skip = false; + + kKeywords.emplace_back("if", CompilerKit::kKeywordKindIf); + kKeywords.emplace_back("else", CompilerKit::kKeywordKindElse); + kKeywords.emplace_back("else if", CompilerKit::kKeywordKindElseIf); + + kKeywords.emplace_back("class", CompilerKit::kKeywordKindClass); + kKeywords.emplace_back("struct", CompilerKit::kKeywordKindClass); + kKeywords.emplace_back("namespace", CompilerKit::kKeywordKindNamespace); + kKeywords.emplace_back("typedef", CompilerKit::kKeywordKindTypedef); + kKeywords.emplace_back("using", CompilerKit::kKeywordKindTypedef); + kKeywords.emplace_back("{", CompilerKit::kKeywordKindBodyStart); + kKeywords.emplace_back("}", CompilerKit::kKeywordKindBodyEnd); + kKeywords.emplace_back("auto", CompilerKit::kKeywordKindVariable); + kKeywords.emplace_back("int", CompilerKit::kKeywordKindType); + kKeywords.emplace_back("bool", CompilerKit::kKeywordKindType); + kKeywords.emplace_back("unsigned", CompilerKit::kKeywordKindType); + kKeywords.emplace_back("short", CompilerKit::kKeywordKindType); + kKeywords.emplace_back("char", CompilerKit::kKeywordKindType); + kKeywords.emplace_back("long", CompilerKit::kKeywordKindType); + kKeywords.emplace_back("float", CompilerKit::kKeywordKindType); + kKeywords.emplace_back("double", CompilerKit::kKeywordKindType); + kKeywords.emplace_back("void", CompilerKit::kKeywordKindType); + + kKeywords.emplace_back("auto*", CompilerKit::kKeywordKindVariablePtr); + kKeywords.emplace_back("int*", CompilerKit::kKeywordKindTypePtr); + kKeywords.emplace_back("bool*", CompilerKit::kKeywordKindTypePtr); + kKeywords.emplace_back("unsigned*", CompilerKit::kKeywordKindTypePtr); + kKeywords.emplace_back("short*", CompilerKit::kKeywordKindTypePtr); + kKeywords.emplace_back("char*", CompilerKit::kKeywordKindTypePtr); + kKeywords.emplace_back("long*", CompilerKit::kKeywordKindTypePtr); + kKeywords.emplace_back("float*", CompilerKit::kKeywordKindTypePtr); + kKeywords.emplace_back("double*", CompilerKit::kKeywordKindTypePtr); + kKeywords.emplace_back("void*", CompilerKit::kKeywordKindTypePtr); + + kKeywords.emplace_back("(", CompilerKit::kKeywordKindFunctionStart); + kKeywords.emplace_back(")", CompilerKit::kKeywordKindFunctionEnd); + kKeywords.emplace_back("=", CompilerKit::kKeywordKindVariableAssign); + kKeywords.emplace_back("+=", CompilerKit::kKeywordKindVariableInc); + kKeywords.emplace_back("-=", CompilerKit::kKeywordKindVariableDec); + kKeywords.emplace_back("const", CompilerKit::kKeywordKindConstant); + kKeywords.emplace_back("*", CompilerKit::kKeywordKindPtr); + kKeywords.emplace_back("->", CompilerKit::kKeywordKindPtrAccess); + kKeywords.emplace_back(".", CompilerKit::kKeywordKindAccess); + kKeywords.emplace_back(",", CompilerKit::kKeywordKindArgSeparator); + kKeywords.emplace_back(";", CompilerKit::kKeywordKindEndInstr); + kKeywords.emplace_back(":", CompilerKit::kKeywordKindSpecifier); + kKeywords.emplace_back("public:", CompilerKit::kKeywordKindSpecifier); + kKeywords.emplace_back("private:", CompilerKit::kKeywordKindSpecifier); + kKeywords.emplace_back("protected:", CompilerKit::kKeywordKindSpecifier); + kKeywords.emplace_back("final", CompilerKit::kKeywordKindSpecifier); + kKeywords.emplace_back("return", CompilerKit::kKeywordKindReturn); + kKeywords.emplace_back("/*", CompilerKit::kKeywordKindCommentMultiLineStart); + kKeywords.emplace_back("*/", CompilerKit::kKeywordKindCommentMultiLineEnd); + kKeywords.emplace_back("//", CompilerKit::kKeywordKindCommentInline); + kKeywords.emplace_back("==", CompilerKit::kKeywordKindEq); + kKeywords.emplace_back("!=", CompilerKit::kKeywordKindNotEq); + kKeywords.emplace_back(">=", CompilerKit::kKeywordKindGreaterEq); + kKeywords.emplace_back("<=", CompilerKit::kKeywordKindLessEq); + + kErrorLimit = 0; + + kCompilerFrontend = new CompilerFrontendCPlusPlusAMD64(); + kFactory.Mount(new AssemblyCPlusPlusInterfaceAMD64()); + + CompilerKit::install_signal(SIGSEGV, Detail::drvi_crash_handler); + + for (auto index = 1UL; index < argc; ++index) { + if (!argv[index]) break; + + if (argv[index][0] == '-') { + if (skip) { + skip = false; + continue; + } + + if (strcmp(argv[index], "-cxx-verbose") == 0) { + kVerbose = true; + + continue; + } + + if (strcmp(argv[index], "-cxx-dialect") == 0) { + if (kCompilerFrontend) std::cout << kCompilerFrontend->Language() << "\n"; + + return LIBCOMPILER_SUCCESS; + } + + if (strcmp(argv[index], "-cxx-max-err") == 0) { + try { + kErrorLimit = std::strtol(argv[index + 1], nullptr, 10); + } + // catch anything here + catch (...) { + kErrorLimit = 0; + } + + skip = true; + + continue; + } + + CompilerKit::STLString err = "Unknown option: "; + err += argv[index]; + + Detail::print_error(err, "cxxdrv"); + + continue; + } + + CompilerKit::STLString argv_i = argv[index]; + + std::vector<CompilerKit::STLString> exts = kExtListCxx; + + for (CompilerKit::STLString ext : exts) { + if (argv_i.ends_with(ext)) { + if (kFactory.Compile(argv_i, kMachine) != kExitOK) { + return LIBCOMPILER_INVALID_DATA; + } + + break; + } + } + } + + kFactory.Unmount(); + + return LIBCOMPILER_SUCCESS; +} + +// +// Last rev 23-5-25 +// diff --git a/dev/CompilerKit/src/Linker/DynamicLinker64PEF.cc b/dev/CompilerKit/src/Linker/DynamicLinker64PEF.cc new file mode 100644 index 0000000..3fde11b --- /dev/null +++ b/dev/CompilerKit/src/Linker/DynamicLinker64PEF.cc @@ -0,0 +1,672 @@ +/* ------------------------------------------- + + Copyright (C) 2024-2025 Amlal El Mahrouss, all rights reserved + + @file DynamicLinker64PEF.cc + @brief: C++ 64-Bit PEF Linker for NeKernel.org + +------------------------------------------- */ + +/// @author EL Mahrouss Amlal (amlal@nekernel.org) +/// @brief NeKernel.org 64-bit PEF Linker. +/// Last Rev: Sat Apr 19 CET 2025 +/// @note Do not look up for anything with .code64/.data64/.zero64! +/// It will be loaded when the program loader will start the image. + + +#include <CompilerKit/Defines.h> +#include <CompilerKit/ErrorID.h> +#include <CompilerKit/CodeGen.h> +#include <CompilerKit/PEF.h> +#include <CompilerKit/UUID.h> +#include <CompilerKit/Version.h> +#include <CompilerKit/AE.h> +#include <CompilerKit/utils/CompilerUtils.h> + +#define kLinkerVersionStr \ + "NeKernel.org 64-Bit Linker (Preferred Executable Format) %s, (c) Amlal El Mahrouss " \ + "2024-2025 " \ + "all rights reserved.\n" + +#define MemoryCopy(DST, SRC, SZ) memcpy(DST, SRC, SZ) +#define StringCompare(DST, SRC) strcmp(DST, SRC) + +#define kPefNoCpu (0U) +#define kPefNoSubCpu (0U) + +#define kLinkerDefaultOrigin kPefBaseOrigin +#define kLinkerId (0x5046FF) +#define kLinkerAbiContainer "__PEFContainer:ABI:" + +#define kPrintF printf +#define kLinkerSplash() kConsoleOut << std::printf(kLinkerVersionStr, kDistVersion) + +/// @brief PEF stack size symbol. +#define kLinkerStackSizeSymbol "__PEFSizeOfReserveStack" + +#define kConsoleOut \ + (std::cout << "\e[0;31m" \ + << "ld64: " \ + << "\e[0;97m") + +enum { + kABITypeNull = 0, + kABITypeStart = 0x1010, /* The start of ABI list. */ + kABITypeNE = 0x5046, /* PF (NeKernel.org's PEF ABI) */ + kABITypeInvalid = 0xFFFF, +}; + +static CompilerKit::STLString kOutput = "a" kPefExt; +static Int32 kAbi = kABITypeNE; +static Int32 kSubArch = kPefNoSubCpu; +static Int32 kArch = CompilerKit::kPefArchInvalid; +static Bool kFatBinaryEnable = false; +static Bool kStartFound = false; +static Bool kDuplicateSymbols = false; + +/* ld64 is to be found, mld is to be found at runtime. */ +static const Char* kLdDefineSymbol = ":UndefinedSymbol:"; +static const Char* kLdDynamicSym = ":RuntimeSymbol:"; + +/* object code and list. */ +static std::vector<CompilerKit::STLString> kObjectList; +static std::vector<Detail::DynamicLinkerBlob> kObjectBytes; + +/// @brief NE 64-bit Linker. +/// @note This linker is made for PEF executable, thus NE based OSes. +LIBCOMPILER_MODULE(DynamicLinker64PEF) { + bool is_executable = true; + + ::signal(SIGSEGV, Detail::drvi_crash_handler); + + /** + * @brief parse flags and trigger options. + */ + for (size_t linker_arg = 1; linker_arg < argc; ++linker_arg) { + if (StringCompare(argv[linker_arg], "-help") == 0) { + kLinkerSplash(); + + kConsoleOut << "-version: Show linker version.\n"; + kConsoleOut << "-help: Show linker help.\n"; + kConsoleOut << "-verbose: Enable linker trace.\n"; + kConsoleOut << "-dylib: Output as a Dynamic PEF.\n"; + kConsoleOut << "-fat: Output as a FAT PEF.\n"; + kConsoleOut << "-32k: Output as a 32x0 PEF.\n"; + kConsoleOut << "-64k: Output as a 64x0 PEF.\n"; + kConsoleOut << "-amd64: Output as a AMD64 PEF.\n"; + kConsoleOut << "-rv64: Output as a RISC-V PEF.\n"; + kConsoleOut << "-power64: Output as a POWER PEF.\n"; + kConsoleOut << "-arm64: Output as a ARM64 PEF.\n"; + kConsoleOut << "-output: Select the output file name.\n"; + + return LIBCOMPILER_SUCCESS; + } else if (StringCompare(argv[linker_arg], "-version") == 0) { + kLinkerSplash(); + + return LIBCOMPILER_SUCCESS; + } else if (StringCompare(argv[linker_arg], "-fat") == 0) { + kFatBinaryEnable = true; + + continue; + } else if (StringCompare(argv[linker_arg], "-64k") == 0) { + kArch = CompilerKit::kPefArch64000; + + continue; + } else if (StringCompare(argv[linker_arg], "-amd64") == 0) { + kArch = CompilerKit::kPefArchAMD64; + + continue; + } else if (StringCompare(argv[linker_arg], "-32k") == 0) { + kArch = CompilerKit::kPefArch32000; + + continue; + } else if (StringCompare(argv[linker_arg], "-power64") == 0) { + kArch = CompilerKit::kPefArchPowerPC; + + continue; + } else if (StringCompare(argv[linker_arg], "-riscv64") == 0) { + kArch = CompilerKit::kPefArchRISCV; + + continue; + } else if (StringCompare(argv[linker_arg], "-arm64") == 0) { + kArch = CompilerKit::kPefArchARM64; + + continue; + } else if (StringCompare(argv[linker_arg], "-verbose") == 0) { + kVerbose = true; + + continue; + } else if (StringCompare(argv[linker_arg], "-dylib") == 0) { + if (kOutput.empty()) { + continue; + } + + if (kOutput.find(kPefExt) != CompilerKit::STLString::npos) + kOutput.erase(kOutput.find(kPefExt), strlen(kPefExt)); + + kOutput += kPefDylibExt; + + is_executable = false; + + continue; + } else if (StringCompare(argv[linker_arg], "-output") == 0) { + if ((linker_arg + 1) > argc) continue; + + kOutput = argv[linker_arg + 1]; + ++linker_arg; + + continue; + } else { + if (argv[linker_arg][0] == '-') { + kConsoleOut << "unknown flag: " << argv[linker_arg] << "\n"; + return EXIT_FAILURE; + } + + kObjectList.emplace_back(argv[linker_arg]); + + continue; + } + } + + if (kOutput.empty()) { + kConsoleOut << "no output filename set." << std::endl; + return LIBCOMPILER_EXEC_ERROR; + } else if (kObjectList.empty()) { + kConsoleOut << "no input files." << std::endl; + return LIBCOMPILER_EXEC_ERROR; + } else { + namespace FS = std::filesystem; + + // check for existing files, if they don't throw an error. + for (auto& obj : kObjectList) { + if (!FS::exists(obj)) { + // if filesystem doesn't find file + // -> throw error. + kConsoleOut << "no such file: " << obj << std::endl; + return LIBCOMPILER_EXEC_ERROR; + } + } + } + + // PEF expects a valid target architecture when outputing a binary. + if (kArch == CompilerKit::kPefArchInvalid) { + kConsoleOut << "no target architecture set, can't continue." << std::endl; + return LIBCOMPILER_EXEC_ERROR; + } + + CompilerKit::PEFContainer pef_container{}; + + int32_t archs = kArch; + + pef_container.Count = 0UL; + pef_container.Kind = is_executable ? CompilerKit::kPefKindExec : CompilerKit::kPefKindDylib; + pef_container.SubCpu = kSubArch; + pef_container.Linker = kLinkerId; // Amlal El Mahrouss Linker + pef_container.Abi = kAbi; // Multi-Processor UX ABI + pef_container.Magic[0] = kPefMagic[kFatBinaryEnable ? 2 : 0]; + pef_container.Magic[1] = kPefMagic[1]; + pef_container.Magic[2] = kPefMagic[kFatBinaryEnable ? 0 : 2]; + pef_container.Magic[3] = kPefMagic[3]; + pef_container.Version = kPefVersion; + + // specify the start address, can be 0x10000 + pef_container.Start = kLinkerDefaultOrigin; + pef_container.HdrSz = sizeof(CompilerKit::PEFContainer); + pef_container.Checksum = 0UL; + + std::ofstream output_fc(kOutput, std::ofstream::binary); + + if (output_fc.bad()) { + if (kVerbose) { + kConsoleOut << "error: " << strerror(errno) << "\n"; + } + + return LIBCOMPILER_FILE_NOT_FOUND; + } + + //! Read AE to convert as PEF. + + std::vector<CompilerKit::PEFCommandHeader> command_headers; + CompilerKit::Utils::AEReadableProtocol reader_protocol{}; + + for (const auto& objectFile : kObjectList) { + if (!std::filesystem::exists(objectFile)) continue; + + CompilerKit::AEHeader hdr{}; + + reader_protocol._Fp = std::ifstream(objectFile, std::ifstream::binary); + reader_protocol._Fp >> hdr; + + if (hdr.fMagic[0] == kAEMag0 && hdr.fMagic[1] == kAEMag1 && + hdr.fSize == sizeof(CompilerKit::AEHeader)) { + if (hdr.fArch != kArch) { + if (kVerbose) kConsoleOut << "is this a FAT binary? : "; + + if (!kFatBinaryEnable) { + if (kVerbose) kConsoleOut << "not a FAT binary.\n"; + + kConsoleOut << "object " << objectFile + << " is a different kind of architecture and output isn't " + "treated as a FAT binary." + << std::endl; + + return LIBCOMPILER_FAT_ERROR; + } else { + if (kVerbose) { + kConsoleOut << "Architecture matches what we expect.\n"; + } + } + } + + // append arch type to archs varaible. + archs |= hdr.fArch; + std::size_t cnt = hdr.fCount; + + if (kVerbose) kConsoleOut << "header found, record count: " << cnt << "\n"; + + pef_container.Count = cnt; + + char_type* raw_ae_records = new char_type[cnt * sizeof(CompilerKit::AERecordHeader)]; + + if (!raw_ae_records) { + if (kVerbose) kConsoleOut << "allocation failed for records of count: " << cnt << "\n"; + } + + std::memset(raw_ae_records, 0, cnt * sizeof(CompilerKit::AERecordHeader)); + + auto* ae_records = reader_protocol.Read(raw_ae_records, cnt); + + size_t org = kLinkerDefaultOrigin; + + for (size_t ae_record_index = 0; ae_record_index < cnt; ++ae_record_index) { + CompilerKit::PEFCommandHeader command_header{0}; + std::size_t offset_of_obj = ae_records[ae_record_index].fOffset; + + MemoryCopy(command_header.Name, ae_records[ae_record_index].fName, kPefNameLen); + + CompilerKit::STLString cmd_hdr_name(command_header.Name); + + // check this header if it's any valid. + if (cmd_hdr_name.find(kPefCode64) == CompilerKit::STLString::npos && + cmd_hdr_name.find(kPefData64) == CompilerKit::STLString::npos && + cmd_hdr_name.find(kPefZero64) == CompilerKit::STLString::npos) { + if (cmd_hdr_name.find(kPefStart) == CompilerKit::STLString::npos && + *command_header.Name == 0) { + if (cmd_hdr_name.find(kLdDefineSymbol) != CompilerKit::STLString::npos) { + goto ld_mark_header; + } else { + continue; + } + } + } + + if (cmd_hdr_name.find(kPefStart) != CompilerKit::STLString::npos && + cmd_hdr_name.find(kPefCode64) != CompilerKit::STLString::npos) { + kStartFound = true; + } + + ld_mark_header: + command_header.Offset = offset_of_obj; + command_header.Kind = ae_records[ae_record_index].fKind; + command_header.Size = ae_records[ae_record_index].fSize; + command_header.Cpu = hdr.fArch; + command_header.VMAddress = org; /// TODO: + command_header.SubCpu = hdr.fSubArch; + + org += command_header.Size; + + if (kVerbose) { + kConsoleOut << "Record: " << ae_records[ae_record_index].fName << " is marked.\n"; + + kConsoleOut << "Offset: " << command_header.Offset << "\n"; + } + + command_headers.emplace_back(command_header); + } + + delete[] raw_ae_records; + raw_ae_records = nullptr; + + std::vector<char> bytes; + bytes.resize(hdr.fCodeSize); + + reader_protocol._Fp.seekg(std::streamsize(hdr.fStartCode)); + reader_protocol._Fp.read(bytes.data(), std::streamsize(hdr.fCodeSize)); + + kObjectBytes.push_back({.mBlob = bytes, .mOffset = hdr.fStartCode}); + + // Blob was written, close fp. + + reader_protocol._Fp.close(); + + continue; + } + + kConsoleOut << "not an object container: " << objectFile << std::endl; + + // don't continue, it is a fatal error. + return LIBCOMPILER_EXEC_ERROR; + } + + pef_container.Cpu = archs; + + output_fc << pef_container; + + if (kVerbose) { + kConsoleOut << "wrote container to: " << output_fc.tellp() << ".\n"; + } + + output_fc.seekp(std::streamsize(pef_container.HdrSz)); + + std::vector<CompilerKit::STLString> not_found; + std::vector<CompilerKit::STLString> symbols; + + // step 2: check for errors (multiple symbols, undefined ones) + + for (auto& command_hdr : command_headers) { + // check if this symbol needs to be resolved. + if (CompilerKit::STLString(command_hdr.Name).find(kLdDefineSymbol) != + CompilerKit::STLString::npos && + CompilerKit::STLString(command_hdr.Name).find(kLdDynamicSym) == + CompilerKit::STLString::npos) { + if (kVerbose) kConsoleOut << "Found undefined symbol: " << command_hdr.Name << "\n"; + + if (auto it = std::find(not_found.begin(), not_found.end(), + CompilerKit::STLString(command_hdr.Name)); + it == not_found.end()) { + not_found.emplace_back(command_hdr.Name); + } + } + + symbols.emplace_back(command_hdr.Name); + } + + // Now try to solve these symbols. + + for (size_t not_found_idx = 0; not_found_idx < command_headers.size(); ++not_found_idx) { + if (const auto it = std::find(not_found.begin(), not_found.end(), + CompilerKit::STLString(command_headers[not_found_idx].Name)); + it != not_found.end()) { + CompilerKit::STLString symbol_imp = *it; + + if (symbol_imp.find(kLdDefineSymbol) == CompilerKit::STLString::npos) continue; + + // erase the lookup prefix. + symbol_imp.erase(0, symbol_imp.find(kLdDefineSymbol) + strlen(kLdDefineSymbol)); + + // demangle everything. + while (symbol_imp.find('$') != CompilerKit::STLString::npos) + symbol_imp.erase(symbol_imp.find('$'), 1); + + // the reason we do is because, this may not match the symbol, and we need + // to look for other matching symbols. + for (auto& command_hdr : command_headers) { + if (CompilerKit::STLString(command_hdr.Name).find(symbol_imp) != + CompilerKit::STLString::npos && + CompilerKit::STLString(command_hdr.Name).find(kLdDefineSymbol) == + CompilerKit::STLString::npos) { + CompilerKit::STLString undefined_symbol = command_hdr.Name; + auto result_of_sym = undefined_symbol.substr(undefined_symbol.find(symbol_imp)); + + for (int i = 0; result_of_sym[i] != 0; ++i) { + if (result_of_sym[i] != symbol_imp[i]) goto ld_continue_search; + } + + not_found.erase(it); + + if (kVerbose) kConsoleOut << "Found symbol: " << command_hdr.Name << "\n"; + + break; + } + } + + ld_continue_search: + continue; + } + } + + // step 3: check for errors (recheck if we have those symbols.) + + if (!kStartFound && is_executable) { + if (kVerbose) + kConsoleOut << "Undefined entrypoint: " << kPefStart + << ", you may have forget to link " + "against the C++ runtime library.\n"; + + kConsoleOut << "Undefined entrypoint " << kPefStart << " for executable: " << kOutput << "\n"; + } + + // step 4: write all PEF commands. + + CompilerKit::PEFCommandHeader date_cmd_hdr{}; + + time_t timestamp = time(nullptr); + + CompilerKit::STLString timeStampStr = "Container:BuildEpoch:"; + timeStampStr += std::to_string(timestamp); + + strncpy(date_cmd_hdr.Name, timeStampStr.c_str(), timeStampStr.size()); + + date_cmd_hdr.Flags = 0; + date_cmd_hdr.Kind = CompilerKit::kPefZero; + date_cmd_hdr.Offset = output_fc.tellp(); + date_cmd_hdr.Size = timeStampStr.size(); + + command_headers.push_back(date_cmd_hdr); + + CompilerKit::PEFCommandHeader abi_cmd_hdr{}; + + CompilerKit::STLString abi = kLinkerAbiContainer; + + switch (kArch) { + case CompilerKit::kPefArchAMD64: { + abi += "MSFT"; + break; + } + case CompilerKit::kPefArchPowerPC: { + abi += "SYSV"; + break; + } + case CompilerKit::kPefArch32000: + case CompilerKit::kPefArch64000: { + abi += "_NEP"; + break; + } + default: { + abi += "_IDK"; + break; + } + } + + MemoryCopy(abi_cmd_hdr.Name, abi.c_str(), abi.size()); + + abi_cmd_hdr.Size = abi.size(); + abi_cmd_hdr.Offset = output_fc.tellp(); + abi_cmd_hdr.Flags = 0; + abi_cmd_hdr.Kind = CompilerKit::kPefLinkerID; + + command_headers.push_back(abi_cmd_hdr); + + CompilerKit::PEFCommandHeader stack_cmd_hdr{0}; + + stack_cmd_hdr.Cpu = kArch; + stack_cmd_hdr.Flags = 0; + stack_cmd_hdr.Size = sizeof(uintptr_t); + stack_cmd_hdr.Offset = 0; + + MemoryCopy(stack_cmd_hdr.Name, kLinkerStackSizeSymbol, strlen(kLinkerStackSizeSymbol)); + + command_headers.push_back(stack_cmd_hdr); + + CompilerKit::PEFCommandHeader uuid_cmd_hdr{}; + + std::random_device rd; + + auto seedData = std::array<int, std::mt19937::state_size>{}; + std::generate(std::begin(seedData), std::end(seedData), std::ref(rd)); + std::seed_seq seq(std::begin(seedData), std::end(seedData)); + std::mt19937 generator(seq); + + auto gen = uuids::uuid_random_generator{generator}; + uuids::uuid id = gen(); + auto uuidStr = uuids::to_string(id); + + MemoryCopy(uuid_cmd_hdr.Name, "Container:GUID:4:", strlen("Container:GUID:4:")); + MemoryCopy(uuid_cmd_hdr.Name + strlen("Container:GUID:4:"), uuidStr.c_str(), uuidStr.size()); + + uuid_cmd_hdr.Size = strlen(uuid_cmd_hdr.Name); + uuid_cmd_hdr.Offset = output_fc.tellp(); + uuid_cmd_hdr.Flags = CompilerKit::kPefLinkerID; + uuid_cmd_hdr.Kind = CompilerKit::kPefZero; + + command_headers.push_back(uuid_cmd_hdr); + + // prepare a symbol vector. + std::vector<CompilerKit::STLString> undef_symbols; + std::vector<CompilerKit::STLString> dupl_symbols; + std::vector<CompilerKit::STLString> resolve_symbols; + + constexpr Int32 kPaddingOffset = 16; + + size_t previous_offset = + (command_headers.size() * sizeof(CompilerKit::PEFCommandHeader)) + kPaddingOffset; + + CompilerKit::PEFCommandHeader end_exec_hdr; + + end_exec_hdr.Offset = output_fc.tellp(); + end_exec_hdr.Flags = CompilerKit::kPefLinkerID; + end_exec_hdr.Kind = CompilerKit::kPefZero; + + MemoryCopy(end_exec_hdr.Name, "Container:Exec:END", strlen("Container:Exec:END")); + + end_exec_hdr.Size = strlen(end_exec_hdr.Name); + + command_headers.push_back(end_exec_hdr); + + // Finally write down the command headers. + // And check for any duplications + for (size_t commandHeaderIndex = 0UL; commandHeaderIndex < command_headers.size(); + ++commandHeaderIndex) { + if (CompilerKit::STLString(command_headers[commandHeaderIndex].Name).find(kLdDefineSymbol) != + CompilerKit::STLString::npos && + CompilerKit::STLString(command_headers[commandHeaderIndex].Name).find(kLdDynamicSym) == + CompilerKit::STLString::npos) { + // ignore :UndefinedSymbol: headers, they do not contain code. + continue; + } + + CompilerKit::STLString symbol_name = command_headers[commandHeaderIndex].Name; + + if (!symbol_name.empty()) { + undef_symbols.emplace_back(symbol_name); + } + + command_headers[commandHeaderIndex].Offset += previous_offset; + previous_offset += command_headers[commandHeaderIndex].Size; + + CompilerKit::STLString name = command_headers[commandHeaderIndex].Name; + + /// so this is valid when we get to the entrypoint. + /// it is always a code64 container. And should equal to kPefStart as well. + /// this chunk of code updates the pef_container.Start with the updated offset. + if (name.find(kPefStart) != CompilerKit::STLString::npos && + name.find(kPefCode64) != CompilerKit::STLString::npos) { + pef_container.Start = command_headers[commandHeaderIndex].Offset; + auto tellCurPos = output_fc.tellp(); + + output_fc.seekp(0); + output_fc << pef_container; + + output_fc.seekp(tellCurPos); + } + + if (kVerbose) { + kConsoleOut << "Command name: " << name << "\n"; + kConsoleOut << "VMAddress of command content: " << command_headers[commandHeaderIndex].Offset + << "\n"; + } + + output_fc << command_headers[commandHeaderIndex]; + + for (size_t sub_command_header_index = 0UL; sub_command_header_index < command_headers.size(); + ++sub_command_header_index) { + if (sub_command_header_index == commandHeaderIndex) continue; + + if (CompilerKit::STLString(command_headers[sub_command_header_index].Name) + .find(kLdDefineSymbol) != CompilerKit::STLString::npos && + CompilerKit::STLString(command_headers[sub_command_header_index].Name) + .find(kLdDynamicSym) == CompilerKit::STLString::npos) { + if (kVerbose) { + kConsoleOut << "Ignoring :UndefinedSymbol: headers...\n"; + } + + // ignore :UndefinedSymbol: headers, they do not contain code. + continue; + } + + auto& command_hdr = command_headers[sub_command_header_index]; + + if (command_hdr.Name == CompilerKit::STLString(command_headers[commandHeaderIndex].Name)) { + if (std::find(dupl_symbols.cbegin(), dupl_symbols.cend(), command_hdr.Name) == + dupl_symbols.cend()) { + dupl_symbols.emplace_back(command_hdr.Name); + } + + if (kVerbose) kConsoleOut << "Found duplicate symbols of: " << command_hdr.Name << "\n"; + + kDuplicateSymbols = true; + } + } + } + + if (!dupl_symbols.empty()) { + for (auto& symbol : dupl_symbols) { + kConsoleOut << "Multiple symbols of: " << symbol << " detected, cannot continue.\n"; + } + + return LIBCOMPILER_EXEC_ERROR; + } + + // step 2.5: write program bytes. + + for (auto& struct_of_blob : kObjectBytes) { + output_fc.write(struct_of_blob.mBlob.data(), struct_of_blob.mBlob.size()); + } + + if (kVerbose) { + kConsoleOut << "Wrote contents of: " << kOutput << "\n"; + } + + // step 3: check if we have those symbols + + std::vector<CompilerKit::STLString> unreferenced_symbols; + + for (auto& command_hdr : command_headers) { + if (auto it = + std::find(not_found.begin(), not_found.end(), CompilerKit::STLString(command_hdr.Name)); + it != not_found.end()) { + unreferenced_symbols.emplace_back(command_hdr.Name); + } + } + + if (!unreferenced_symbols.empty()) { + for (auto& unreferenced_symbol : unreferenced_symbols) { + kConsoleOut << "Undefined symbol " << unreferenced_symbol << "\n"; + } + + return LIBCOMPILER_EXEC_ERROR; + } + + if ((!kStartFound || kDuplicateSymbols) && + (std::filesystem::exists(kOutput) || !unreferenced_symbols.empty())) { + if (kVerbose) { + kConsoleOut << "File: " << kOutput << ", is corrupt, removing file...\n"; + } + + return LIBCOMPILER_EXEC_ERROR; + } + + return LIBCOMPILER_SUCCESS; +} + +// Last rev 13-1-24 diff --git a/dev/CompilerKit/src/Macro/CPlusPlusCompilerPreProcessor.cc b/dev/CompilerKit/src/Macro/CPlusPlusCompilerPreProcessor.cc new file mode 100644 index 0000000..12a69c8 --- /dev/null +++ b/dev/CompilerKit/src/Macro/CPlusPlusCompilerPreProcessor.cc @@ -0,0 +1,893 @@ +/* + * ======================================================== + * + * C++ Preprocessor Driver + * Copyright (C) 2024-2025 Amlal El Mahrouss, all rights reserved. + * + * ======================================================== + */ + +/// BUGS: 0 + +#include <CompilerKit/ErrorID.h> +#include <CompilerKit/Frontend.h> +#include <algorithm> +#include <filesystem> +#include <fstream> +#include <iostream> +#include <stdexcept> +#include <vector> + +#define kMacroPrefix '#' + +/// @author EL Mahrouss Amlal (amlel) +/// @file bpp.cxx +/// @brief Preprocessor. + +typedef Int32 (*bpp_parser_fn_t)(CompilerKit::STLString& line, std::ifstream& hdr_file, + std::ofstream& pp_out); + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief Preprocessor internal types. + +///////////////////////////////////////////////////////////////////////////////////////// + +namespace Detail { +enum { + kInvalid = 0, + kEqual = 100, + kGreaterEqThan, + kLesserEqThan, + kGreaterThan, + kLesserThan, + kNotEqual, + kCount = 6, +}; + +struct bpp_macro_condition final { + int32_t fType; + CompilerKit::STLString fTypeName; + + void Print() { + std::cout << "type: " << fType << "\n"; + std::cout << "type_name: " << fTypeName << "\n"; + } +}; + +struct bpp_macro final { + std::vector<CompilerKit::STLString> fArgs; + CompilerKit::STLString fName; + CompilerKit::STLString fValue; + + void Print() { + std::cout << "name: " << fName << "\n"; + std::cout << "value: " << fValue << "\n"; + + for (auto& arg : fArgs) { + std::cout << "arg: " << arg << "\n"; + } + } +}; +} // namespace Detail + +static std::vector<CompilerKit::STLString> kFiles; +static std::vector<Detail::bpp_macro> kMacros; +static std::vector<CompilerKit::STLString> kIncludes; + +static CompilerKit::STLString kWorkingDir = ""; + +///////////////////////////////////////////////////////////////////////////////////////// + +// @name bpp_parse_if_condition +// @brief parse #if condition + +///////////////////////////////////////////////////////////////////////////////////////// + +int32_t bpp_parse_if_condition(Detail::bpp_macro_condition& cond, Detail::bpp_macro& macro, + bool& inactive_code, bool& defined, + CompilerKit::STLString& macro_str) { + if (cond.fType == Detail::kEqual) { + auto substr_macro = macro_str.substr(macro_str.find(macro.fName) + macro.fName.size()); + + if (substr_macro.find(macro.fValue) != CompilerKit::STLString::npos) { + if (macro.fValue == "0") { + defined = false; + inactive_code = true; + + return 1; + } + + defined = true; + inactive_code = false; + + return 1; + } + } else if (cond.fType == Detail::kNotEqual) { + auto substr_macro = macro_str.substr(macro_str.find(macro.fName) + macro.fName.size()); + + if (substr_macro.find(macro.fName) != CompilerKit::STLString::npos) { + if (substr_macro.find(macro.fValue) != CompilerKit::STLString::npos) { + defined = false; + inactive_code = true; + + return 1; + } + + defined = true; + inactive_code = false; + + return 1; + } + + return 0; + } + + auto substr_macro = macro_str.substr(macro_str.find(macro.fName) + macro.fName.size()); + + CompilerKit::STLString number; + + for (auto& macro_num : kMacros) { + if (substr_macro.find(macro_num.fName) != CompilerKit::STLString::npos) { + for (size_t i = 0; i < macro_num.fName.size(); ++i) { + if (isdigit(macro_num.fValue[i])) { + number += macro_num.fValue[i]; + } else { + number.clear(); + break; + } + } + + break; + } + } + + size_t y = 2; + + /* last try */ + for (; y < macro_str.size(); y++) { + if (isdigit(macro_str[y])) { + for (size_t x = y; x < macro_str.size(); x++) { + if (macro_str[x] == ' ') break; + + number += macro_str[x]; + } + + break; + } + } + + size_t rhs = atol(macro.fValue.c_str()); + size_t lhs = atol(number.c_str()); + + if (lhs == 0) { + number.clear(); + ++y; + + for (; y < macro_str.size(); y++) { + if (isdigit(macro_str[y])) { + for (size_t x = y; x < macro_str.size(); x++) { + if (macro_str[x] == ' ') break; + + number += macro_str[x]; + } + + break; + } + } + + lhs = atol(number.c_str()); + } + + if (cond.fType == Detail::kGreaterThan) { + if (lhs < rhs) { + defined = true; + inactive_code = false; + + return 1; + } + + return 0; + } + + if (cond.fType == Detail::kGreaterEqThan) { + if (lhs <= rhs) { + defined = true; + inactive_code = false; + + return 1; + } + + return 0; + } + + if (cond.fType == Detail::kLesserEqThan) { + if (lhs >= rhs) { + defined = true; + inactive_code = false; + + return 1; + } + + return 0; + } + + if (cond.fType == Detail::kLesserThan) { + if (lhs > rhs) { + defined = true; + inactive_code = false; + + return 1; + } + + return 0; + } + + return 0; +} + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief stores every included file here. + +///////////////////////////////////////////////////////////////////////////////////////// + +std::vector<CompilerKit::STLString> kAllIncludes; + +///////////////////////////////////////////////////////////////////////////////////////// + +// @name bpp_parse_file +// @brief parse file to preprocess it. + +///////////////////////////////////////////////////////////////////////////////////////// + +void bpp_parse_file(std::ifstream& hdr_file, std::ofstream& pp_out) { + CompilerKit::STLString hdr_line; + CompilerKit::STLString line_after_include; + + bool inactive_code = false; + bool defined = false; + + try { + while (std::getline(hdr_file, hdr_line)) { + if (inactive_code) { + if (hdr_line.find("#endif") == CompilerKit::STLString::npos) { + continue; + } else if (hdr_line[0] == kMacroPrefix && + hdr_line.find("#endif") != CompilerKit::STLString::npos) { + inactive_code = false; + } + } + + if (hdr_line.find("*/") != CompilerKit::STLString::npos) { + hdr_line.erase(hdr_line.find("*/"), strlen("*/")); + } + + if (hdr_line.find("/*") != CompilerKit::STLString::npos) { + inactive_code = true; + + // get rid of comment. + hdr_line.erase(hdr_line.find("/*")); + } + + if (hdr_line[0] == kMacroPrefix && hdr_line.find("endif") != CompilerKit::STLString::npos) { + if (!defined && inactive_code) { + inactive_code = false; + defined = false; + + continue; + } + + continue; + } + + if (!defined && inactive_code) { + continue; + } + + if (defined && inactive_code) { + continue; + } + + for (auto macro : kMacros) { + if (CompilerKit::find_word(hdr_line, macro.fName)) { + if (hdr_line.substr(hdr_line.find(macro.fName)).find(macro.fName + '(') != + CompilerKit::STLString::npos) { + if (!macro.fArgs.empty()) { + CompilerKit::STLString symbol_val = macro.fValue; + std::vector<CompilerKit::STLString> args; + + size_t x_arg_indx = 0; + + CompilerKit::STLString line_after_define = hdr_line; + CompilerKit::STLString str_arg; + + if (line_after_define.find("(") != CompilerKit::STLString::npos) { + line_after_define.erase(0, line_after_define.find("(") + 1); + + for (auto& subc : line_after_define) { + if (subc == ' ' || subc == '\t') continue; + + if (subc == ',' || subc == ')') { + if (str_arg.empty()) continue; + + args.push_back(str_arg); + + str_arg.clear(); + + continue; + } + + str_arg.push_back(subc); + } + } + + for (auto arg : macro.fArgs) { + if (symbol_val.find(macro.fArgs[x_arg_indx]) != CompilerKit::STLString::npos) { + symbol_val.replace(symbol_val.find(macro.fArgs[x_arg_indx]), + macro.fArgs[x_arg_indx].size(), args[x_arg_indx]); + ++x_arg_indx; + } else { + throw std::runtime_error("cppdrv: Internal error."); + } + } + + auto len = macro.fName.size(); + len += symbol_val.size(); + len += 2; // ( and ) + + hdr_line.erase(hdr_line.find(")"), 1); + + hdr_line.replace(hdr_line.find(hdr_line.substr(hdr_line.find(macro.fName + '('))), + len, symbol_val); + } else { + auto value = macro.fValue; + + hdr_line.replace(hdr_line.find(macro.fName), macro.fName.size(), value); + } + } + } + } + + if (hdr_line[0] == kMacroPrefix && hdr_line.find("define ") != CompilerKit::STLString::npos) { + auto line_after_define = hdr_line.substr(hdr_line.find("define ") + strlen("define ")); + + CompilerKit::STLString macro_value; + CompilerKit::STLString macro_key; + + std::size_t pos = 0UL; + + std::vector<CompilerKit::STLString> args; + bool on_args = false; + + for (auto& ch : line_after_define) { + ++pos; + + if (ch == '(') { + on_args = true; + continue; + } + + if (ch == ')') { + on_args = false; + continue; + } + + if (ch == '\\') continue; + + if (on_args) continue; + + if (ch == ' ') { + for (size_t i = pos; i < line_after_define.size(); i++) { + macro_value += line_after_define[i]; + } + + break; + } + + macro_key += ch; + } + + CompilerKit::STLString str; + + if (line_after_define.find("(") != CompilerKit::STLString::npos) { + line_after_define.erase(0, line_after_define.find("(") + 1); + + for (auto& subc : line_after_define) { + if (subc == ',' || subc == ')') { + if (str.empty()) continue; + + args.push_back(str); + + str.clear(); + + continue; + } + + str.push_back(subc); + } + } + + Detail::bpp_macro macro; + + macro.fArgs = args; + macro.fName = macro_key; + macro.fValue = macro_value; + + kMacros.emplace_back(macro); + + continue; + } + + if (hdr_line[0] != kMacroPrefix) { + if (inactive_code) { + continue; + } + + pp_out << hdr_line << std::endl; + + continue; + } + + if (hdr_line[0] == kMacroPrefix && hdr_line.find("ifndef") != CompilerKit::STLString::npos) { + auto line_after_ifndef = hdr_line.substr(hdr_line.find("ifndef") + strlen("ifndef") + 1); + CompilerKit::STLString macro; + + for (auto& ch : line_after_ifndef) { + if (ch == ' ') { + break; + } + + macro += ch; + } + + if (macro == "0") { + defined = true; + inactive_code = false; + continue; + } + + if (macro == "1") { + defined = false; + inactive_code = true; + + continue; + } + + bool found = false; + + defined = true; + inactive_code = false; + + for (auto& macro_ref : kMacros) { + if (hdr_line.find(macro_ref.fName) != CompilerKit::STLString::npos) { + found = true; + break; + } + } + + if (found) { + defined = false; + inactive_code = true; + + continue; + } + } else if (hdr_line[0] == kMacroPrefix && + hdr_line.find("else") != CompilerKit::STLString::npos) { + if (!defined && inactive_code) { + inactive_code = false; + defined = true; + + continue; + } else { + defined = false; + inactive_code = true; + + continue; + } + } else if (hdr_line[0] == kMacroPrefix && + hdr_line.find("ifdef") != CompilerKit::STLString::npos) { + auto line_after_ifdef = hdr_line.substr(hdr_line.find("ifdef") + strlen("ifdef") + 1); + CompilerKit::STLString macro; + + for (auto& ch : line_after_ifdef) { + if (ch == ' ') { + break; + } + + macro += ch; + } + + if (macro == "0") { + defined = false; + inactive_code = true; + + continue; + } + + if (macro == "1") { + defined = true; + inactive_code = false; + + continue; + } + + defined = false; + inactive_code = true; + + for (auto& macro_ref : kMacros) { + if (hdr_line.find(macro_ref.fName) != CompilerKit::STLString::npos) { + defined = true; + inactive_code = false; + + break; + } + } + } else if (hdr_line[0] == kMacroPrefix && + hdr_line.find("if") != CompilerKit::STLString::npos) { + inactive_code = true; + + std::vector<Detail::bpp_macro_condition> bpp_macro_condition_list = { + { + .fType = Detail::kEqual, + .fTypeName = "==", + }, + { + .fType = Detail::kNotEqual, + .fTypeName = "!=", + }, + { + .fType = Detail::kLesserThan, + .fTypeName = "<", + }, + { + .fType = Detail::kGreaterThan, + .fTypeName = ">", + }, + { + .fType = Detail::kLesserEqThan, + .fTypeName = "<=", + }, + { + .fType = Detail::kGreaterEqThan, + .fTypeName = ">=", + }, + }; + + int32_t good_to_go = 0; + + for (auto& macro_condition : bpp_macro_condition_list) { + if (hdr_line.find(macro_condition.fTypeName) != CompilerKit::STLString::npos) { + for (auto& found_macro : kMacros) { + if (hdr_line.find(found_macro.fName) != CompilerKit::STLString::npos) { + good_to_go = bpp_parse_if_condition(macro_condition, found_macro, inactive_code, + defined, hdr_line); + + break; + } + } + } + } + + if (good_to_go) continue; + + auto line_after_if = hdr_line.substr(hdr_line.find("if") + strlen("if") + 1); + CompilerKit::STLString macro; + + for (auto& ch : line_after_if) { + if (ch == ' ') { + break; + } + + macro += ch; + } + + if (macro == "0") { + defined = false; + inactive_code = true; + continue; + } + + if (macro == "1") { + defined = true; + inactive_code = false; + + continue; + } + + // last try, is it defined to be one? + for (auto& macro_ref : kMacros) { + if (macro_ref.fName.find(macro) != CompilerKit::STLString::npos && + macro_ref.fValue == "1") { + inactive_code = false; + defined = true; + + break; + } + } + } else if (hdr_line[0] == kMacroPrefix && + hdr_line.find("warning") != CompilerKit::STLString::npos) { + auto line_after_warning = hdr_line.substr(hdr_line.find("warning") + strlen("warning") + 1); + CompilerKit::STLString message; + + for (auto& ch : line_after_warning) { + if (ch == '\r' || ch == '\n') { + break; + } + + message += ch; + } + + std::cout << "warn: " << message << std::endl; + } else if (hdr_line[0] == kMacroPrefix && + hdr_line.find("error") != CompilerKit::STLString::npos) { + auto line_after_warning = hdr_line.substr(hdr_line.find("error") + strlen("error") + 1); + CompilerKit::STLString message; + + for (auto& ch : line_after_warning) { + if (ch == '\r' || ch == '\n') { + break; + } + + message += ch; + } + + throw std::runtime_error("error: " + message); + } else if (hdr_line[0] == kMacroPrefix && + hdr_line.find("include ") != CompilerKit::STLString::npos) { + line_after_include = hdr_line.substr(hdr_line.find("include ") + strlen("include ")); + + kIncludeFile: + auto it = std::find(kAllIncludes.cbegin(), kAllIncludes.cend(), line_after_include); + + if (it != kAllIncludes.cend()) { + continue; + } + + CompilerKit::STLString path; + + kAllIncludes.push_back(line_after_include); + + bool enable = false; + bool not_local = false; + + for (auto& ch : line_after_include) { + if (ch == ' ') continue; + + if (ch == '<') { + not_local = true; + enable = true; + + continue; + } + + if (ch == '\"') { + not_local = false; + enable = true; + continue; + } + + if (enable) { + path += ch; + } + } + + if (not_local) { + bool open = false; + + if (path.ends_with('>')) { + path.erase(path.find('>')); + } + + if (path.ends_with('"')) { + path.erase(path.find('"')); + } + + for (auto& include : kIncludes) { + CompilerKit::STLString header_path = include; + header_path.push_back('/'); + header_path += path; + + std::ifstream header(header_path); + + if (!header.is_open()) continue; + + open = true; + + bpp_parse_file(header, pp_out); + + break; + } + + if (!open) { + throw std::runtime_error("cppdrv: no such include file: " + path); + } + } else { + std::ifstream header(path); + + if (!header.is_open()) throw std::runtime_error("cppdrv: no such include file: " + path); + + bpp_parse_file(header, pp_out); + } + } else { + std::cerr << ("cppdrv: unknown pre-processor directive, " + hdr_line) << "\n"; + continue; + } + } + } catch (std::out_of_range& oor) { + return; + } +} + +///////////////////////////////////////////////////////////////////////////////////////// + +// @brief main entrypoint of app. + +///////////////////////////////////////////////////////////////////////////////////////// + +LIBCOMPILER_MODULE(CPlusPlusPreprocessorMain) { + try { + bool skip = false; + bool double_skip = false; + + Detail::bpp_macro macro_1; + + macro_1.fName = "__true"; + macro_1.fValue = "1"; + + kMacros.push_back(macro_1); + + Detail::bpp_macro macro_unreachable; + + macro_unreachable.fName = "__unreachable"; + macro_unreachable.fValue = "__libcompiler_unreachable"; + + kMacros.push_back(macro_unreachable); + + Detail::bpp_macro macro_unused; + + macro_unreachable.fName = "__unused"; + macro_unreachable.fValue = "__libcompiler_unused"; + + kMacros.push_back(macro_unused); + + Detail::bpp_macro macro_0; + + macro_0.fName = "__false"; + macro_0.fValue = "0"; + + kMacros.push_back(macro_0); + + Detail::bpp_macro macro_zka; + + macro_zka.fName = "__LIBCOMPILER__"; + macro_zka.fValue = "1"; + + kMacros.push_back(macro_zka); + + Detail::bpp_macro macro_cxx; + + macro_cxx.fName = "__cplusplus"; + macro_cxx.fValue = "202302L"; + + kMacros.push_back(macro_cxx); + + Detail::bpp_macro macro_size_t; + macro_size_t.fName = "__SIZE_TYPE__"; + macro_size_t.fValue = "unsigned long long int"; + + kMacros.push_back(macro_size_t); + + macro_size_t.fName = "__UINT32_TYPE__"; + macro_size_t.fValue = "unsigned int"; + + kMacros.push_back(macro_size_t); + + macro_size_t.fName = "__UINTPTR_TYPE__"; + macro_size_t.fValue = "unsigned long long int"; + + kMacros.push_back(macro_size_t); + + for (auto index = 1UL; index < argc; ++index) { + if (skip) { + skip = false; + continue; + } + + if (double_skip) { + ++index; + double_skip = false; + continue; + } + + if (argv[index][0] == '-') { + if (strcmp(argv[index], "-cpp-ver") == 0) { + printf("%s\n", + "NeKernel Preprocessor Driver v1.11, (c) Amlal El Mahrouss 2024-2025 all rights " + "reserved."); + + return LIBCOMPILER_SUCCESS; + } + + if (strcmp(argv[index], "-cpp-help") == 0) { + printf("%s\n", + "NeKernel Preprocessor Driver v1.11, (c) Amlal El Mahrouss 2024-2025 all rights " + "reserved."); + printf("%s\n", "-cpp-working-dir <path>: set directory to working path."); + printf("%s\n", "-cpp-include-dir <path>: add directory to include path."); + printf("%s\n", "-cpp-def <name> <value>: define a macro."); + printf("%s\n", "-cpp-ver: print the version."); + printf("%s\n", "-cpp-help: show help (this current command)."); + + return LIBCOMPILER_SUCCESS; + } + + if (strcmp(argv[index], "-cpp-include-dir") == 0) { + CompilerKit::STLString inc = argv[index + 1]; + + skip = true; + + kIncludes.push_back(inc); + } + + if (strcmp(argv[index], "-cpp-working-dir") == 0) { + CompilerKit::STLString inc = argv[index + 1]; + skip = true; + kWorkingDir = inc; + } + + if (strcmp(argv[index], "-cpp-def") == 0 && argv[index + 1] != nullptr && + argv[index + 2] != nullptr) { + CompilerKit::STLString macro_key = argv[index + 1]; + + CompilerKit::STLString macro_value; + bool is_string = false; + + for (int argv_find_len = 0; argv_find_len < strlen(argv[index]); ++argv_find_len) { + if (!isdigit(argv[index][argv_find_len])) { + is_string = true; + macro_value += "\""; + + break; + } + } + + macro_value += argv[index + 2]; + + if (is_string) macro_value += "\""; + + Detail::bpp_macro macro; + macro.fName = macro_key; + macro.fValue = macro_value; + + kMacros.push_back(macro); + + double_skip = true; + } + + continue; + } + + kFiles.emplace_back(argv[index]); + } + + if (kFiles.empty()) return LIBCOMPILER_EXEC_ERROR; + + for (auto& file : kFiles) { + if (!std::filesystem::exists(file)) continue; + + std::ifstream file_descriptor(file); + std::ofstream file_descriptor_pp(file + ".pp"); + + bpp_parse_file(file_descriptor, file_descriptor_pp); + } + + return LIBCOMPILER_SUCCESS; + } catch (const std::runtime_error& e) { + std::cout << e.what() << '\n'; + } + + return LIBCOMPILER_EXEC_ERROR; +} + +// Last rev 8-1-24 |
