220 lines
8.8 KiB
C++
220 lines
8.8 KiB
C++
#include "lang/preprocessor.hpp"
|
|
|
|
namespace hex::lang {
|
|
|
|
Preprocessor::Preprocessor() {
|
|
|
|
}
|
|
|
|
std::optional<std::string> Preprocessor::preprocess(const std::string& code, bool initialRun) {
|
|
u32 offset = 0;
|
|
u32 lineNumber = 1;
|
|
|
|
if (initialRun) {
|
|
this->m_defines.clear();
|
|
this->m_pragmas.clear();
|
|
}
|
|
|
|
std::string output;
|
|
output.reserve(code.length());
|
|
|
|
try {
|
|
while (offset < code.length()) {
|
|
if (code[offset] == '#') {
|
|
offset += 1;
|
|
|
|
if (code.substr(offset, 7) == "include") {
|
|
offset += 7;
|
|
|
|
while (std::isblank(code[offset]) || std::isspace(code[offset]))
|
|
offset += 1;
|
|
|
|
if (code[offset] != '<' && code[offset] != '"')
|
|
throwPreprocessorError("expected '<' or '\"' before file name", lineNumber);
|
|
|
|
char endChar = code[offset];
|
|
if (endChar == '<') endChar = '>';
|
|
|
|
offset += 1;
|
|
|
|
std::string includeFile;
|
|
while (code[offset] != endChar) {
|
|
includeFile += code[offset];
|
|
offset += 1;
|
|
|
|
if (offset >= code.length())
|
|
throwPreprocessorError(hex::format("missing terminating '%c' character", endChar), lineNumber);
|
|
}
|
|
offset += 1;
|
|
|
|
if (includeFile[0] != '/')
|
|
includeFile = "include/" + includeFile;
|
|
|
|
FILE *file = fopen(includeFile.c_str(), "r");
|
|
if (file == nullptr)
|
|
throwPreprocessorError(hex::format("%s: No such file or directory", includeFile.c_str()), lineNumber);
|
|
|
|
fseek(file, 0, SEEK_END);
|
|
size_t size = ftell(file);
|
|
char *buffer = new char[size + 1];
|
|
rewind(file);
|
|
|
|
fread(buffer, size, 1, file);
|
|
buffer[size] = 0x00;
|
|
|
|
auto preprocessedInclude = this->preprocess(buffer, false);
|
|
if (!preprocessedInclude.has_value())
|
|
throw this->m_error;
|
|
|
|
auto content = preprocessedInclude.value();
|
|
|
|
std::replace(content.begin(), content.end(), '\n', ' ');
|
|
std::replace(content.begin(), content.end(), '\r', ' ');
|
|
|
|
output += content;
|
|
|
|
delete[] buffer;
|
|
|
|
fclose(file);
|
|
} else if (code.substr(offset, 6) == "define") {
|
|
offset += 6;
|
|
|
|
while (std::isblank(code[offset])) {
|
|
offset += 1;
|
|
}
|
|
|
|
std::string defineName;
|
|
while (!std::isblank(code[offset])) {
|
|
defineName += code[offset];
|
|
|
|
if (offset >= code.length() || code[offset] == '\n' || code[offset] == '\r')
|
|
throwPreprocessorError("no value given in #define directive", lineNumber);
|
|
offset += 1;
|
|
}
|
|
|
|
while (std::isblank(code[offset])) {
|
|
offset += 1;
|
|
if (offset >= code.length())
|
|
throwPreprocessorError("no value given in #define directive", lineNumber);
|
|
}
|
|
|
|
std::string replaceValue;
|
|
while (code[offset] != '\n' && code[offset] != '\r') {
|
|
if (offset >= code.length())
|
|
throwPreprocessorError("missing new line after #define directive", lineNumber);
|
|
|
|
replaceValue += code[offset];
|
|
offset += 1;
|
|
}
|
|
|
|
if (replaceValue.empty())
|
|
throwPreprocessorError("no value given in #define directive", lineNumber);
|
|
|
|
this->m_defines.emplace(defineName, replaceValue);
|
|
} else if (code.substr(offset, 6) == "pragma") {
|
|
offset += 6;
|
|
|
|
while (std::isblank(code[offset]))
|
|
offset += 1;
|
|
|
|
std::string pragmaKey;
|
|
while (!std::isblank(code[offset])) {
|
|
pragmaKey += code[offset];
|
|
|
|
if (offset >= code.length() || code[offset] == '\n' || code[offset] == '\r')
|
|
throwPreprocessorError("no instruction given in #pragma directive", lineNumber);
|
|
|
|
offset += 1;
|
|
}
|
|
|
|
while (std::isblank(code[offset]))
|
|
offset += 1;
|
|
|
|
std::string pragmaValue;
|
|
while (code[offset] != '\n' && code[offset] != '\r') {
|
|
if (offset >= code.length())
|
|
throwPreprocessorError("missing new line after #pragma directive", lineNumber);
|
|
|
|
pragmaValue += code[offset];
|
|
offset += 1;
|
|
}
|
|
|
|
if (pragmaValue.empty())
|
|
throwPreprocessorError("missing value in #pragma directive", lineNumber);
|
|
|
|
this->m_pragmas.emplace(pragmaKey, pragmaValue);
|
|
} else
|
|
throwPreprocessorError("unknown preprocessor directive", lineNumber);
|
|
} else if (code.substr(offset, 2) == "//") {
|
|
while (code[offset] != '\n' && offset < code.length())
|
|
offset += 1;
|
|
} else if (code.substr(offset, 2) == "/*") {
|
|
while (code.substr(offset, 2) != "*/" && offset < code.length()) {
|
|
if (code[offset] == '\n') {
|
|
output += '\n';
|
|
lineNumber++;
|
|
}
|
|
|
|
offset += 1;
|
|
}
|
|
|
|
offset += 2;
|
|
if (offset >= code.length())
|
|
throwPreprocessorError("unterminated comment", lineNumber - 1);
|
|
}
|
|
|
|
if (code[offset] == '\n')
|
|
lineNumber++;
|
|
|
|
output += code[offset];
|
|
offset += 1;
|
|
}
|
|
|
|
if (initialRun) {
|
|
// Apply defines
|
|
std::vector<std::pair<std::string, std::string>> sortedDefines;
|
|
std::copy(this->m_defines.begin(), this->m_defines.end(), std::back_inserter(sortedDefines));
|
|
std::sort(sortedDefines.begin(), sortedDefines.end(), [](const auto &left, const auto &right) {
|
|
return left.first.size() > right.first.size();
|
|
});
|
|
|
|
for (const auto &[define, value] : sortedDefines) {
|
|
s32 index = 0;
|
|
while((index = output.find(define, index)) != std::string::npos) {
|
|
output.replace(index, define.length(), value);
|
|
index += value.length();
|
|
}
|
|
}
|
|
|
|
// Handle pragmas
|
|
for (const auto &[type, value] : this->m_pragmas) {
|
|
if (this->m_pragmaHandlers.contains(type)) {
|
|
if (!this->m_pragmaHandlers[type](value))
|
|
throwPreprocessorError(hex::format("invalid value provided to '%s' #pragma directive", type.c_str()), lineNumber);
|
|
} else
|
|
throwPreprocessorError(hex::format("no #pragma handler registered for type %s", type.c_str()), lineNumber);
|
|
}
|
|
}
|
|
} catch (PreprocessorError &e) {
|
|
this->m_error = e;
|
|
return { };
|
|
}
|
|
|
|
return output;
|
|
}
|
|
|
|
void Preprocessor::addPragmaHandler(const std::string &pragmaType, const std::function<bool(const std::string&)> &function) {
|
|
if (!this->m_pragmaHandlers.contains(pragmaType))
|
|
this->m_pragmaHandlers.emplace(pragmaType, function);
|
|
}
|
|
|
|
void Preprocessor::addDefaultPragmaHandlers() {
|
|
this->addPragmaHandler("MIME", [](const std::string &value) {
|
|
return !std::all_of(value.begin(), value.end(), isspace) && !value.ends_with('\n') && !value.ends_with('\r');
|
|
});
|
|
this->addPragmaHandler("endian", [](const std::string &value) {
|
|
return value == "big" || value == "little" || value == "native";
|
|
});
|
|
}
|
|
|
|
} |