Files
deusdata--codebase-memory-mcp/internal/cbm/vendored/simplecpp/simplecpp.cpp
T
wehub-resource-sync 41cb1c0170
OpenSSF Scorecard / scorecard (push) Failing after 0s
DCO / dco (push) Failing after 0s
CodeQL SAST / analyze (push) Failing after 1s
Deploy Pages / deploy (push) Failing after 1s
chore: import upstream snapshot with attribution
2026-07-13 12:28:05 +08:00

3954 lines
148 KiB
C++

/*
* simplecpp - A simple and high-fidelity C/C++ preprocessor library
* Copyright (C) 2016-2023 simplecpp team
*/
#if defined(_WIN32)
# ifndef _WIN32_WINNT
# define _WIN32_WINNT 0x0602
# endif
# ifndef NOMINMAX
# define NOMINMAX
# endif
# ifndef WIN32_LEAN_AND_MEAN
# define WIN32_LEAN_AND_MEAN
# endif
# include <windows.h>
# undef ERROR
#endif
#include "simplecpp.h"
#if defined(_WIN32) || defined(__CYGWIN__) || defined(__MINGW32__)
# define SIMPLECPP_WINDOWS
#endif
#include <algorithm>
#include <cassert>
#include <cctype>
#include <climits>
#include <cstddef> // IWYU pragma: keep
#include <cstdint>
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <ctime>
#include <fstream>
#include <iostream>
#include <istream>
#include <limits>
#include <list>
#include <map>
#include <set>
#include <sstream>
#include <stack>
#include <stdexcept>
#include <string>
#ifdef SIMPLECPP_WINDOWS
# include <mutex>
#endif
#include <unordered_map>
#include <utility>
#include <vector>
#ifdef _WIN32
# include <direct.h>
#else
# include <sys/stat.h>
#endif
static bool isHex(const std::string &s)
{
return s.size()>2 && (s.compare(0,2,"0x")==0 || s.compare(0,2,"0X")==0);
}
static bool isOct(const std::string &s)
{
return s.size()>1 && (s[0]=='0') && (s[1] >= '0') && (s[1] < '8');
}
static bool isStringLiteral(const std::string &s)
{
return s.size() > 1 && (s[0]=='\"') && (*s.rbegin()=='\"');
}
static bool isCharLiteral(const std::string &s)
{
// char literal patterns can include 'a', '\t', '\000', '\xff', 'abcd', and maybe ''
// This only checks for the surrounding '' but doesn't parse the content.
return s.size() > 1 && (s[0]=='\'') && (*s.rbegin()=='\'');
}
static const simplecpp::TokenString DEFINE("define");
static const simplecpp::TokenString UNDEF("undef");
static const simplecpp::TokenString INCLUDE("include");
static const simplecpp::TokenString ERROR("error");
static const simplecpp::TokenString WARNING("warning");
static const simplecpp::TokenString IF("if");
static const simplecpp::TokenString IFDEF("ifdef");
static const simplecpp::TokenString IFNDEF("ifndef");
static const simplecpp::TokenString DEFINED("defined");
static const simplecpp::TokenString ELSE("else");
static const simplecpp::TokenString ELIF("elif");
static const simplecpp::TokenString ENDIF("endif");
static const simplecpp::TokenString PRAGMA("pragma");
static const simplecpp::TokenString ONCE("once");
static const simplecpp::TokenString HAS_INCLUDE("__has_include");
template<class T> static std::string toString(T t)
{
// NOLINTNEXTLINE(misc-const-correctness) - false positive
std::ostringstream ostr;
ostr << t;
return ostr.str();
}
#ifdef SIMPLECPP_DEBUG_MACRO_EXPANSION
static std::string locstring(const simplecpp::Location &loc)
{
std::ostringstream ostr;
ostr << '[' << loc.file() << ':' << loc.line << ':' << loc.col << ']';
return ostr.str();
}
#endif
static long long stringToLL(const std::string &s)
{
long long ret;
const bool hex = isHex(s);
const bool oct = isOct(s);
std::istringstream istr(hex ? s.substr(2) : oct ? s.substr(1) : s);
if (hex)
istr >> std::hex;
else if (oct)
istr >> std::oct;
istr >> ret;
return ret;
}
static unsigned long long stringToULL(const std::string &s)
{
unsigned long long ret;
const bool hex = isHex(s);
const bool oct = isOct(s);
std::istringstream istr(hex ? s.substr(2) : oct ? s.substr(1) : s);
if (hex)
istr >> std::hex;
else if (oct)
istr >> std::oct;
istr >> ret;
return ret;
}
static bool endsWith(const std::string &s, const std::string &e)
{
return (s.size() >= e.size()) && std::equal(e.rbegin(), e.rend(), s.rbegin());
}
static bool sameline(const simplecpp::Token *tok1, const simplecpp::Token *tok2)
{
return tok1 && tok2 && tok1->location.sameline(tok2->location);
}
static bool isAlternativeBinaryOp(const simplecpp::Token *tok, const std::string &alt)
{
return (tok->name &&
tok->str() == alt &&
tok->previous &&
tok->next &&
(tok->previous->number || tok->previous->name || tok->previous->op == ')') &&
(tok->next->number || tok->next->name || tok->next->op == '('));
}
static bool isAlternativeUnaryOp(const simplecpp::Token *tok, const std::string &alt)
{
return ((tok->name && tok->str() == alt) &&
(!tok->previous || tok->previous->op == '(') &&
(tok->next && (tok->next->name || tok->next->number)));
}
static std::string replaceAll(std::string s, const std::string& from, const std::string& to)
{
for (size_t pos = s.find(from); pos != std::string::npos; pos = s.find(from, pos + to.size()))
s.replace(pos, from.size(), to);
return s;
}
void simplecpp::Location::adjust(const std::string &str)
{
if (strpbrk(str.c_str(), "\r\n") == nullptr) {
col += str.size();
return;
}
for (std::size_t i = 0U; i < str.size(); ++i) {
col++;
if (str[i] == '\n' || str[i] == '\r') {
col = 1;
line++;
if (str[i] == '\r' && (i+1)<str.size() && str[i+1]=='\n')
++i;
}
}
}
bool simplecpp::Token::isOneOf(const char ops[]) const
{
return (op != '\0') && (std::strchr(ops, op) != nullptr);
}
bool simplecpp::Token::startsWithOneOf(const char c[]) const
{
return std::strchr(c, string[0]) != nullptr;
}
bool simplecpp::Token::endsWithOneOf(const char c[]) const
{
return std::strchr(c, string[string.size() - 1U]) != nullptr;
}
void simplecpp::Token::printAll() const
{
const Token *tok = this;
while (tok->previous)
tok = tok->previous;
for (; tok; tok = tok->next) {
if (tok->previous) {
std::cout << (sameline(tok, tok->previous) ? ' ' : '\n');
}
std::cout << tok->str();
}
std::cout << std::endl;
}
void simplecpp::Token::printOut() const
{
for (const Token *tok = this; tok; tok = tok->next) {
if (tok != this) {
std::cout << (sameline(tok, tok->previous) ? ' ' : '\n');
}
std::cout << tok->str();
}
std::cout << std::endl;
}
// cppcheck-suppress noConstructor - we call init() in the inherited to initialize the private members
class simplecpp::TokenList::Stream {
public:
virtual ~Stream() = default;
virtual int get() = 0;
virtual int peek() = 0;
virtual void unget() = 0;
virtual bool good() = 0;
unsigned char readChar() {
auto ch = static_cast<unsigned char>(get());
// For UTF-16 encoded files the BOM is 0xfeff/0xfffe. If the
// character is non-ASCII character then replace it with 0xff
if (isUtf16) {
const auto ch2 = static_cast<unsigned char>(get());
const int ch16 = makeUtf16Char(ch, ch2);
ch = static_cast<unsigned char>(((ch16 >= 0x80) ? 0xff : ch16));
}
// Handling of newlines..
if (ch == '\r') {
ch = '\n';
int ch2 = get();
if (isUtf16) {
const int c2 = get();
ch2 = makeUtf16Char(ch2, c2);
}
if (ch2 != '\n')
ungetChar();
}
return ch;
}
unsigned char peekChar() {
auto ch = static_cast<unsigned char>(peek());
// For UTF-16 encoded files the BOM is 0xfeff/0xfffe. If the
// character is non-ASCII character then replace it with 0xff
if (isUtf16) {
(void)get();
const auto ch2 = static_cast<unsigned char>(peek());
unget();
const int ch16 = makeUtf16Char(ch, ch2);
ch = static_cast<unsigned char>(((ch16 >= 0x80) ? 0xff : ch16));
}
// Handling of newlines..
if (ch == '\r')
ch = '\n';
return ch;
}
void ungetChar() {
unget();
if (isUtf16)
unget();
}
protected:
void init() {
// initialize since we use peek() in getAndSkipBOM()
isUtf16 = false;
bom = getAndSkipBOM();
isUtf16 = (bom == 0xfeff || bom == 0xfffe);
}
private:
inline int makeUtf16Char(const unsigned char ch, const unsigned char ch2) const {
return (bom == 0xfeff) ? (ch<<8 | ch2) : (ch2<<8 | ch);
}
unsigned short getAndSkipBOM() {
const int ch1 = peek();
// The UTF-16 BOM is 0xfffe or 0xfeff.
if (ch1 >= 0xfe) {
(void)get();
const unsigned short byte = (static_cast<unsigned char>(ch1) << 8);
if (peek() >= 0xfe)
return byte | static_cast<unsigned char>(get());
unget();
return 0;
}
// Skip UTF-8 BOM 0xefbbbf
if (ch1 == 0xef) {
(void)get();
if (peek() == 0xbb) {
(void)get();
if (peek() == 0xbf) {
(void)get();
return 0;
}
unget();
}
unget();
}
return 0;
}
unsigned short bom;
protected:
bool isUtf16;
};
namespace {
class StdIStream : public simplecpp::TokenList::Stream {
public:
// cppcheck-suppress uninitDerivedMemberVar - we call Stream::init() to initialize the private members
explicit StdIStream(std::istream &istr)
: istr(istr) {
assert(istr.good());
init();
}
int get() override {
return istr.get();
}
int peek() override {
return istr.peek();
}
void unget() override {
istr.unget();
}
bool good() override {
return istr.good();
}
private:
std::istream &istr;
};
class StdCharBufStream : public simplecpp::TokenList::Stream {
public:
// cppcheck-suppress uninitDerivedMemberVar - we call Stream::init() to initialize the private members
StdCharBufStream(const unsigned char* str, std::size_t size)
: str(str)
, size(size)
{
init();
}
int get() override {
if (pos >= size)
return lastStatus = EOF;
return str[pos++];
}
int peek() override {
if (pos >= size)
return lastStatus = EOF;
return str[pos];
}
void unget() override {
--pos;
}
bool good() override {
return lastStatus != EOF;
}
private:
const unsigned char *str;
const std::size_t size;
std::size_t pos{};
int lastStatus{};
};
class FileStream : public simplecpp::TokenList::Stream {
public:
/**
* @throws simplecpp::Output thrown if file is not found
*/
// cppcheck-suppress uninitDerivedMemberVar - we call Stream::init() to initialize the private members
explicit FileStream(const std::string &filename, std::vector<std::string> &files)
: file(fopen(filename.c_str(), "rb"))
{
if (!file) {
files.emplace_back(filename);
throw simplecpp::Output(simplecpp::Output::FILE_NOT_FOUND, {}, "File is missing: " + filename);
}
init();
}
FileStream(const FileStream&) = delete;
FileStream &operator=(const FileStream&) = delete;
~FileStream() override {
fclose(file);
file = nullptr;
}
int get() override {
lastStatus = lastCh = fgetc(file);
return lastCh;
}
int peek() override {
// keep lastCh intact
const int ch = fgetc(file);
unget_internal(ch);
return ch;
}
void unget() override {
unget_internal(lastCh);
}
bool good() override {
return lastStatus != EOF;
}
private:
void unget_internal(int ch) {
if (isUtf16) {
// TODO: use ungetc() as well
// UTF-16 has subsequent unget() calls
fseek(file, -1, SEEK_CUR);
} else {
ungetc(ch, file);
}
}
FILE *file;
int lastCh{};
int lastStatus{};
};
}
simplecpp::TokenList::TokenList(std::vector<std::string> &filenames) : frontToken(nullptr), backToken(nullptr), files(filenames) {}
simplecpp::TokenList::TokenList(std::istream &istr, std::vector<std::string> &filenames, const std::string &filename, OutputList *outputList)
: frontToken(nullptr), backToken(nullptr), files(filenames)
{
StdIStream stream(istr);
readfile(stream,filename,outputList);
}
simplecpp::TokenList::TokenList(const unsigned char* data, std::size_t size, std::vector<std::string> &filenames, const std::string &filename, OutputList *outputList, int /*unused*/)
: frontToken(nullptr), backToken(nullptr), files(filenames)
{
StdCharBufStream stream(data, size);
readfile(stream,filename,outputList);
}
simplecpp::TokenList::TokenList(const std::string &filename, std::vector<std::string> &filenames, OutputList *outputList)
: frontToken(nullptr), backToken(nullptr), files(filenames)
{
try {
FileStream stream(filename, filenames);
readfile(stream,filename,outputList);
} catch (const simplecpp::Output & e) {
outputList->emplace_back(e);
}
}
simplecpp::TokenList::TokenList(const TokenList &other) : frontToken(nullptr), backToken(nullptr), files(other.files)
{
*this = other;
}
simplecpp::TokenList::TokenList(TokenList &&other) : frontToken(nullptr), backToken(nullptr), files(other.files)
{
*this = std::move(other);
}
simplecpp::TokenList::~TokenList()
{
clear();
}
simplecpp::TokenList &simplecpp::TokenList::operator=(const TokenList &other)
{
if (this != &other) {
clear();
files = other.files;
for (const Token *tok = other.cfront(); tok; tok = tok->next)
push_back(new Token(*tok));
sizeOfType = other.sizeOfType;
}
return *this;
}
simplecpp::TokenList &simplecpp::TokenList::operator=(TokenList &&other)
{
if (this != &other) {
clear();
frontToken = other.frontToken;
other.frontToken = nullptr;
backToken = other.backToken;
other.backToken = nullptr;
files = other.files;
sizeOfType = std::move(other.sizeOfType);
}
return *this;
}
void simplecpp::TokenList::clear()
{
backToken = nullptr;
while (frontToken) {
Token * const next = frontToken->next;
delete frontToken;
frontToken = next;
}
sizeOfType.clear();
}
void simplecpp::TokenList::push_back(Token *tok)
{
if (!frontToken)
frontToken = tok;
else
backToken->next = tok;
tok->previous = backToken;
backToken = tok;
}
void simplecpp::TokenList::dump(bool linenrs) const
{
std::cout << stringify(linenrs) << std::endl;
}
std::string simplecpp::TokenList::stringify(bool linenrs) const
{
std::ostringstream ret;
Location loc;
loc.line = 1;
bool filechg = true;
for (const Token *tok = cfront(); tok; tok = tok->next) {
if (tok->location.line < loc.line || tok->location.fileIndex != loc.fileIndex) {
ret << "\n#line " << tok->location.line << " \"" << file(tok->location) << "\"\n";
loc = tok->location;
filechg = true;
}
if (linenrs && filechg) {
ret << loc.line << ": ";
filechg = false;
}
while (tok->location.line > loc.line) {
ret << '\n';
loc.line++;
if (linenrs)
ret << loc.line << ": ";
}
if (sameline(tok->previous, tok))
ret << ' ';
ret << tok->str();
loc.adjust(tok->str());
}
return ret.str();
}
static bool isNameChar(unsigned char ch)
{
return std::isalnum(ch) || ch == '_' || ch == '$';
}
static std::string escapeString(const std::string &str)
{
std::ostringstream ostr;
ostr << '\"';
for (std::size_t i = 1U; i < str.size() - 1; ++i) {
const char c = str[i];
if (c == '\\' || c == '\"' || c == '\'')
ostr << '\\';
ostr << c;
}
ostr << '\"';
return ostr.str();
}
static void portabilityBackslash(simplecpp::OutputList *outputList, const simplecpp::Location &location)
{
if (!outputList)
return;
simplecpp::Output err{
simplecpp::Output::PORTABILITY_BACKSLASH,
location,
"Combination 'backslash space newline' is not portable."
};
outputList->emplace_back(std::move(err));
}
static bool isStringLiteralPrefix(const std::string &str)
{
return str == "u" || str == "U" || str == "L" || str == "u8" ||
str == "R" || str == "uR" || str == "UR" || str == "LR" || str == "u8R";
}
void simplecpp::TokenList::lineDirective(unsigned int fileIndex, unsigned int line, Location &location)
{
if (fileIndex != location.fileIndex || line >= location.line) {
location.fileIndex = fileIndex;
location.line = line;
return;
}
if (line + 2 >= location.line) {
location.line = line;
while (cback()->op != '#')
deleteToken(back());
deleteToken(back());
return;
}
}
static const std::string COMMENT_END("*/");
void simplecpp::TokenList::readfile(Stream &stream, const std::string &filename, OutputList *outputList)
{
std::stack<simplecpp::Location> loc;
unsigned int multiline = 0U;
const Token *oldLastToken = nullptr;
Location location(fileIndex(filename), 1, 1);
while (stream.good()) {
unsigned char ch = stream.readChar();
if (!stream.good())
break;
if (ch >= 0x80) {
if (outputList) {
simplecpp::Output err{
simplecpp::Output::UNHANDLED_CHAR_ERROR,
location,
"The code contains unhandled character(s) (character code=" + std::to_string(static_cast<int>(ch)) + "). Neither unicode nor extended ascii is supported."
};
outputList->emplace_back(std::move(err));
}
clear();
return;
}
if (ch == '\n') {
if (cback() && cback()->op == '\\') {
if (location.col > cback()->location.col + 1U)
portabilityBackslash(outputList, cback()->location);
++multiline;
deleteToken(back());
} else {
location.line += multiline + 1;
multiline = 0U;
}
if (!multiline)
location.col = 1;
if (oldLastToken != cback()) {
oldLastToken = cback();
const Token * const llTok = isLastLinePreprocessor();
if (!llTok)
continue;
const Token * const llNextToken = llTok->next;
if (!llTok->next)
continue;
if (llNextToken->next) {
// #file "file.c"
if (llNextToken->str() == "file" &&
llNextToken->next->str()[0] == '\"')
{
const Token *strtok = cback();
while (strtok->comment)
strtok = strtok->previous;
loc.push(location);
location.fileIndex = fileIndex(strtok->str().substr(1U, strtok->str().size() - 2U));
location.line = 1U;
}
// TODO: add support for "# 3"
// #3 "file.c"
// #line 3 "file.c"
else if ((llNextToken->number &&
llNextToken->next->str()[0] == '\"') ||
(llNextToken->str() == "line" &&
llNextToken->next->number &&
llNextToken->next->next &&
llNextToken->next->next->str()[0] == '\"'))
{
const Token *strtok = cback();
while (strtok->comment)
strtok = strtok->previous;
const Token *numtok = strtok->previous;
while (numtok->comment)
numtok = numtok->previous;
lineDirective(fileIndex(replaceAll(strtok->str().substr(1U, strtok->str().size() - 2U),"\\\\","\\")),
std::atol(numtok->str().c_str()), location);
}
// #line 3
else if (llNextToken->str() == "line" &&
llNextToken->next->number)
{
const Token *numtok = cback();
while (numtok->comment)
numtok = numtok->previous;
lineDirective(location.fileIndex, std::atol(numtok->str().c_str()), location);
}
}
// #endfile
else if (llNextToken->str() == "endfile" && !loc.empty())
{
location = loc.top();
loc.pop();
}
}
continue;
}
if (ch <= ' ') {
location.col++;
continue;
}
TokenString currentToken;
if (cback() && cback()->location.line == location.line && cback()->previous && cback()->previous->op == '#') {
const Token* const ppTok = cback()->previous;
if (ppTok->next && (ppTok->next->str() == "error" || ppTok->next->str() == "warning")) {
char prev = ' ';
while (stream.good() && (prev == '\\' || (ch != '\r' && ch != '\n'))) {
currentToken += ch;
prev = ch;
ch = stream.readChar();
}
stream.ungetChar();
push_back(new Token(currentToken, location));
location.adjust(currentToken);
continue;
}
}
// number or name
if (isNameChar(ch)) {
const bool num = !!std::isdigit(ch);
while (stream.good() && isNameChar(ch)) {
currentToken += ch;
ch = stream.readChar();
if (num && ch=='\'' && isNameChar(stream.peekChar()))
ch = stream.readChar();
}
stream.ungetChar();
}
// comment
else if (ch == '/' && stream.peekChar() == '/') {
while (stream.good() && ch != '\n') {
currentToken += ch;
ch = stream.readChar();
if (ch == '\\') {
TokenString tmp;
char tmp_ch = ch;
while ((stream.good()) && (tmp_ch == '\\' || tmp_ch == ' ' || tmp_ch == '\t')) {
tmp += tmp_ch;
tmp_ch = stream.readChar();
}
if (!stream.good()) {
break;
}
if (tmp_ch != '\n') {
currentToken += tmp;
} else {
const TokenString check_portability = currentToken + tmp;
const std::string::size_type pos = check_portability.find_last_not_of(" \t");
if (pos < check_portability.size() - 1U && check_portability[pos] == '\\')
portabilityBackslash(outputList, location);
++multiline;
tmp_ch = stream.readChar();
currentToken += '\n';
}
ch = tmp_ch;
}
}
if (ch == '\n') {
stream.ungetChar();
}
}
// comment
else if (ch == '/' && stream.peekChar() == '*') {
currentToken = "/*";
(void)stream.readChar();
ch = stream.readChar();
while (stream.good()) {
currentToken += ch;
if (currentToken.size() >= 4U && endsWith(currentToken, COMMENT_END))
break;
ch = stream.readChar();
}
// multiline..
std::string::size_type pos = 0;
while ((pos = currentToken.find("\\\n",pos)) != std::string::npos) {
currentToken.erase(pos,2);
++multiline;
}
if (multiline || isLastLinePreprocessor()) {
pos = 0;
while ((pos = currentToken.find('\n',pos)) != std::string::npos) {
currentToken.erase(pos,1);
++multiline;
}
}
}
// string / char literal
else if (ch == '\"' || ch == '\'') {
std::string prefix;
if (cback() && cback()->name && isStringLiteralPrefix(cback()->str()) &&
((cback()->location.col + cback()->str().size()) == location.col) &&
(cback()->location.line == location.line)) {
prefix = cback()->str();
}
// C++11 raw string literal
if (ch == '\"' && !prefix.empty() && *cback()->str().rbegin() == 'R') {
std::string delim;
currentToken = ch;
prefix.resize(prefix.size() - 1);
ch = stream.readChar();
while (stream.good() && ch != '(' && ch != '\n') {
delim += ch;
ch = stream.readChar();
}
if (!stream.good() || ch == '\n') {
if (outputList) {
Output err{
Output::SYNTAX_ERROR,
location,
"Invalid newline in raw string delimiter."
};
outputList->emplace_back(std::move(err));
}
return;
}
const std::string endOfRawString(')' + delim + currentToken);
while (stream.good() && (!endsWith(currentToken, endOfRawString) || currentToken.size() <= 1))
currentToken += stream.readChar();
if (!endsWith(currentToken, endOfRawString)) {
if (outputList) {
Output err{
Output::SYNTAX_ERROR,
location,
"Raw string missing terminating delimiter."
};
outputList->emplace_back(std::move(err));
}
return;
}
currentToken.erase(currentToken.size() - endOfRawString.size(), endOfRawString.size() - 1U);
currentToken = escapeString(currentToken);
currentToken.insert(0, prefix);
back()->setstr(currentToken);
location.adjust(currentToken);
if (currentToken.find_first_of("\r\n") == std::string::npos)
location.col += 2 + (2 * delim.size());
else
location.col += 1 + delim.size();
continue;
}
currentToken = readUntil(stream,location,ch,ch,outputList);
if (currentToken.size() < 2U)
// Error is reported by readUntil()
return;
std::string s = currentToken;
std::string::size_type pos;
int newlines = 0;
while ((pos = s.find_first_of("\r\n")) != std::string::npos) {
s.erase(pos,1);
newlines++;
}
if (prefix.empty())
push_back(new Token(s, location, !!std::isspace(stream.peekChar()))); // push string without newlines
else
back()->setstr(prefix + s);
if (newlines > 0) {
const Token * const llTok = lastLineTok();
if (llTok && llTok->op == '#' && llTok->next && (llTok->next->str() == "define" || llTok->next->str() == "pragma") && llTok->next->next) {
multiline += newlines;
location.adjust(s);
continue;
}
}
location.adjust(currentToken);
continue;
}
else {
currentToken += ch;
}
if (*currentToken.begin() == '<') {
const Token * const llTok = lastLineTok();
if (llTok && llTok->op == '#' && llTok->next && llTok->next->str() == "include") {
currentToken = readUntil(stream, location, '<', '>', outputList);
if (currentToken.size() < 2U)
return;
}
}
push_back(new Token(currentToken, location, !!std::isspace(stream.peekChar())));
if (multiline)
location.col += currentToken.size();
else
location.adjust(currentToken);
}
combineOperators();
}
void simplecpp::TokenList::constFold()
{
while (cfront()) {
// goto last '('
Token *tok = back();
while (tok && tok->op != '(')
tok = tok->previous;
// no '(', goto first token
if (!tok)
tok = front();
// Constant fold expression
constFoldUnaryNotPosNeg(tok);
constFoldMulDivRem(tok);
constFoldAddSub(tok);
constFoldShift(tok);
constFoldComparison(tok);
constFoldBitwise(tok);
constFoldLogicalOp(tok);
constFoldQuestionOp(tok);
// If there is no '(' we are done with the constant folding
if (tok->op != '(')
break;
if (!tok->next || !tok->next->next || tok->next->next->op != ')')
break;
tok = tok->next;
deleteToken(tok->previous);
deleteToken(tok->next);
}
}
static bool isFloatSuffix(const simplecpp::Token *tok)
{
if (!tok || tok->str().size() != 1U)
return false;
const char c = std::tolower(tok->str()[0]);
return c == 'f' || c == 'l';
}
static const std::string AND("and");
static const std::string BITAND("bitand");
static const std::string BITOR("bitor");
static bool isAlternativeAndBitandBitor(const simplecpp::Token* tok)
{
return isAlternativeBinaryOp(tok, AND) || isAlternativeBinaryOp(tok, BITAND) || isAlternativeBinaryOp(tok, BITOR);
}
void simplecpp::TokenList::combineOperators()
{
std::stack<bool> executableScope;
executableScope.push(false);
for (Token *tok = front(); tok; tok = tok->next) {
if (tok->op == '{') {
if (executableScope.top()) {
executableScope.push(true);
continue;
}
const Token *prev = tok->previous;
while (prev && prev->isOneOf(";{}()"))
prev = prev->previous;
executableScope.push(prev && prev->op == ')');
continue;
}
if (tok->op == '}') {
if (executableScope.size() > 1)
executableScope.pop();
continue;
}
if (tok->op == '.') {
// ellipsis ...
if (tok->next && tok->next->op == '.' && tok->next->location.col == (tok->location.col + 1) &&
tok->next->next && tok->next->next->op == '.' && tok->next->next->location.col == (tok->location.col + 2)) {
tok->setstr("...");
deleteToken(tok->next);
deleteToken(tok->next);
continue;
}
// float literals..
if (tok->previous && tok->previous->number && sameline(tok->previous, tok) && tok->previous->str().find_first_of("._") == std::string::npos) {
tok->setstr(tok->previous->str() + '.');
deleteToken(tok->previous);
if (sameline(tok, tok->next) && (isFloatSuffix(tok->next) || (tok->next && tok->next->startsWithOneOf("AaBbCcDdEeFfPp") && !isAlternativeAndBitandBitor(tok->next)))) {
tok->setstr(tok->str() + tok->next->str());
deleteToken(tok->next);
}
}
if (tok->next && tok->next->number) {
tok->setstr(tok->str() + tok->next->str());
deleteToken(tok->next);
}
}
// match: [0-9.]+E [+-] [0-9]+
const char lastChar = tok->str()[tok->str().size() - 1];
if (tok->number && !isOct(tok->str()) &&
((!isHex(tok->str()) && (lastChar == 'E' || lastChar == 'e')) ||
(isHex(tok->str()) && (lastChar == 'P' || lastChar == 'p'))) &&
tok->next && tok->next->isOneOf("+-") && tok->next->next && tok->next->next->number) {
tok->setstr(tok->str() + tok->next->op + tok->next->next->str());
deleteToken(tok->next);
deleteToken(tok->next);
}
if (tok->op == '\0' || !tok->next || tok->next->op == '\0')
continue;
if (!sameline(tok,tok->next))
continue;
if (tok->location.col + 1U != tok->next->location.col)
continue;
if (tok->next->op == '=' && tok->isOneOf("=!<>+-*/%&|^")) {
if (tok->op == '&' && !executableScope.top()) {
// don't combine &= if it is a anonymous reference parameter with default value:
// void f(x&=2)
int indentlevel = 0;
const Token *start = tok;
while (indentlevel >= 0 && start) {
if (start->op == ')')
++indentlevel;
else if (start->op == '(')
--indentlevel;
else if (start->isOneOf(";{}"))
break;
start = start->previous;
}
if (indentlevel == -1 && start) {
const Token * const ftok = start;
bool isFuncDecl = ftok->name;
while (isFuncDecl) {
if (!start->name && start->str() != "::" && start->op != '*' && start->op != '&')
isFuncDecl = false;
if (!start->previous)
break;
if (start->previous->isOneOf(";{}:"))
break;
start = start->previous;
}
isFuncDecl &= start != ftok && start->name;
if (isFuncDecl) {
// TODO: we could loop through the parameters here and check if they are correct.
continue;
}
}
}
tok->setstr(tok->str() + "=");
deleteToken(tok->next);
} else if ((tok->op == '|' || tok->op == '&') && tok->op == tok->next->op) {
tok->setstr(tok->str() + tok->next->str());
deleteToken(tok->next);
} else if (tok->op == ':' && tok->next->op == ':') {
tok->setstr(tok->str() + tok->next->str());
deleteToken(tok->next);
} else if (tok->op == '-' && tok->next->op == '>') {
tok->setstr(tok->str() + tok->next->str());
deleteToken(tok->next);
} else if ((tok->op == '<' || tok->op == '>') && tok->op == tok->next->op) {
tok->setstr(tok->str() + tok->next->str());
deleteToken(tok->next);
if (tok->next && tok->next->op == '=' && tok->next->next && tok->next->next->op != '=') {
tok->setstr(tok->str() + tok->next->str());
deleteToken(tok->next);
}
} else if ((tok->op == '+' || tok->op == '-') && tok->op == tok->next->op) {
if (tok->location.col + 1U != tok->next->location.col)
continue;
if (tok->previous && tok->previous->number)
continue;
if (tok->next->next && tok->next->next->number)
continue;
tok->setstr(tok->str() + tok->next->str());
deleteToken(tok->next);
}
}
}
static const std::string COMPL("compl");
static const std::string NOT("not");
void simplecpp::TokenList::constFoldUnaryNotPosNeg(simplecpp::Token *tok)
{
for (; tok && tok->op != ')'; tok = tok->next) {
// "not" might be !
if (isAlternativeUnaryOp(tok, NOT))
tok->op = '!';
// "compl" might be ~
else if (isAlternativeUnaryOp(tok, COMPL))
tok->op = '~';
if (tok->op == '!' && tok->next && tok->next->number) {
tok->setstr(tok->next->str() == "0" ? "1" : "0");
deleteToken(tok->next);
} else if (tok->op == '~' && tok->next && tok->next->number) {
tok->setstr(toString(~stringToLL(tok->next->str())));
deleteToken(tok->next);
} else {
if (tok->previous && (tok->previous->number || tok->previous->name))
continue;
if (!tok->next || !tok->next->number)
continue;
switch (tok->op) {
case '+':
tok->setstr(tok->next->str());
deleteToken(tok->next);
break;
case '-':
tok->setstr(tok->op + tok->next->str());
deleteToken(tok->next);
break;
}
}
}
}
void simplecpp::TokenList::constFoldMulDivRem(Token *tok)
{
for (; tok && tok->op != ')'; tok = tok->next) {
if (!tok->previous || !tok->previous->number)
continue;
if (!tok->next || !tok->next->number)
continue;
long long result;
if (tok->op == '*') {
result = (stringToLL(tok->previous->str()) * stringToLL(tok->next->str()));
}
else if (tok->op == '/' || tok->op == '%') {
const long long rhs = stringToLL(tok->next->str());
if (rhs == 0)
throw std::overflow_error("division/modulo by zero");
const long long lhs = stringToLL(tok->previous->str());
if (rhs == -1 && lhs == std::numeric_limits<long long>::min())
throw std::overflow_error("division overflow");
if (tok->op == '/')
result = (lhs / rhs);
else
result = (lhs % rhs);
} else {
continue;
}
tok = tok->previous;
tok->setstr(toString(result));
deleteToken(tok->next);
deleteToken(tok->next);
}
}
void simplecpp::TokenList::constFoldAddSub(Token *tok)
{
for (; tok && tok->op != ')'; tok = tok->next) {
if (!tok->previous || !tok->previous->number)
continue;
if (!tok->next || !tok->next->number)
continue;
long long result;
if (tok->op == '+')
result = stringToLL(tok->previous->str()) + stringToLL(tok->next->str());
else if (tok->op == '-')
result = stringToLL(tok->previous->str()) - stringToLL(tok->next->str());
else
continue;
tok = tok->previous;
tok->setstr(toString(result));
deleteToken(tok->next);
deleteToken(tok->next);
}
}
void simplecpp::TokenList::constFoldShift(Token *tok)
{
for (; tok && tok->op != ')'; tok = tok->next) {
if (!tok->previous || !tok->previous->number)
continue;
if (!tok->next || !tok->next->number)
continue;
long long result;
if (tok->str() == "<<")
result = stringToLL(tok->previous->str()) << stringToLL(tok->next->str());
else if (tok->str() == ">>")
result = stringToLL(tok->previous->str()) >> stringToLL(tok->next->str());
else
continue;
tok = tok->previous;
tok->setstr(toString(result));
deleteToken(tok->next);
deleteToken(tok->next);
}
}
static const std::string NOTEQ("not_eq");
void simplecpp::TokenList::constFoldComparison(Token *tok)
{
for (; tok && tok->op != ')'; tok = tok->next) {
if (isAlternativeBinaryOp(tok,NOTEQ))
tok->setstr("!=");
if (!tok->startsWithOneOf("<>=!"))
continue;
if (!tok->previous || !tok->previous->number)
continue;
if (!tok->next || !tok->next->number)
continue;
int result;
if (tok->str() == "==")
result = (stringToLL(tok->previous->str()) == stringToLL(tok->next->str()));
else if (tok->str() == "!=")
result = (stringToLL(tok->previous->str()) != stringToLL(tok->next->str()));
else if (tok->str() == ">")
result = (stringToLL(tok->previous->str()) > stringToLL(tok->next->str()));
else if (tok->str() == ">=")
result = (stringToLL(tok->previous->str()) >= stringToLL(tok->next->str()));
else if (tok->str() == "<")
result = (stringToLL(tok->previous->str()) < stringToLL(tok->next->str()));
else if (tok->str() == "<=")
result = (stringToLL(tok->previous->str()) <= stringToLL(tok->next->str()));
else
continue;
tok = tok->previous;
tok->setstr(toString(result));
deleteToken(tok->next);
deleteToken(tok->next);
}
}
static const std::string XOR("xor");
void simplecpp::TokenList::constFoldBitwise(Token *tok)
{
Token * const tok1 = tok;
for (const char *op = "&^|"; *op; op++) {
const std::string* alternativeOp;
if (*op == '&')
alternativeOp = &BITAND;
else if (*op == '|')
alternativeOp = &BITOR;
else
alternativeOp = &XOR;
for (tok = tok1; tok && tok->op != ')'; tok = tok->next) {
if (tok->op != *op && !isAlternativeBinaryOp(tok, *alternativeOp))
continue;
if (!tok->previous || !tok->previous->number)
continue;
if (!tok->next || !tok->next->number)
continue;
long long result;
if (*op == '&')
result = (stringToLL(tok->previous->str()) & stringToLL(tok->next->str()));
else if (*op == '^')
result = (stringToLL(tok->previous->str()) ^ stringToLL(tok->next->str()));
else /*if (*op == '|')*/
result = (stringToLL(tok->previous->str()) | stringToLL(tok->next->str()));
tok = tok->previous;
tok->setstr(toString(result));
deleteToken(tok->next);
deleteToken(tok->next);
}
}
}
static const std::string OR("or");
void simplecpp::TokenList::constFoldLogicalOp(Token *tok)
{
for (; tok && tok->op != ')'; tok = tok->next) {
if (tok->name) {
if (isAlternativeBinaryOp(tok,AND))
tok->setstr("&&");
else if (isAlternativeBinaryOp(tok,OR))
tok->setstr("||");
}
if (tok->str() != "&&" && tok->str() != "||")
continue;
if (!tok->previous || !tok->previous->number)
continue;
if (!tok->next || !tok->next->number)
continue;
int result;
if (tok->str() == "||")
result = (stringToLL(tok->previous->str()) || stringToLL(tok->next->str()));
else /*if (tok->str() == "&&")*/
result = (stringToLL(tok->previous->str()) && stringToLL(tok->next->str()));
tok = tok->previous;
tok->setstr(toString(result));
deleteToken(tok->next);
deleteToken(tok->next);
}
}
void simplecpp::TokenList::constFoldQuestionOp(Token *&tok1)
{
bool gotoTok1 = false;
// NOLINTNEXTLINE(misc-const-correctness) - technically correct but used to access non-const data
for (Token *tok = tok1; tok && tok->op != ')'; tok = gotoTok1 ? tok1 : tok->next) {
gotoTok1 = false;
if (tok->str() != "?")
continue;
if (!tok->previous || !tok->next || !tok->next->next)
throw std::runtime_error("invalid expression");
if (!tok->previous->number)
continue;
if (tok->next->next->op != ':')
continue;
Token * const condTok = tok->previous;
Token * const trueTok = tok->next;
Token * const falseTok = trueTok->next->next;
if (!falseTok)
throw std::runtime_error("invalid expression");
if (condTok == tok1)
tok1 = (condTok->str() != "0" ? trueTok : falseTok);
deleteToken(condTok->next); // ?
deleteToken(trueTok->next); // :
deleteToken(condTok->str() == "0" ? trueTok : falseTok);
deleteToken(condTok);
gotoTok1 = true;
}
}
void simplecpp::TokenList::removeComments()
{
Token *tok = frontToken;
while (tok) {
Token * const tok1 = tok;
tok = tok->next;
if (tok1->comment)
deleteToken(tok1);
}
}
std::string simplecpp::TokenList::readUntil(Stream &stream, const Location &location, const char start, const char end, OutputList *outputList)
{
std::string ret;
ret += start;
bool backslash = false;
char ch = 0;
while (ch != end && ch != '\r' && ch != '\n' && stream.good()) {
ch = stream.readChar();
if (backslash && ch == '\n') {
ch = 0;
backslash = false;
continue;
}
backslash = false;
ret += ch;
if (ch == '\\') {
bool update_ch = false;
char next = 0;
do {
next = stream.readChar();
if (next == '\r' || next == '\n') {
ret.erase(ret.size()-1U);
backslash = (next == '\r');
update_ch = false;
} else if (next == '\\') {
update_ch = !update_ch;
}
ret += next;
} while (next == '\\');
if (update_ch)
ch = next;
}
}
if (!stream.good() || ch != end) {
clear();
if (outputList) {
Output err{
Output::SYNTAX_ERROR,
location,
std::string("No pair for character (") + start + "). Can't process file. File is either invalid or unicode, which is currently not supported."
};
outputList->emplace_back(std::move(err));
}
return "";
}
return ret;
}
const simplecpp::Token* simplecpp::TokenList::lastLineTok(int maxsize) const
{
const Token* prevTok = nullptr;
int count = 0;
for (const Token *tok = cback(); ; tok = tok->previous) {
if (!sameline(tok, cback()))
break;
if (tok->comment)
continue;
if (++count > maxsize)
return nullptr;
prevTok = tok;
}
return prevTok;
}
const simplecpp::Token* simplecpp::TokenList::isLastLinePreprocessor(int maxsize) const
{
const Token * const prevTok = lastLineTok(maxsize);
if (prevTok && prevTok->op == '#')
return prevTok;
return nullptr;
}
unsigned int simplecpp::TokenList::fileIndex(const std::string &filename)
{
for (unsigned int i = 0; i < files.size(); ++i) {
if (files[i] == filename)
return i;
}
files.emplace_back(filename);
return files.size() - 1U;
}
const std::string& simplecpp::TokenList::file(const Location& loc) const
{
static const std::string s_emptyFileName;
return loc.fileIndex < files.size() ? files[loc.fileIndex] : s_emptyFileName;
}
namespace simplecpp {
class Macro;
using MacroMap = std::unordered_map<TokenString,Macro>;
class Macro {
public:
explicit Macro(std::vector<std::string> &f) : nameTokDef(nullptr), valueToken(nullptr), endToken(nullptr), files(f), tokenListDefine(f), variadic(false), variadicOpt(false), valueDefinedInCode_(false) {}
/**
* @throws std::runtime_error thrown on bad macro syntax
*/
Macro(const Token *tok, std::vector<std::string> &f) : nameTokDef(nullptr), files(f), tokenListDefine(f), valueDefinedInCode_(true) {
if (sameline(tok->previousSkipComments(), tok))
throw std::runtime_error("bad macro syntax");
if (tok->op != '#')
throw std::runtime_error("bad macro syntax");
const Token * const hashtok = tok;
tok = tok->next;
if (!tok || tok->str() != DEFINE)
throw std::runtime_error("bad macro syntax");
tok = tok->next;
if (!tok || !tok->name || !sameline(hashtok,tok))
throw std::runtime_error("bad macro syntax");
if (!parseDefine(tok))
throw std::runtime_error("bad macro syntax");
}
/**
* @throws std::runtime_error thrown on bad macro syntax
*/
Macro(const std::string &name, const std::string &value, std::vector<std::string> &f) : nameTokDef(nullptr), files(f), tokenListDefine(f), valueDefinedInCode_(false) {
const std::string def(name + ' ' + value);
StdCharBufStream stream(reinterpret_cast<const unsigned char*>(def.data()), def.size());
tokenListDefine.readfile(stream);
if (!parseDefine(tokenListDefine.cfront()))
throw std::runtime_error("bad macro syntax. macroname=" + name + " value=" + value);
}
Macro(const Macro &other) : nameTokDef(nullptr), files(other.files), tokenListDefine(other.files), valueDefinedInCode_(other.valueDefinedInCode_) {
// TODO: remove the try-catch - see #537
// avoid bugprone-exception-escape clang-tidy warning
try {
*this = other;
}
catch (const Error&) {} // NOLINT(bugprone-empty-catch)
}
~Macro() {
delete optExpandValue;
delete optNoExpandValue;
}
Macro &operator=(const Macro &other) {
if (this != &other) {
files = other.files;
valueDefinedInCode_ = other.valueDefinedInCode_;
if (other.tokenListDefine.empty()) {
parseDefine(other.nameTokDef);
}
else {
tokenListDefine = other.tokenListDefine;
parseDefine(tokenListDefine.cfront());
}
usageList = other.usageList;
}
return *this;
}
bool valueDefinedInCode() const {
return valueDefinedInCode_;
}
/**
* Expand macro. This will recursively expand inner macros.
* @param output destination tokenlist
* @param rawtok macro token
* @param macros list of macros
* @param inputFiles the input files
* @return token after macro
* @throws Error thrown on missing or invalid preprocessor directives
* @throws wrongNumberOfParameters thrown on invalid number of parameters
* @throws invalidHashHash thrown on invalid ## usage
*/
const Token * expand(TokenList & output,
const Token * rawtok,
const MacroMap &macros,
std::vector<std::string> &inputFiles) const {
std::set<TokenString> expandedmacros;
#ifdef SIMPLECPP_DEBUG_MACRO_EXPANSION
std::cout << "expand " << name() << " " << locstring(rawtok->location) << std::endl;
#endif
TokenList output2(inputFiles);
if (functionLike() && rawtok->next && rawtok->next->op == '(') {
// Copy macro call to a new tokenlist with no linebreaks
const Token * const rawtok1 = rawtok;
TokenList rawtokens2(inputFiles);
rawtokens2.push_back(new Token(rawtok->str(), rawtok1->location, rawtok->whitespaceahead));
rawtok = rawtok->next;
rawtokens2.push_back(new Token(rawtok->str(), rawtok1->location, rawtok->whitespaceahead));
rawtok = rawtok->next;
int par = 1;
while (rawtok && par > 0) {
if (rawtok->op == '(')
++par;
else if (rawtok->op == ')')
--par;
else if (rawtok->op == '#' && !sameline(rawtok->previous, rawtok))
throw Error(rawtok->location, "it is invalid to use a preprocessor directive as macro parameter");
rawtokens2.push_back(new Token(rawtok->str(), rawtok1->location, rawtok->whitespaceahead));
rawtok = rawtok->next;
}
if (expand(output2, rawtok1->location, rawtokens2.cfront(), macros, expandedmacros))
rawtok = rawtok1->next;
} else {
rawtok = expand(output2, rawtok->location, rawtok, macros, expandedmacros);
}
while (output2.cback() && rawtok) {
unsigned int par = 0;
Token* macro2tok = output2.back();
while (macro2tok) {
if (macro2tok->op == '(') {
if (par==0)
break;
--par;
} else if (macro2tok->op == ')') {
++par;
}
macro2tok = macro2tok->previous;
}
if (macro2tok) { // macro2tok->op == '('
macro2tok = macro2tok->previous;
expandedmacros.insert(name());
} else if (rawtok->op == '(') {
macro2tok = output2.back();
}
if (!macro2tok || !macro2tok->name)
break;
if (output2.cfront() != output2.cback() && macro2tok->str() == this->name())
break;
const MacroMap::const_iterator macro = macros.find(macro2tok->str());
if (macro == macros.end() || !macro->second.functionLike())
break;
TokenList rawtokens2(inputFiles);
const Location loc(macro2tok->location);
while (macro2tok) {
Token * const next = macro2tok->next;
rawtokens2.push_back(new Token(macro2tok->str(), loc));
output2.deleteToken(macro2tok);
macro2tok = next;
}
par = (rawtokens2.cfront() != rawtokens2.cback()) ? 1U : 0U;
const Token *rawtok2 = rawtok;
for (; rawtok2; rawtok2 = rawtok2->next) {
rawtokens2.push_back(new Token(rawtok2->str(), loc));
if (rawtok2->op == '(') {
++par;
}
else if (rawtok2->op == ')') {
if (par <= 1U)
break;
--par;
}
}
if (!rawtok2 || par != 1U)
break;
if (macro->second.expand(output2, rawtok->location, rawtokens2.cfront(), macros, expandedmacros) != nullptr)
break;
rawtok = rawtok2->next;
}
output.takeTokens(output2);
return rawtok;
}
/** macro name */
const TokenString &name() const {
return nameTokDef->str();
}
/** location for macro definition */
const Location &defineLocation() const {
return nameTokDef->location;
}
/** how has this macro been used so far */
const std::list<Location> &usage() const {
return usageList;
}
/** is this a function like macro */
bool functionLike() const {
return nameTokDef->next &&
nameTokDef->next->op == '(' &&
sameline(nameTokDef, nameTokDef->next) &&
nameTokDef->next->location.col == nameTokDef->location.col + nameTokDef->str().size();
}
/** base class for errors */
struct Error {
Error(const Location &loc, const std::string &s) : location(loc), what(s) {}
const Location location;
const std::string what;
};
/** Struct that is thrown when macro is expanded with wrong number of parameters */
struct wrongNumberOfParameters : public Error {
wrongNumberOfParameters(const Location &loc, const std::string &macroName) : Error(loc, "Wrong number of parameters for macro \'" + macroName + "\'.") {}
};
/** Struct that is thrown when there is invalid ## usage */
struct invalidHashHash : public Error {
static inline std::string format(const std::string &macroName, const std::string &message) {
return "Invalid ## usage when expanding \'" + macroName + "\': " + message;
}
invalidHashHash(const Location &loc, const std::string &macroName, const std::string &message)
: Error(loc, format(macroName, message)) {}
static inline invalidHashHash unexpectedToken(const Location &loc, const std::string &macroName, const Token *tokenA) {
return {loc, macroName, "Unexpected token '"+ tokenA->str()+"'"};
}
static inline invalidHashHash cannotCombine(const Location &loc, const std::string &macroName, const Token *tokenA, const Token *tokenB) {
return {loc, macroName, "Combining '"+ tokenA->str()+ "' and '"+ tokenB->str() + "' yields an invalid token."};
}
static inline invalidHashHash unexpectedNewline(const Location &loc, const std::string &macroName) {
return {loc, macroName, "Unexpected newline"};
}
static inline invalidHashHash universalCharacterUB(const Location &loc, const std::string &macroName, const Token* tokenA, const std::string& strAB) {
return {loc, macroName, "Combining '\\"+ tokenA->str()+ "' and '"+ strAB.substr(tokenA->str().size()) + "' yields universal character '\\" + strAB + "'. This is undefined behavior according to C standard chapter 5.1.1.2, paragraph 4."};
}
};
private:
/** Create new token where Token::macro is set for replaced tokens */
Token *newMacroToken(const TokenString &str, const Location &loc, bool replaced, const Token *expandedFromToken=nullptr) const {
auto *tok = new Token(str,loc);
if (replaced)
tok->macro = nameTokDef->str();
if (expandedFromToken)
tok->setExpandedFrom(expandedFromToken, this);
return tok;
}
bool parseDefine(const Token *nametoken) {
nameTokDef = nametoken;
variadic = false;
variadicOpt = false;
delete optExpandValue;
optExpandValue = nullptr;
delete optNoExpandValue;
optNoExpandValue = nullptr;
if (!nameTokDef) {
valueToken = endToken = nullptr;
args.clear();
return false;
}
// function like macro..
if (functionLike()) {
args.clear();
const Token *argtok = nameTokDef->next->next;
while (sameline(nametoken, argtok) && argtok->op != ')') {
if (argtok->str() == "..." &&
argtok->next && argtok->next->op == ')') {
variadic = true;
if (!argtok->previous->name)
args.emplace_back("__VA_ARGS__");
argtok = argtok->next; // goto ')'
break;
}
if (argtok->op != ',')
args.emplace_back(argtok->str());
argtok = argtok->next;
}
if (!sameline(nametoken, argtok)) {
endToken = argtok ? argtok->previous : argtok;
valueToken = nullptr;
return false;
}
valueToken = argtok ? argtok->next : nullptr;
} else {
args.clear();
valueToken = nameTokDef->next;
}
if (!sameline(valueToken, nameTokDef))
valueToken = nullptr;
endToken = valueToken;
while (sameline(endToken, nameTokDef)) {
if (variadic && endToken->str() == "__VA_OPT__")
variadicOpt = true;
endToken = endToken->next;
}
if (variadicOpt) {
TokenList expandValue(files);
TokenList noExpandValue(files);
for (const Token *tok = valueToken; tok && tok != endToken;) {
if (tok->str() == "__VA_OPT__") {
if (!sameline(tok, tok->next) || tok->next->op != '(')
throw Error(tok->location, "In definition of '" + nameTokDef->str() + "': Missing opening parenthesis for __VA_OPT__");
tok = tok->next->next;
int par = 1;
while (tok && tok != endToken) {
if (tok->op == '(')
par++;
else if (tok->op == ')')
par--;
else if (tok->str() == "__VA_OPT__")
throw Error(tok->location, "In definition of '" + nameTokDef->str() + "': __VA_OPT__ cannot be nested");
if (par == 0) {
tok = tok->next;
break;
}
expandValue.push_back(new Token(*tok));
tok = tok->next;
}
if (par != 0) {
const Token *const lastTok = expandValue.back() ? expandValue.back() : valueToken->next;
throw Error(lastTok->location, "In definition of '" + nameTokDef->str() + "': Missing closing parenthesis for __VA_OPT__");
}
} else {
expandValue.push_back(new Token(*tok));
noExpandValue.push_back(new Token(*tok));
tok = tok->next;
}
}
optExpandValue = new TokenList(std::move(expandValue));
optNoExpandValue = new TokenList(std::move(noExpandValue));
}
return true;
}
unsigned int getArgNum(const TokenString &str) const {
unsigned int par = 0;
while (par < args.size()) {
if (str == args[par])
return par;
par++;
}
return ~0U;
}
std::vector<const Token *> getMacroParameters(const Token *nameTokInst, bool calledInDefine) const {
if (!nameTokInst->next || nameTokInst->next->op != '(' || !functionLike())
return {};
std::vector<const Token *> parametertokens;
parametertokens.emplace_back(nameTokInst->next);
unsigned int par = 0U;
for (const Token *tok = nameTokInst->next->next; calledInDefine ? sameline(tok, nameTokInst) : (tok != nullptr); tok = tok->next) {
if (tok->op == '(') {
++par;
}
else if (tok->op == ')') {
if (par == 0U) {
parametertokens.emplace_back(tok);
break;
}
--par;
} else if (par == 0U && tok->op == ',' && (!variadic || parametertokens.size() < args.size())) {
parametertokens.emplace_back(tok);
}
}
return parametertokens;
}
const Token *appendTokens(TokenList &tokens,
const Location &rawloc,
const Token * const lpar,
const MacroMap &macros,
const std::set<TokenString> &expandedmacros,
const std::vector<const Token*> &parametertokens) const {
if (!lpar || lpar->op != '(')
return nullptr;
unsigned int par = 0;
const Token *tok = lpar;
while (sameline(lpar, tok)) {
if (tok->op == '#' && sameline(tok,tok->next) && tok->next->op == '#' && sameline(tok,tok->next->next)) {
// A##B => AB
tok = expandHashHash(tokens, rawloc, tok, macros, expandedmacros, parametertokens, false);
} else if (tok->op == '#' && sameline(tok, tok->next) && tok->next->op != '#') {
tok = expandHash(tokens, rawloc, tok, expandedmacros, parametertokens);
} else {
if (!expandArg(tokens, tok, rawloc, macros, expandedmacros, parametertokens)) {
tokens.push_back(new Token(*tok));
if (tok->macro.empty() && (par > 0 || tok->str() != "("))
tokens.back()->macro = name();
}
if (tok->op == '(') {
++par;
}
else if (tok->op == ')') {
--par;
if (par == 0U)
break;
}
tok = tok->next;
}
}
for (Token *tok2 = tokens.front(); tok2; tok2 = tok2->next)
tok2->location = lpar->location;
return sameline(lpar,tok) ? tok : nullptr;
}
const Token * expand(TokenList & output, const Location &loc, const Token * const nameTokInst, const MacroMap &macros, std::set<TokenString> expandedmacros) const {
expandedmacros.insert(nameTokInst->str());
#ifdef SIMPLECPP_DEBUG_MACRO_EXPANSION
std::cout << " expand " << name() << " " << locstring(defineLocation()) << std::endl;
#endif
usageList.emplace_back(loc);
if (nameTokInst->str() == "__FILE__") {
output.push_back(new Token('\"'+output.file(loc)+'\"', loc));
return nameTokInst->next;
}
if (nameTokInst->str() == "__LINE__") {
output.push_back(new Token(toString(loc.line), loc));
return nameTokInst->next;
}
if (nameTokInst->str() == "__COUNTER__") {
output.push_back(new Token(toString(usageList.size()-1U), loc));
return nameTokInst->next;
}
const bool calledInDefine = (loc.fileIndex != nameTokInst->location.fileIndex ||
loc.line < nameTokInst->location.line);
std::vector<const Token*> parametertokens1(getMacroParameters(nameTokInst, calledInDefine));
if (functionLike()) {
// No arguments => not macro expansion
if (nameTokInst->next && nameTokInst->next->op != '(') {
output.push_back(new Token(nameTokInst->str(), loc));
return nameTokInst->next;
}
// Parse macro-call
if (variadic) {
if (parametertokens1.size() < args.size()) {
throw wrongNumberOfParameters(nameTokInst->location, name());
}
} else {
if (parametertokens1.size() != args.size() + (args.empty() ? 2U : 1U))
throw wrongNumberOfParameters(nameTokInst->location, name());
}
}
// If macro call uses __COUNTER__ then expand that first
TokenList tokensparams(files);
std::vector<const Token *> parametertokens2;
if (!parametertokens1.empty()) {
bool counter = false;
for (const Token *tok = parametertokens1[0]; tok != parametertokens1.back(); tok = tok->next) {
if (tok->str() == "__COUNTER__") {
counter = true;
break;
}
}
const MacroMap::const_iterator m = macros.find("__COUNTER__");
if (!counter || m == macros.end()) {
parametertokens2.swap(parametertokens1);
}
else {
const Macro &counterMacro = m->second;
unsigned int par = 0;
for (const Token *tok = parametertokens1[0]; tok && par < parametertokens1.size(); tok = tok->next) {
if (tok->str() == "__COUNTER__") {
tokensparams.push_back(new Token(toString(counterMacro.usageList.size()), tok->location));
counterMacro.usageList.emplace_back(tok->location);
} else {
tokensparams.push_back(new Token(*tok));
if (tok == parametertokens1[par]) {
parametertokens2.emplace_back(tokensparams.cback());
par++;
}
}
}
}
}
// NOLINTNEXTLINE(misc-const-correctness) - technically correct but used to access non-const data
Token * const output_end_1 = output.back();
const Token *valueToken2;
const Token *endToken2;
if (variadicOpt) {
if (parametertokens2.size() > args.size() && parametertokens2[args.size() - 1]->next->op != ')')
valueToken2 = optExpandValue->cfront();
else
valueToken2 = optNoExpandValue->cfront();
endToken2 = nullptr;
} else {
valueToken2 = valueToken;
endToken2 = endToken;
}
// expand
for (const Token *tok = valueToken2; tok != endToken2;) {
if (tok->op != '#') {
// A##B => AB
if (sameline(tok, tok->next) && tok->next && tok->next->op == '#' && tok->next->next && tok->next->next->op == '#') {
if (!sameline(tok, tok->next->next->next))
throw invalidHashHash::unexpectedNewline(tok->location, name());
if (variadic && tok->op == ',' && tok->next->next->next->str() == args.back()) {
Token *const comma = newMacroToken(tok->str(), loc, isReplaced(expandedmacros), tok);
output.push_back(comma);
tok = expandToken(output, loc, tok->next->next->next, macros, expandedmacros, parametertokens2);
if (output.back() == comma)
output.deleteToken(comma);
continue;
}
TokenList new_output(files);
if (!expandArg(new_output, tok, parametertokens2))
output.push_back(newMacroToken(tok->str(), loc, isReplaced(expandedmacros), tok));
else if (new_output.empty()) // placemarker token
output.push_back(newMacroToken("", loc, isReplaced(expandedmacros)));
else
for (const Token *tok2 = new_output.cfront(); tok2; tok2 = tok2->next)
output.push_back(newMacroToken(tok2->str(), loc, isReplaced(expandedmacros), tok2));
tok = tok->next;
} else {
tok = expandToken(output, loc, tok, macros, expandedmacros, parametertokens2);
}
continue;
}
int numberOfHash = 1;
const Token *hashToken = tok->next;
while (sameline(tok,hashToken) && hashToken->op == '#') {
hashToken = hashToken->next;
++numberOfHash;
}
if (numberOfHash == 4 && tok->next->location.col + 1 == tok->next->next->location.col) {
// # ## # => ##
output.push_back(newMacroToken("##", loc, isReplaced(expandedmacros)));
tok = hashToken;
continue;
}
if (numberOfHash >= 2 && tok->location.col + 1 < tok->next->location.col) {
output.push_back(new Token(*tok));
tok = tok->next;
continue;
}
tok = tok->next;
if (tok == endToken2) {
if (tok) {
output.push_back(new Token(*tok->previous));
}
else {
output.push_back(new Token(*nameTokInst));
output.back()->setstr("\"\"");
}
break;
}
if (tok->op == '#') {
// A##B => AB
tok = expandHashHash(output, loc, tok->previous, macros, expandedmacros, parametertokens2);
} else {
// #123 => "123"
tok = expandHash(output, loc, tok->previous, expandedmacros, parametertokens2);
}
}
if (!functionLike()) {
for (Token *tok = output_end_1 ? output_end_1->next : output.front(); tok; tok = tok->next) {
tok->macro = nameTokInst->str();
}
}
if (!parametertokens1.empty())
parametertokens1.swap(parametertokens2);
return functionLike() ? parametertokens2.back()->next : nameTokInst->next;
}
const Token *recursiveExpandToken(TokenList &output, TokenList &temp, const Location &loc, const Token *tok, const MacroMap &macros, const std::set<TokenString> &expandedmacros, const std::vector<const Token*> &parametertokens) const {
if (!temp.cback() || !temp.cback()->name || !tok->next || tok->next->op != '(') {
output.takeTokens(temp);
return tok->next;
}
if (!sameline(tok, tok->next)) {
output.takeTokens(temp);
return tok->next;
}
const MacroMap::const_iterator it = macros.find(temp.cback()->str());
if (it == macros.end() || expandedmacros.find(temp.cback()->str()) != expandedmacros.end()) {
output.takeTokens(temp);
return tok->next;
}
const Macro &calledMacro = it->second;
if (!calledMacro.functionLike()) {
output.takeTokens(temp);
return tok->next;
}
TokenList temp2(files);
temp2.push_back(new Token(temp.cback()->str(), tok->location));
const Token * const tok2 = appendTokens(temp2, loc, tok->next, macros, expandedmacros, parametertokens);
if (!tok2)
return tok->next;
output.takeTokens(temp);
output.deleteToken(output.back());
calledMacro.expand(output, loc, temp2.cfront(), macros, expandedmacros);
return tok2->next;
}
const Token *expandToken(TokenList &output, const Location &loc, const Token *tok, const MacroMap &macros, const std::set<TokenString> &expandedmacros, const std::vector<const Token*> &parametertokens) const {
// Not name..
if (!tok->name) {
output.push_back(newMacroToken(tok->str(), loc, true, tok));
return tok->next;
}
// Macro parameter..
{
TokenList temp(files);
if (expandArg(temp, tok, loc, macros, expandedmacros, parametertokens)) {
if (tok->str() == "__VA_ARGS__" && temp.empty() && output.cback() && output.cback()->str() == "," &&
tok->nextSkipComments() && tok->nextSkipComments()->str() == ")")
output.deleteToken(output.back());
return recursiveExpandToken(output, temp, loc, tok, macros, expandedmacros, parametertokens);
}
}
// Macro..
const MacroMap::const_iterator it = macros.find(tok->str());
if (it != macros.end() && expandedmacros.find(tok->str()) == expandedmacros.end()) {
std::set<std::string> expandedmacros2(expandedmacros);
expandedmacros2.insert(tok->str());
const Macro &calledMacro = it->second;
if (!calledMacro.functionLike()) {
TokenList temp(files);
calledMacro.expand(temp, loc, tok, macros, expandedmacros);
return recursiveExpandToken(output, temp, loc, tok, macros, expandedmacros2, parametertokens);
}
if (!sameline(tok, tok->next)) {
output.push_back(newMacroToken(tok->str(), loc, true, tok));
return tok->next;
}
TokenList tokens(files);
tokens.push_back(new Token(*tok));
const Token * tok2 = nullptr;
if (tok->next->op == '(') {
tok2 = appendTokens(tokens, loc, tok->next, macros, expandedmacros, parametertokens);
}
else if (expandArg(tokens, tok->next, loc, macros, expandedmacros, parametertokens)) {
tokens.front()->location = loc;
if (tokens.cfront()->next && tokens.cfront()->next->op == '(')
tok2 = tok->next;
}
if (!tok2) {
output.push_back(newMacroToken(tok->str(), loc, true, tok));
return tok->next;
}
TokenList temp(files);
calledMacro.expand(temp, loc, tokens.cfront(), macros, expandedmacros);
return recursiveExpandToken(output, temp, loc, tok2, macros, expandedmacros, parametertokens);
}
if (tok->str() == DEFINED) {
const Token * const tok2 = tok->next;
const Token * const tok3 = tok2 ? tok2->next : nullptr;
const Token * const tok4 = tok3 ? tok3->next : nullptr;
const Token *defToken = nullptr;
const Token *lastToken = nullptr;
if (sameline(tok, tok4) && tok2->op == '(' && tok3->name && tok4->op == ')') {
defToken = tok3;
lastToken = tok4;
} else if (sameline(tok,tok2) && tok2->name) {
defToken = lastToken = tok2;
}
if (defToken) {
std::string macroName = defToken->str();
if (defToken->next && defToken->next->op == '#' && defToken->next->next && defToken->next->next->op == '#' && defToken->next->next->next && defToken->next->next->next->name && sameline(defToken,defToken->next->next->next)) {
TokenList temp(files);
if (expandArg(temp, defToken, parametertokens))
macroName = temp.cback()->str();
if (expandArg(temp, defToken->next->next->next, parametertokens))
macroName += temp.cback() ? temp.cback()->str() : "";
else
macroName += defToken->next->next->next->str();
lastToken = defToken->next->next->next;
}
const bool def = (macros.find(macroName) != macros.end());
output.push_back(newMacroToken(def ? "1" : "0", loc, true));
return lastToken->next;
}
}
output.push_back(newMacroToken(tok->str(), loc, true, tok));
return tok->next;
}
bool expandArg(TokenList &output, const Token *tok, const std::vector<const Token*> &parametertokens) const {
if (!tok->name)
return false;
const unsigned int argnr = getArgNum(tok->str());
if (argnr >= args.size())
return false;
// empty variadic parameter
if (variadic && argnr + 1U >= parametertokens.size())
return true;
for (const Token *partok = parametertokens[argnr]->next; partok != parametertokens[argnr + 1U]; partok = partok->next)
output.push_back(new Token(*partok));
return true;
}
bool expandArg(TokenList &output, const Token *tok, const Location &loc, const MacroMap &macros, const std::set<TokenString> &expandedmacros, const std::vector<const Token*> &parametertokens) const {
if (!tok->name)
return false;
const unsigned int argnr = getArgNum(tok->str());
if (argnr >= args.size())
return false;
if (variadic && argnr + 1U >= parametertokens.size()) // empty variadic parameter
return true;
for (const Token *partok = parametertokens[argnr]->next; partok != parametertokens[argnr + 1U];) {
const MacroMap::const_iterator it = macros.find(partok->str());
if (it != macros.end() && !partok->isExpandedFrom(&it->second) && (partok->str() == name() || expandedmacros.find(partok->str()) == expandedmacros.end())) {
std::set<TokenString> expandedmacros2(expandedmacros); // temporary amnesia to allow reexpansion of currently expanding macros during argument evaluation
expandedmacros2.erase(name());
partok = it->second.expand(output, loc, partok, macros, std::move(expandedmacros2));
} else {
output.push_back(newMacroToken(partok->str(), loc, isReplaced(expandedmacros), partok));
output.back()->macro = partok->macro;
partok = partok->next;
}
}
if (tok->whitespaceahead && output.back())
output.back()->whitespaceahead = true;
return true;
}
/**
* Expand #X => "X"
* @param output destination tokenlist
* @param loc location for expanded token
* @param tok The # token
* @param expandedmacros set with expanded macros, with this macro
* @param parametertokens parameters given when expanding this macro
* @return token after the X
*/
const Token *expandHash(TokenList &output, const Location &loc, const Token *tok, const std::set<TokenString> &expandedmacros, const std::vector<const Token*> &parametertokens) const {
TokenList tokenListHash(files);
const MacroMap macros2; // temporarily bypass macro expansion
tok = expandToken(tokenListHash, loc, tok->next, macros2, expandedmacros, parametertokens);
std::ostringstream ostr;
ostr << '\"';
for (const Token *hashtok = tokenListHash.cfront(), *next; hashtok; hashtok = next) {
next = hashtok->next;
ostr << hashtok->str();
if (next && hashtok->whitespaceahead)
ostr << ' ';
}
ostr << '\"';
output.push_back(newMacroToken(escapeString(ostr.str()), loc, isReplaced(expandedmacros)));
return tok;
}
/**
* Expand A##B => AB
* The A should already be expanded. Call this when you reach the first # token
* @param output destination tokenlist
* @param loc location for expanded token
* @param tok first # token
* @param macros all macros
* @param expandedmacros set with expanded macros, with this macro
* @param parametertokens parameters given when expanding this macro
* @param expandResult expand ## result i.e. "AB"?
* @return token after B
*/
const Token *expandHashHash(TokenList &output, const Location &loc, const Token *tok, const MacroMap &macros, const std::set<TokenString> &expandedmacros, const std::vector<const Token*> &parametertokens, bool expandResult=true) const {
Token *A = output.back();
if (!A)
throw invalidHashHash(tok->location, name(), "Missing first argument");
if (!sameline(tok, tok->next) || !sameline(tok, tok->next->next))
throw invalidHashHash::unexpectedNewline(tok->location, name());
const bool canBeConcatenatedWithEqual = A->isOneOf("+-*/%&|^") || A->str() == "<<" || A->str() == ">>";
const bool canBeConcatenatedStringOrChar = isStringLiteral(A->str()) || isCharLiteral(A->str());
const bool unexpectedA = (!A->name && !A->number && !A->str().empty() && !canBeConcatenatedWithEqual && !canBeConcatenatedStringOrChar);
const Token * const B = tok->next->next;
if (!B->name && !B->number && B->op && !B->isOneOf("#="))
throw invalidHashHash::unexpectedToken(tok->location, name(), B);
if ((canBeConcatenatedWithEqual && B->op != '=') ||
(!canBeConcatenatedWithEqual && B->op == '='))
throw invalidHashHash::cannotCombine(tok->location, name(), A, B);
// Superficial check; more in-depth would in theory be possible _after_ expandArg
if (canBeConcatenatedStringOrChar && (B->number || !B->name))
throw invalidHashHash::cannotCombine(tok->location, name(), A, B);
TokenList tokensB(files);
const Token *nextTok = B->next;
if (canBeConcatenatedStringOrChar) {
if (unexpectedA)
throw invalidHashHash::unexpectedToken(tok->location, name(), A);
// It seems clearer to handle this case separately even though the code is similar-ish, but we don't want to merge here.
// TODO The question is whether the ## or varargs may still apply, and how to provoke?
if (expandArg(tokensB, B, parametertokens)) {
for (Token *b = tokensB.front(); b; b = b->next)
b->location = loc;
} else {
tokensB.push_back(new Token(*B));
tokensB.back()->location = loc;
}
output.takeTokens(tokensB);
} else {
std::string strAB;
const bool varargs = variadic && !args.empty() && B->str() == args[args.size()-1U];
if (expandArg(tokensB, B, parametertokens)) {
if (tokensB.empty()) {
strAB = A->str();
}
else if (varargs && A->op == ',') {
strAB = ",";
}
else if (varargs && unexpectedA) {
throw invalidHashHash::unexpectedToken(tok->location, name(), A);
}
else {
strAB = A->str() + tokensB.cfront()->str();
tokensB.deleteToken(tokensB.front());
}
} else {
if (unexpectedA)
throw invalidHashHash::unexpectedToken(tok->location, name(), A);
strAB = A->str() + B->str();
}
// producing universal character is undefined behavior
if (A->previous && A->previous->str() == "\\") {
if (strAB[0] == 'u' && strAB.size() == 5)
throw invalidHashHash::universalCharacterUB(tok->location, name(), A, strAB);
if (strAB[0] == 'U' && strAB.size() == 9)
throw invalidHashHash::universalCharacterUB(tok->location, name(), A, strAB);
}
if (varargs && tokensB.empty() && tok->previous->str() == ",") {
output.deleteToken(A);
}
else if (strAB != "," && macros.find(strAB) == macros.end()) {
A->setstr(strAB);
for (Token *b = tokensB.front(); b; b = b->next)
b->location = loc;
output.takeTokens(tokensB);
} else if (sameline(B, nextTok) && sameline(B, nextTok->next) && nextTok->op == '#' && nextTok->next->op == '#') {
TokenList output2(files);
output2.push_back(new Token(strAB, tok->location));
nextTok = expandHashHash(output2, loc, nextTok, macros, expandedmacros, parametertokens);
output.deleteToken(A);
output.takeTokens(output2);
} else {
output.deleteToken(A);
TokenList tokens(files);
tokens.push_back(new Token(strAB, tok->location));
// for function like macros, push the (...)
if (tokensB.empty() && sameline(B,B->next) && B->next->op=='(') {
const MacroMap::const_iterator it = macros.find(strAB);
if (it != macros.end() && expandedmacros.find(strAB) == expandedmacros.end() && it->second.functionLike()) {
const Token * const tok2 = appendTokens(tokens, loc, B->next, macros, expandedmacros, parametertokens);
if (tok2)
nextTok = tok2->next;
}
}
if (expandResult)
expandToken(output, loc, tokens.cfront(), macros, expandedmacros, parametertokens);
else
output.takeTokens(tokens);
for (Token *b = tokensB.front(); b; b = b->next)
b->location = loc;
output.takeTokens(tokensB);
}
}
return nextTok;
}
static bool isReplaced(const std::set<std::string> &expandedmacros) {
// return true if size > 1
auto it = expandedmacros.cbegin();
if (it == expandedmacros.cend())
return false;
++it;
return (it != expandedmacros.cend());
}
/** name token in definition */
const Token *nameTokDef;
/** arguments for macro */
std::vector<TokenString> args;
/** first token in replacement string */
const Token *valueToken;
/** token after replacement string */
const Token *endToken;
/** files */
std::vector<std::string> &files;
/** this is used for -D where the definition is not seen anywhere in code */
TokenList tokenListDefine;
/** usage of this macro */
mutable std::list<Location> usageList;
/** is macro variadic? */
bool variadic;
/** does the macro expansion have __VA_OPT__? */
bool variadicOpt;
/** Expansion value for varadic macros with __VA_OPT__ expanded and discarded respectively */
const TokenList *optExpandValue{};
const TokenList *optNoExpandValue{};
/** was the value of this macro actually defined in the code? */
bool valueDefinedInCode_;
};
}
namespace simplecpp {
#ifdef __CYGWIN__
static bool startsWith(const std::string &s, const std::string &p)
{
return (s.size() >= p.size()) && std::equal(p.begin(), p.end(), s.begin());
}
std::string convertCygwinToWindowsPath(const std::string &cygwinPath)
{
std::string windowsPath;
std::string::size_type pos = 0;
if (cygwinPath.size() >= 11 && startsWith(cygwinPath, "/cygdrive/")) {
const unsigned char driveLetter = cygwinPath[10];
if (std::isalpha(driveLetter)) {
if (cygwinPath.size() == 11) {
windowsPath = toupper(driveLetter);
windowsPath += ":\\"; // volume root directory
pos = 11;
} else if (cygwinPath[11] == '/') {
windowsPath = toupper(driveLetter);
windowsPath += ":";
pos = 11;
}
}
}
for (; pos < cygwinPath.size(); ++pos) {
unsigned char c = cygwinPath[pos];
if (c == '/')
c = '\\';
windowsPath += c;
}
return windowsPath;
}
#endif
bool isAbsolutePath(const std::string &path)
{
#ifdef SIMPLECPP_WINDOWS
// C:\\path\\file
// C:/path/file
if (path.length() >= 3 && std::isalpha(path[0]) && path[1] == ':' && (path[2] == '\\' || path[2] == '/'))
return true;
// \\host\path\file
// //host/path/file
if (path.length() >= 2 && (path[0] == '\\' || path[0] == '/') && (path[1] == '\\' || path[1] == '/'))
return true;
return false;
#else
return !path.empty() && path[0] == '/';
#endif
}
}
namespace simplecpp {
/**
* perform path simplifications for . and ..
*/
std::string simplifyPath(std::string path)
{
if (path.empty())
return path;
std::string::size_type pos;
// replace backslash separators
std::replace(path.begin(), path.end(), '\\', '/');
const bool unc(path.compare(0,2,"//") == 0);
// replace "//" with "/"
pos = 0;
while ((pos = path.find("//",pos)) != std::string::npos) {
path.erase(pos,1);
}
// remove "./"
pos = 0;
while ((pos = path.find("./",pos)) != std::string::npos) {
if (pos == 0 || path[pos - 1U] == '/')
path.erase(pos,2);
else
pos += 2;
}
// remove trailing dot if path ends with "/."
if (endsWith(path,"/."))
path.erase(path.size()-1);
// simplify ".."
pos = 1; // don't simplify ".." if path starts with that
while ((pos = path.find("/..", pos)) != std::string::npos) {
// not end of path, then string must be "/../"
if (pos + 3 < path.size() && path[pos + 3] != '/') {
++pos;
continue;
}
// get previous subpath
std::string::size_type pos1 = path.rfind('/', pos - 1U);
if (pos1 == std::string::npos) {
pos1 = 0;
} else {
pos1 += 1U;
}
const std::string previousSubPath = path.substr(pos1, pos - pos1);
if (previousSubPath == "..") {
// don't simplify
++pos;
} else {
// remove previous subpath and ".."
path.erase(pos1, pos - pos1 + 4);
if (path.empty())
path = ".";
// update pos
pos = (pos1 == 0) ? 1 : (pos1 - 1);
}
}
// Remove trailing '/'?
//if (path.size() > 1 && endsWith(path, "/"))
// path.erase(path.size()-1);
if (unc)
path = '/' + path;
return path;
}
}
/** Evaluate sizeof(type)
* @throws std::runtime_error thrown on missing arguments or invalid expression
*/
static void simplifySizeof(simplecpp::TokenList &expr, const std::map<std::string, std::size_t> &sizeOfType)
{
for (simplecpp::Token *tok = expr.front(); tok; tok = tok->next) {
if (tok->str() != "sizeof")
continue;
const simplecpp::Token *tok1 = tok->next;
if (!tok1) {
throw std::runtime_error("missing sizeof argument");
}
const simplecpp::Token *tok2 = tok1->next;
if (!tok2) {
throw std::runtime_error("missing sizeof argument");
}
if (tok1->op == '(') {
tok1 = tok1->next;
while (tok2->op != ')') {
tok2 = tok2->next;
if (!tok2) {
throw std::runtime_error("invalid sizeof expression");
}
}
}
std::string type;
for (const simplecpp::Token *typeToken = tok1; typeToken != tok2; typeToken = typeToken->next) {
if ((typeToken->str() == "unsigned" || typeToken->str() == "signed") && typeToken->next->name)
continue;
if (typeToken->str() == "*" && type.find('*') != std::string::npos)
continue;
if (!type.empty())
type += ' ';
type += typeToken->str();
}
const std::map<std::string, std::size_t>::const_iterator it = sizeOfType.find(type);
if (it != sizeOfType.end())
tok->setstr(toString(it->second));
else
continue;
tok2 = tok2->next;
while (tok->next != tok2)
expr.deleteToken(tok->next);
}
}
static bool isCpp17OrLater(const simplecpp::DUI &dui)
{
const std::string std_ver = simplecpp::getCppStdString(dui.std);
return std_ver.empty() || (std_ver >= "201703L");
}
static bool isGnu(const simplecpp::DUI &dui)
{
return dui.std.rfind("gnu", 0) != std::string::npos;
}
static std::string dirPath(const std::string& path, bool withTrailingSlash=true)
{
const std::size_t lastSlash = path.find_last_of("\\/");
if (lastSlash == std::string::npos) {
return "";
}
return path.substr(0, lastSlash + (withTrailingSlash ? 1U : 0U));
}
static std::string openHeader(std::ifstream &f, const simplecpp::DUI &dui, const std::string &sourcefile, const std::string &header, bool systemheader);
/** Evaluate __has_include(include)
* @throws std::runtime_error thrown on missing arguments or invalid expression
*/
static void simplifyHasInclude(simplecpp::TokenList &expr, const simplecpp::DUI &dui)
{
if (!isCpp17OrLater(dui) && !isGnu(dui))
return;
for (simplecpp::Token *tok = expr.front(); tok; tok = tok->next) {
if (tok->str() != HAS_INCLUDE)
continue;
const simplecpp::Token *tok1 = tok->next;
if (!tok1) {
throw std::runtime_error("missing __has_include argument");
}
const simplecpp::Token *tok2 = tok1->next;
if (!tok2) {
throw std::runtime_error("missing __has_include argument");
}
if (tok1->op == '(') {
tok1 = tok1->next;
while (tok2->op != ')') {
tok2 = tok2->next;
if (!tok2) {
throw std::runtime_error("invalid __has_include expression");
}
}
}
const std::string &sourcefile = expr.file(tok->location);
const bool systemheader = (tok1 && tok1->op == '<');
std::string header;
if (systemheader) {
const simplecpp::Token *tok3 = tok1->next;
if (!tok3) {
throw std::runtime_error("missing __has_include closing angular bracket");
}
while (tok3->op != '>') {
tok3 = tok3->next;
if (!tok3) {
throw std::runtime_error("invalid __has_include expression");
}
}
for (const simplecpp::Token *headerToken = tok1->next; headerToken != tok3; headerToken = headerToken->next)
header += headerToken->str();
} else {
header = tok1->str().substr(1U, tok1->str().size() - 2U);
}
std::ifstream f;
const std::string header2 = openHeader(f,dui,sourcefile,header,systemheader);
tok->setstr(header2.empty() ? "0" : "1");
tok2 = tok2->next;
while (tok->next != tok2)
expr.deleteToken(tok->next);
}
}
/** Evaluate name
* @throws std::runtime_error thrown on undefined function-like macro
*/
static void simplifyName(simplecpp::TokenList &expr)
{
for (simplecpp::Token *tok = expr.front(); tok; tok = tok->next) {
if (tok->name) {
static const std::set<std::string> altop{"and","or","bitand","bitor","compl","not","not_eq","xor"};
if (altop.find(tok->str()) != altop.end()) {
bool alt;
if (tok->str() == "not" || tok->str() == "compl") {
alt = isAlternativeUnaryOp(tok,tok->str());
} else {
alt = isAlternativeBinaryOp(tok,tok->str());
}
if (alt)
continue;
}
if (tok->next && tok->next->str() == "(")
throw std::runtime_error("undefined function-like macro invocation: " + tok->str() + "( ... )");
tok->setstr("0");
}
}
}
/*
* Reads at least minlen and at most maxlen digits (inc. prefix) in base base
* from s starting at position pos and converts them to a
* unsigned long long value, updating pos to point to the first
* unused element of s.
* Returns ULLONG_MAX if the result is not representable and
* @throws std::runtime_error thrown if the above requirements were not possible to satisfy.
*/
static unsigned long long stringToULLbounded(
const std::string& s,
std::size_t& pos,
int base = 0,
std::ptrdiff_t minlen = 1,
std::size_t maxlen = std::string::npos
)
{
const std::string sub = s.substr(pos, maxlen);
const char * const start = sub.c_str();
char* end;
const unsigned long long value = std::strtoull(start, &end, base);
pos += end - start;
if (end - start < minlen)
throw std::runtime_error("expected digit");
return value;
}
long long simplecpp::characterLiteralToLL(const std::string& str)
{
// default is wide/utf32
bool narrow = false;
bool utf8 = false;
bool utf16 = false;
std::size_t pos;
if (!str.empty() && str[0] == '\'') {
narrow = true;
pos = 1;
} else if (str.size() >= 2 && str[0] == 'u' && str[1] == '\'') {
utf16 = true;
pos = 2;
} else if (str.size() >= 3 && str[0] == 'u' && str[1] == '8' && str[2] == '\'') {
utf8 = true;
pos = 3;
} else if (str.size() >= 2 && (str[0] == 'L' || str[0] == 'U') && str[1] == '\'') {
pos = 2;
} else {
throw std::runtime_error("expected a character literal");
}
unsigned long long multivalue = 0;
std::size_t nbytes = 0;
while (pos + 1 < str.size()) {
if (str[pos] == '\'' || str[pos] == '\n')
throw std::runtime_error("raw single quotes and newlines not allowed in character literals");
if (nbytes >= 1 && !narrow)
throw std::runtime_error("multiple characters only supported in narrow character literals");
unsigned long long value;
if (str[pos] == '\\') {
pos++;
const char escape = str[pos++];
if (pos >= str.size())
throw std::runtime_error("unexpected end of character literal");
switch (escape) {
// obscure GCC extensions
case '%':
case '(':
case '[':
case '{':
// standard escape sequences
case '\'':
case '"':
case '?':
case '\\':
value = static_cast<unsigned char>(escape);
break;
case 'a':
value = static_cast<unsigned char>('\a');
break;
case 'b':
value = static_cast<unsigned char>('\b');
break;
case 'f':
value = static_cast<unsigned char>('\f');
break;
case 'n':
value = static_cast<unsigned char>('\n');
break;
case 'r':
value = static_cast<unsigned char>('\r');
break;
case 't':
value = static_cast<unsigned char>('\t');
break;
case 'v':
value = static_cast<unsigned char>('\v');
break;
// GCC extension for ESC character
case 'e':
case 'E':
value = static_cast<unsigned char>('\x1b');
break;
case '0':
case '1':
case '2':
case '3':
case '4':
case '5':
case '6':
case '7':
// octal escape sequences consist of 1 to 3 digits
value = stringToULLbounded(str, --pos, 8, 1, 3);
break;
case 'x':
// hexadecimal escape sequences consist of at least 1 digit
value = stringToULLbounded(str, pos, 16);
break;
case 'u':
case 'U': {
// universal character names have exactly 4 or 8 digits
const std::size_t ndigits = (escape == 'u' ? 4 : 8);
value = stringToULLbounded(str, pos, 16, ndigits, ndigits);
// UTF-8 encodes code points above 0x7f in multiple code units
// code points above 0x10ffff are not allowed
if (((narrow || utf8) && value > 0x7f) || (utf16 && value > 0xffff) || value > 0x10ffff)
throw std::runtime_error("code point too large");
if (value >= 0xd800 && value <= 0xdfff)
throw std::runtime_error("surrogate code points not allowed in universal character names");
break;
}
default:
throw std::runtime_error("invalid escape sequence");
}
} else {
value = static_cast<unsigned char>(str[pos++]);
if (!narrow && value >= 0x80) {
// Assuming this is a UTF-8 encoded code point.
// This decoder may not completely validate the input.
// Noncharacters are neither rejected nor replaced.
int additional_bytes;
if (value >= 0xf5) // higher values would result in code points above 0x10ffff
throw std::runtime_error("assumed UTF-8 encoded source, but sequence is invalid");
if (value >= 0xf0)
additional_bytes = 3;
else if (value >= 0xe0)
additional_bytes = 2;
else if (value >= 0xc2) // 0xc0 and 0xc1 are always overlong 2-bytes encodings
additional_bytes = 1;
else
throw std::runtime_error("assumed UTF-8 encoded source, but sequence is invalid");
value &= (1 << (6 - additional_bytes)) - 1;
while (additional_bytes--) {
if (pos + 1 >= str.size())
throw std::runtime_error("assumed UTF-8 encoded source, but character literal ends unexpectedly");
const unsigned char c = str[pos++];
if (((c >> 6) != 2) // ensure c has form 0xb10xxxxxx
|| (!value && additional_bytes == 1 && c < 0xa0) // overlong 3-bytes encoding
|| (!value && additional_bytes == 2 && c < 0x90)) // overlong 4-bytes encoding
throw std::runtime_error("assumed UTF-8 encoded source, but sequence is invalid");
value = (value << 6) | (c & ((1 << 7) - 1));
}
if (value >= 0xd800 && value <= 0xdfff)
throw std::runtime_error("assumed UTF-8 encoded source, but sequence is invalid");
if ((utf8 && value > 0x7f) || (utf16 && value > 0xffff) || value > 0x10ffff)
throw std::runtime_error("code point too large");
}
}
if (((narrow || utf8) && value > std::numeric_limits<unsigned char>::max()) || (utf16 && value >> 16) || value >> 32)
throw std::runtime_error("numeric escape sequence too large");
multivalue <<= CHAR_BIT;
multivalue |= value;
nbytes++;
}
if (pos + 1 != str.size() || str[pos] != '\'')
throw std::runtime_error("missing closing quote in character literal");
if (!nbytes)
throw std::runtime_error("empty character literal");
// ordinary narrow character literal's value is determined by (possibly signed) char
if (narrow && nbytes == 1)
return static_cast<char>(multivalue);
// while multi-character literal's value is determined by (signed) int
if (narrow)
return static_cast<int>(multivalue);
// All other cases are unsigned. Since long long is at least 64bit wide,
// while the literals at most 32bit wide, the conversion preserves all values.
return multivalue;
}
/**
* @throws std::runtime_error thrown on invalid literal
*/
static void simplifyNumbers(simplecpp::TokenList &expr)
{
for (simplecpp::Token *tok = expr.front(); tok; tok = tok->next) {
if (tok->str().size() == 1U)
continue;
if (tok->str().compare(0,2,"0x") == 0)
tok->setstr(toString(stringToULL(tok->str())));
else if (!tok->number && tok->str().find('\'') != std::string::npos)
tok->setstr(toString(simplecpp::characterLiteralToLL(tok->str())));
}
}
static void simplifyComments(simplecpp::TokenList &expr)
{
for (simplecpp::Token *tok = expr.front(); tok;) {
simplecpp::Token * const d = tok;
tok = tok->next;
if (d->comment)
expr.deleteToken(d);
}
}
/**
* @throws std::runtime_error thrown on invalid literals, missing sizeof arguments or invalid expressions,
* missing __has_include() arguments or expressions, undefined function-like macros, invalid number literals
* @throws std::overflow_error thrown on overflow or division by zero
*/
static long long evaluate(simplecpp::TokenList &expr, const simplecpp::DUI &dui, const std::map<std::string, std::size_t> &sizeOfType)
{
simplifyComments(expr);
simplifySizeof(expr, sizeOfType);
simplifyHasInclude(expr, dui);
simplifyName(expr);
simplifyNumbers(expr);
expr.constFold();
// TODO: handle invalid expressions
return expr.cfront() && expr.cfront() == expr.cback() && expr.cfront()->number ? stringToLL(expr.cfront()->str()) : 0LL;
}
static const simplecpp::Token *gotoNextLine(const simplecpp::Token *tok)
{
const unsigned int line = tok->location.line;
const unsigned int file = tok->location.fileIndex;
while (tok && tok->location.line == line && tok->location.fileIndex == file)
tok = tok->next;
return tok;
}
#ifdef SIMPLECPP_WINDOWS
class NonExistingFilesCache {
public:
NonExistingFilesCache() {}
bool contains(const std::string& path) {
std::lock_guard<std::mutex> lock(m_mutex);
return (m_pathSet.find(path) != m_pathSet.end());
}
void add(const std::string& path) {
std::lock_guard<std::mutex> lock(m_mutex);
m_pathSet.insert(path);
}
void clear() {
std::lock_guard<std::mutex> lock(m_mutex);
m_pathSet.clear();
}
private:
std::set<std::string> m_pathSet;
std::mutex m_mutex;
};
static NonExistingFilesCache nonExistingFilesCache;
#endif
static std::string openHeaderDirect(std::ifstream &f, const std::string &path)
{
#ifdef SIMPLECPP_WINDOWS
if (nonExistingFilesCache.contains(path))
return ""; // file is known not to exist, skip expensive file open call
#endif
f.open(path.c_str());
if (f.is_open())
return path;
#ifdef SIMPLECPP_WINDOWS
nonExistingFilesCache.add(path);
#endif
return "";
}
static std::string openHeader(std::ifstream &f, const simplecpp::DUI &dui, const std::string &sourcefile, const std::string &header, bool systemheader)
{
if (simplecpp::isAbsolutePath(header))
return openHeaderDirect(f, simplecpp::simplifyPath(header));
// prefer first to search the header relatively to source file if found, when not a system header
if (!systemheader) {
std::string path = openHeaderDirect(f, simplecpp::simplifyPath(dirPath(sourcefile) + header));
if (!path.empty()) {
return path;
}
}
// search the header on the include paths (provided by the flags "-I...")
for (const auto &includePath : dui.includePaths) {
std::string path = openHeaderDirect(f, simplecpp::simplifyPath(includePath + "/" + header));
if (!path.empty())
return path;
}
return "";
}
std::pair<simplecpp::FileData *, bool> simplecpp::FileDataCache::tryload(FileDataCache::name_map_type::iterator &name_it, const simplecpp::DUI &dui, std::vector<std::string> &filenames, simplecpp::OutputList *outputList)
{
const std::string &path = name_it->first;
FileID fileId;
if (!getFileId(path, fileId))
return {nullptr, false};
const auto id_it = mIdMap.find(fileId);
if (id_it != mIdMap.end()) {
name_it->second = id_it->second;
return {id_it->second, false};
}
auto *const data = new FileData {path, TokenList(path, filenames, outputList)};
if (dui.removeComments)
data->tokens.removeComments();
name_it->second = data;
mIdMap.emplace(fileId, data);
mData.emplace_back(data);
return {data, true};
}
std::pair<simplecpp::FileData *, bool> simplecpp::FileDataCache::get(const std::string &sourcefile, const std::string &header, const simplecpp::DUI &dui, bool systemheader, std::vector<std::string> &filenames, simplecpp::OutputList *outputList)
{
if (isAbsolutePath(header)) {
auto ins = mNameMap.emplace(simplecpp::simplifyPath(header), nullptr);
if (ins.second) {
const auto ret = tryload(ins.first, dui, filenames, outputList);
if (ret.first != nullptr) {
return ret;
}
} else {
return {ins.first->second, false};
}
return {nullptr, false};
}
if (!systemheader) {
auto ins = mNameMap.emplace(simplecpp::simplifyPath(dirPath(sourcefile) + header), nullptr);
if (ins.second) {
const auto ret = tryload(ins.first, dui, filenames, outputList);
if (ret.first != nullptr) {
return ret;
}
} else if (ins.first->second != nullptr) {
return {ins.first->second, false};
}
}
for (const auto &includePath : dui.includePaths) {
auto ins = mNameMap.emplace(simplecpp::simplifyPath(includePath + "/" + header), nullptr);
if (ins.second) {
const auto ret = tryload(ins.first, dui, filenames, outputList);
if (ret.first != nullptr) {
return ret;
}
} else if (ins.first->second != nullptr) {
return {ins.first->second, false};
}
}
return {nullptr, false};
}
bool simplecpp::FileDataCache::getFileId(const std::string &path, FileID &id)
{
#ifdef _WIN32
HANDLE hFile = CreateFileA(path.c_str(), 0, FILE_SHARE_READ | FILE_SHARE_WRITE, nullptr, OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, nullptr);
if (hFile == INVALID_HANDLE_VALUE)
return false;
BOOL ret = GetFileInformationByHandleEx(hFile, FileIdInfo, &id.fileIdInfo, sizeof(id.fileIdInfo));
if (!ret) {
const DWORD err = GetLastError();
if (err == ERROR_INVALID_PARAMETER || // encountered when using a non-NTFS filesystem e.g. exFAT
err == ERROR_NOT_SUPPORTED) // encountered on Windows Server Core (used as a Docker container)
{
BY_HANDLE_FILE_INFORMATION fileInfo;
ret = GetFileInformationByHandle(hFile, &fileInfo);
if (ret) {
id.fileIdInfo.VolumeSerialNumber = static_cast<std::uint64_t>(fileInfo.dwVolumeSerialNumber);
id.fileIdInfo.FileId.IdentifierHi = static_cast<std::uint64_t>(fileInfo.nFileIndexHigh);
id.fileIdInfo.FileId.IdentifierLo = static_cast<std::uint64_t>(fileInfo.nFileIndexLow);
}
}
}
CloseHandle(hFile);
return ret == TRUE;
#else
struct stat statbuf;
if (stat(path.c_str(), &statbuf) != 0)
return false;
id.dev = statbuf.st_dev;
id.ino = statbuf.st_ino;
return true;
#endif
}
simplecpp::FileDataCache simplecpp::load(const simplecpp::TokenList &rawtokens, std::vector<std::string> &filenames, const simplecpp::DUI &dui, simplecpp::OutputList *outputList, FileDataCache cache)
{
#ifdef SIMPLECPP_WINDOWS
if (dui.clearIncludeCache)
nonExistingFilesCache.clear();
#endif
std::list<const Token *> filelist;
// -include files
for (auto it = dui.includes.cbegin(); it != dui.includes.cend(); ++it) {
const std::string &filename = *it;
const auto loadResult = cache.get("", filename, dui, false, filenames, outputList);
const bool loaded = loadResult.second;
FileData *const filedata = loadResult.first;
if (filedata == nullptr) {
if (outputList) {
simplecpp::Output err{
simplecpp::Output::EXPLICIT_INCLUDE_NOT_FOUND,
{},
"Can not open include file '" + filename + "' that is explicitly included."
};
outputList->emplace_back(std::move(err));
}
continue;
}
if (!loaded)
continue;
if (!filedata->tokens.front())
continue;
if (dui.removeComments)
filedata->tokens.removeComments();
filelist.emplace_back(filedata->tokens.front());
}
for (const Token *rawtok = rawtokens.cfront(); rawtok || !filelist.empty(); rawtok = rawtok ? rawtok->next : nullptr) {
if (rawtok == nullptr) {
rawtok = filelist.back();
filelist.pop_back();
}
if (rawtok->op != '#' || sameline(rawtok->previousSkipComments(), rawtok))
continue;
rawtok = rawtok->nextSkipComments();
if (!rawtok || rawtok->str() != INCLUDE)
continue;
const std::string &sourcefile = rawtokens.file(rawtok->location);
const Token * const htok = rawtok->nextSkipComments();
if (!sameline(rawtok, htok))
continue;
const bool systemheader = (htok->str()[0] == '<');
const std::string header(htok->str().substr(1U, htok->str().size() - 2U));
const auto loadResult = cache.get(sourcefile, header, dui, systemheader, filenames, outputList);
const bool loaded = loadResult.second;
if (!loaded)
continue;
FileData *const filedata = loadResult.first;
if (!filedata->tokens.front())
continue;
if (dui.removeComments)
filedata->tokens.removeComments();
filelist.emplace_back(filedata->tokens.front());
}
return cache;
}
static bool preprocessToken(simplecpp::TokenList &output, const simplecpp::Token *&tok1, simplecpp::MacroMap &macros, std::vector<std::string> &files, simplecpp::OutputList *outputList)
{
const simplecpp::Token * const tok = tok1;
const simplecpp::MacroMap::const_iterator it = tok->name ? macros.find(tok->str()) : macros.end();
if (it != macros.end()) {
simplecpp::TokenList value(files);
try {
tok1 = it->second.expand(value, tok, macros, files);
} catch (const simplecpp::Macro::Error &err) {
if (outputList) {
simplecpp::Output out{
simplecpp::Output::SYNTAX_ERROR,
err.location,
"failed to expand \'" + tok->str() + "\', " + err.what
};
outputList->emplace_back(std::move(out));
}
return false;
}
output.takeTokens(value);
} else {
if (!tok->comment)
output.push_back(new simplecpp::Token(*tok));
tok1 = tok->next;
}
return true;
}
static void getLocaltime(struct tm &ltime)
{
time_t t;
time(&t);
#ifndef _WIN32
// NOLINTNEXTLINE(misc-include-cleaner) - false positive
localtime_r(&t, &ltime);
#else
localtime_s(&ltime, &t);
#endif
}
static std::string getDateDefine(const struct tm *timep)
{
char buf[] = "??? ?? ????";
strftime(buf, sizeof(buf), "%b %d %Y", timep);
return std::string("\"").append(buf).append("\"");
}
static std::string getTimeDefine(const struct tm *timep)
{
char buf[] = "??:??:??";
strftime(buf, sizeof(buf), "%H:%M:%S", timep);
return std::string("\"").append(buf).append("\"");
}
void simplecpp::preprocess(simplecpp::TokenList &output, const simplecpp::TokenList &rawtokens, std::vector<std::string> &files, simplecpp::FileDataCache &cache, const simplecpp::DUI &dui, simplecpp::OutputList *outputList, std::list<simplecpp::MacroUsage> *macroUsage, std::list<simplecpp::IfCond> *ifCond)
{
#ifdef SIMPLECPP_WINDOWS
if (dui.clearIncludeCache)
nonExistingFilesCache.clear();
#endif
std::map<std::string, std::size_t> sizeOfType(rawtokens.sizeOfType);
sizeOfType.insert(std::make_pair("char", sizeof(char)));
sizeOfType.insert(std::make_pair("short", sizeof(short)));
sizeOfType.insert(std::make_pair("short int", sizeOfType["short"]));
sizeOfType.insert(std::make_pair("int", sizeof(int)));
sizeOfType.insert(std::make_pair("long", sizeof(long)));
sizeOfType.insert(std::make_pair("long int", sizeOfType["long"]));
sizeOfType.insert(std::make_pair("long long", sizeof(long long)));
sizeOfType.insert(std::make_pair("float", sizeof(float)));
sizeOfType.insert(std::make_pair("double", sizeof(double)));
sizeOfType.insert(std::make_pair("long double", sizeof(long double)));
sizeOfType.insert(std::make_pair("char *", sizeof(char *)));
sizeOfType.insert(std::make_pair("short *", sizeof(short *)));
sizeOfType.insert(std::make_pair("short int *", sizeOfType["short *"]));
sizeOfType.insert(std::make_pair("int *", sizeof(int *)));
sizeOfType.insert(std::make_pair("long *", sizeof(long *)));
sizeOfType.insert(std::make_pair("long int *", sizeOfType["long *"]));
sizeOfType.insert(std::make_pair("long long *", sizeof(long long *)));
sizeOfType.insert(std::make_pair("float *", sizeof(float *)));
sizeOfType.insert(std::make_pair("double *", sizeof(double *)));
sizeOfType.insert(std::make_pair("long double *", sizeof(long double *)));
// use a dummy vector for the macros because as this is not part of the file and would add an empty entry - e.g. /usr/include/poll.h
std::vector<std::string> dummy;
const bool hasInclude = isCpp17OrLater(dui) || isGnu(dui);
MacroMap macros;
bool strictAnsiDefined = false;
for (auto it = dui.defines.cbegin(); it != dui.defines.cend(); ++it) {
const std::string &macrostr = *it;
const std::string::size_type eq = macrostr.find('=');
const std::string::size_type par = macrostr.find('(');
const std::string macroname = macrostr.substr(0, std::min(eq,par));
if (macroname == "__STRICT_ANSI__")
strictAnsiDefined = true;
if (dui.undefined.find(macroname) != dui.undefined.end())
continue;
const std::string lhs(macrostr.substr(0,eq));
const std::string rhs(eq==std::string::npos ? std::string("1") : macrostr.substr(eq+1));
try {
const Macro macro(lhs, rhs, dummy);
macros.insert(std::pair<TokenString,Macro>(macro.name(), macro));
} catch (const std::runtime_error& e) {
if (outputList) {
simplecpp::Output err{
Output::DUI_ERROR,
{},
e.what()
};
outputList->emplace_back(std::move(err));
}
output.clear();
return;
}
}
const bool strictAnsiUndefined = dui.undefined.find("__STRICT_ANSI__") != dui.undefined.cend();
if (!isGnu(dui) && !strictAnsiDefined && !strictAnsiUndefined)
macros.insert(std::pair<TokenString, Macro>("__STRICT_ANSI__", Macro("__STRICT_ANSI__", "1", dummy)));
macros.insert(std::make_pair("__FILE__", Macro("__FILE__", "__FILE__", dummy)));
macros.insert(std::make_pair("__LINE__", Macro("__LINE__", "__LINE__", dummy)));
macros.insert(std::make_pair("__COUNTER__", Macro("__COUNTER__", "__COUNTER__", dummy)));
struct tm ltime {};
getLocaltime(ltime);
macros.insert(std::make_pair("__DATE__", Macro("__DATE__", getDateDefine(&ltime), dummy)));
macros.insert(std::make_pair("__TIME__", Macro("__TIME__", getTimeDefine(&ltime), dummy)));
if (!dui.std.empty()) {
const cstd_t c_std = simplecpp::getCStd(dui.std);
if (c_std != CUnknown) {
const std::string std_def = simplecpp::getCStdString(c_std);
if (!std_def.empty())
macros.insert(std::make_pair("__STDC_VERSION__", Macro("__STDC_VERSION__", std_def, dummy)));
} else {
const cppstd_t cpp_std = simplecpp::getCppStd(dui.std);
if (cpp_std == CPPUnknown) {
if (outputList) {
simplecpp::Output err{
Output::DUI_ERROR,
{},
"unknown standard specified: '" + dui.std + "'"
};
outputList->emplace_back(std::move(err));
}
output.clear();
return;
}
const std::string std_def = simplecpp::getCppStdString(cpp_std);
if (!std_def.empty())
macros.insert(std::make_pair("__cplusplus", Macro("__cplusplus", std_def, dummy)));
}
}
// True => code in current #if block should be kept
// ElseIsTrue => code in current #if block should be dropped. the code in the #else should be kept.
// AlwaysFalse => drop all code in #if and #else
enum IfState : std::uint8_t { True, ElseIsTrue, AlwaysFalse };
std::stack<int> ifstates;
std::stack<const Token *> iftokens;
ifstates.push(True);
std::stack<const Token *> includetokenstack;
std::set<std::string> pragmaOnce;
includetokenstack.push(rawtokens.cfront());
for (auto it = dui.includes.cbegin(); it != dui.includes.cend(); ++it) {
const FileData *const filedata = cache.get("", *it, dui, false, files, outputList).first;
if (filedata != nullptr && filedata->tokens.cfront() != nullptr)
includetokenstack.push(filedata->tokens.cfront());
}
std::map<std::string, std::list<Location>> maybeUsedMacros;
for (const Token *rawtok = nullptr; rawtok || !includetokenstack.empty();) {
if (rawtok == nullptr) {
rawtok = includetokenstack.top();
includetokenstack.pop();
continue;
}
if (rawtok->op == '#' && !sameline(rawtok->previousSkipComments(), rawtok)) {
if (!sameline(rawtok, rawtok->next)) {
rawtok = rawtok->next;
continue;
}
rawtok = rawtok->next;
if (!rawtok->name) {
rawtok = gotoNextLine(rawtok);
continue;
}
if (ifstates.size() <= 1U && (rawtok->str() == ELIF || rawtok->str() == ELSE || rawtok->str() == ENDIF)) {
if (outputList) {
simplecpp::Output err{
Output::SYNTAX_ERROR,
rawtok->location,
"#" + rawtok->str() + " without #if"
};
outputList->emplace_back(std::move(err));
}
output.clear();
return;
}
if (ifstates.top() == True && (rawtok->str() == ERROR || rawtok->str() == WARNING)) {
if (outputList) {
std::string msg;
for (const Token *tok = rawtok->next; tok && sameline(rawtok,tok); tok = tok->next) {
if (!msg.empty() && isNameChar(tok->str()[0]))
msg += ' ';
msg += tok->str();
}
msg = '#' + rawtok->str() + ' ' + msg;
simplecpp::Output err{
rawtok->str() == ERROR ? Output::ERROR : Output::WARNING,
rawtok->location,
std::move(msg)
};
outputList->emplace_back(std::move(err));
}
if (rawtok->str() == ERROR) {
output.clear();
return;
}
}
if (rawtok->str() == DEFINE) {
if (ifstates.top() != True)
continue;
try {
const Macro &macro = Macro(rawtok->previous, files);
if (dui.undefined.find(macro.name()) == dui.undefined.end()) {
const MacroMap::iterator it = macros.find(macro.name());
if (it == macros.end())
macros.insert(std::pair<TokenString, Macro>(macro.name(), macro));
else
it->second = macro;
}
} catch (const std::runtime_error &) {
if (outputList) {
simplecpp::Output err{
Output::SYNTAX_ERROR,
rawtok->location,
"Failed to parse #define"
};
outputList->emplace_back(std::move(err));
}
output.clear();
return;
} catch (const simplecpp::Macro::Error &err) {
if (outputList) {
simplecpp::Output out{
simplecpp::Output::SYNTAX_ERROR,
err.location,
"Failed to parse #define, " + err.what
};
outputList->emplace_back(std::move(out));
}
output.clear();
return;
}
} else if (ifstates.top() == True && rawtok->str() == INCLUDE) {
TokenList inc1(files);
for (const Token *inctok = rawtok->next; sameline(rawtok,inctok); inctok = inctok->next) {
if (!inctok->comment)
inc1.push_back(new Token(*inctok));
}
TokenList inc2(files);
if (!inc1.empty() && inc1.cfront()->name) {
const Token *inctok = inc1.cfront();
if (!preprocessToken(inc2, inctok, macros, files, outputList)) {
output.clear();
return;
}
} else {
inc2.takeTokens(inc1);
}
if (!inc1.empty() && !inc2.empty() && inc2.cfront()->op == '<' && inc2.cback()->op == '>') {
TokenString hdr;
// TODO: Sometimes spaces must be added in the string
// Somehow preprocessToken etc must be told that the location should be source location not destination location
for (const Token *tok = inc2.cfront(); tok; tok = tok->next) {
hdr += tok->str();
}
inc2.clear();
inc2.push_back(new Token(hdr, inc1.cfront()->location));
inc2.front()->op = '<';
}
if (inc2.empty() || inc2.cfront()->str().size() <= 2U) {
if (outputList) {
simplecpp::Output err{
Output::SYNTAX_ERROR,
rawtok->location,
"No header in #include"
};
outputList->emplace_back(std::move(err));
}
output.clear();
return;
}
const Token * const inctok = inc2.cfront();
const bool systemheader = (inctok->str()[0] == '<');
const std::string header(inctok->str().substr(1U, inctok->str().size() - 2U));
const FileData *const filedata = cache.get(rawtokens.file(rawtok->location), header, dui, systemheader, files, outputList).first;
if (filedata == nullptr) {
if (outputList) {
simplecpp::Output out{
simplecpp::Output::MISSING_HEADER,
rawtok->location,
"Header not found: " + inctok->str()
};
outputList->emplace_back(std::move(out));
}
} else if (includetokenstack.size() >= 400) {
if (outputList) {
simplecpp::Output out{
simplecpp::Output::INCLUDE_NESTED_TOO_DEEPLY,
rawtok->location,
"#include nested too deeply"
};
outputList->emplace_back(std::move(out));
}
} else if (pragmaOnce.find(filedata->filename) == pragmaOnce.end()) {
includetokenstack.push(gotoNextLine(rawtok));
rawtok = filedata->tokens.cfront();
continue;
}
} else if (rawtok->str() == IF || rawtok->str() == IFDEF || rawtok->str() == IFNDEF || rawtok->str() == ELIF) {
if (!sameline(rawtok,rawtok->next)) {
if (outputList) {
simplecpp::Output out{
simplecpp::Output::SYNTAX_ERROR,
rawtok->location,
"Syntax error in #" + rawtok->str()
};
outputList->emplace_back(std::move(out));
}
output.clear();
return;
}
bool conditionIsTrue;
if (ifstates.top() == AlwaysFalse || (ifstates.top() == ElseIsTrue && rawtok->str() != ELIF)) {
conditionIsTrue = false;
}
else if (rawtok->str() == IFDEF) {
conditionIsTrue = (macros.find(rawtok->next->str()) != macros.end() || (hasInclude && rawtok->next->str() == HAS_INCLUDE));
maybeUsedMacros[rawtok->next->str()].emplace_back(rawtok->next->location);
} else if (rawtok->str() == IFNDEF) {
conditionIsTrue = (macros.find(rawtok->next->str()) == macros.end() && !(hasInclude && rawtok->next->str() == HAS_INCLUDE));
maybeUsedMacros[rawtok->next->str()].emplace_back(rawtok->next->location);
} else { /*if (rawtok->str() == IF || rawtok->str() == ELIF)*/
TokenList expr(files);
for (const Token *tok = rawtok->next; tok && tok->location.sameline(rawtok->location); tok = tok->next) {
if (!tok->name) {
expr.push_back(new Token(*tok));
continue;
}
if (tok->str() == DEFINED) {
tok = tok->next;
const bool par = (tok && tok->op == '(');
if (par)
tok = tok->next;
maybeUsedMacros[rawtok->next->str()].emplace_back(rawtok->next->location);
if (tok) {
if (macros.find(tok->str()) != macros.end())
expr.push_back(new Token("1", tok->location));
else if (hasInclude && tok->str() == HAS_INCLUDE)
expr.push_back(new Token("1", tok->location));
else
expr.push_back(new Token("0", tok->location));
}
if (par)
tok = tok ? tok->next : nullptr;
if (!tok || !sameline(rawtok,tok) || (par && tok->op != ')')) {
if (outputList) {
Output out{
Output::SYNTAX_ERROR,
rawtok->location,
"failed to evaluate " + std::string(rawtok->str() == IF ? "#if" : "#elif") + " condition"
};
outputList->emplace_back(std::move(out));
}
output.clear();
return;
}
continue;
}
if (hasInclude && tok->str() == HAS_INCLUDE) {
tok = tok->next;
const bool par = (tok && tok->op == '(');
if (par)
tok = tok->next;
bool closingAngularBracket = false;
if (tok) {
const std::string &sourcefile = rawtokens.file(rawtok->location);
const bool systemheader = (tok && tok->op == '<');
std::string header;
if (systemheader) {
while ((tok = tok->next) && tok->op != '>')
header += tok->str();
if (tok && tok->op == '>')
closingAngularBracket = true;
} else {
header = tok->str().substr(1U, tok->str().size() - 2U);
closingAngularBracket = true;
}
if (tok) {
std::ifstream f;
const std::string header2 = openHeader(f,dui,sourcefile,header,systemheader);
expr.push_back(new Token(header2.empty() ? "0" : "1", tok->location));
}
}
if (par)
tok = tok ? tok->next : nullptr;
if (!tok || !sameline(rawtok,tok) || (par && tok->op != ')') || (!closingAngularBracket)) {
if (outputList) {
Output out{
Output::SYNTAX_ERROR,
rawtok->location,
"failed to evaluate " + std::string(rawtok->str() == IF ? "#if" : "#elif") + " condition"
};
outputList->emplace_back(std::move(out));
}
output.clear();
return;
}
continue;
}
maybeUsedMacros[rawtok->next->str()].emplace_back(rawtok->next->location);
const Token *tmp = tok;
if (!preprocessToken(expr, tmp, macros, files, outputList)) {
output.clear();
return;
}
if (!tmp)
break;
tok = tmp->previous;
}
try {
if (ifCond) {
std::string E;
for (const simplecpp::Token *tok = expr.cfront(); tok; tok = tok->next)
E += (E.empty() ? "" : " ") + tok->str();
const long long result = evaluate(expr, dui, sizeOfType);
conditionIsTrue = (result != 0);
ifCond->emplace_back(rawtok->location, E, result);
} else {
const long long result = evaluate(expr, dui, sizeOfType);
conditionIsTrue = (result != 0);
}
} catch (const std::runtime_error &e) {
if (outputList) {
std::string msg = "failed to evaluate " + std::string(rawtok->str() == IF ? "#if" : "#elif") + " condition";
if (e.what() && *e.what())
msg += std::string(", ") + e.what();
Output out{
Output::SYNTAX_ERROR,
rawtok->location,
std::move(msg)
};
outputList->emplace_back(std::move(out));
}
output.clear();
return;
}
}
if (rawtok->str() != ELIF) {
// push a new ifstate..
if (ifstates.top() != True)
ifstates.push(AlwaysFalse);
else
ifstates.push(conditionIsTrue ? True : ElseIsTrue);
iftokens.push(rawtok);
} else {
if (ifstates.top() == True)
ifstates.top() = AlwaysFalse;
else if (ifstates.top() == ElseIsTrue && conditionIsTrue)
ifstates.top() = True;
iftokens.top()->nextcond = rawtok;
iftokens.top() = rawtok;
}
} else if (rawtok->str() == ELSE) {
ifstates.top() = (ifstates.top() == ElseIsTrue) ? True : AlwaysFalse;
iftokens.top()->nextcond = rawtok;
iftokens.top() = rawtok;
} else if (rawtok->str() == ENDIF) {
ifstates.pop();
iftokens.top()->nextcond = rawtok;
iftokens.pop();
} else if (rawtok->str() == UNDEF) {
if (ifstates.top() == True) {
const Token *tok = rawtok->next;
while (sameline(rawtok,tok) && tok->comment)
tok = tok->next;
if (sameline(rawtok, tok))
macros.erase(tok->str());
}
} else if (ifstates.top() == True && rawtok->str() == PRAGMA && rawtok->next && rawtok->next->str() == ONCE && sameline(rawtok,rawtok->next)) {
pragmaOnce.insert(rawtokens.file(rawtok->location));
}
if (ifstates.top() != True && rawtok->nextcond)
rawtok = rawtok->nextcond->previous;
else
rawtok = gotoNextLine(rawtok);
continue;
}
if (ifstates.top() != True) {
// drop code
rawtok = gotoNextLine(rawtok);
continue;
}
bool hash=false, hashhash=false;
if (rawtok->op == '#' && sameline(rawtok,rawtok->next)) {
if (rawtok->next->op != '#') {
hash = true;
rawtok = rawtok->next; // skip '#'
} else if (sameline(rawtok,rawtok->next->next)) {
hashhash = true;
rawtok = rawtok->next->next; // skip '#' '#'
}
}
const Location loc(rawtok->location);
TokenList tokens(files);
if (!preprocessToken(tokens, rawtok, macros, files, outputList)) {
output.clear();
return;
}
if (hash || hashhash) {
std::string s;
for (const Token *hashtok = tokens.cfront(); hashtok; hashtok = hashtok->next)
s += hashtok->str();
if (hash)
output.push_back(new Token('\"' + s + '\"', loc));
else if (output.back())
output.back()->setstr(output.cback()->str() + s);
else
output.push_back(new Token(s, loc));
} else {
output.takeTokens(tokens);
}
}
if (macroUsage) {
for (simplecpp::MacroMap::const_iterator macroIt = macros.begin(); macroIt != macros.end(); ++macroIt) {
const Macro &macro = macroIt->second;
std::list<Location> usage = macro.usage();
const std::list<Location>& temp = maybeUsedMacros[macro.name()];
usage.insert(usage.end(), temp.begin(), temp.end());
for (std::list<Location>::const_iterator usageIt = usage.begin(); usageIt != usage.end(); ++usageIt) {
MacroUsage mu(macro.valueDefinedInCode());
mu.macroName = macro.name();
mu.macroLocation = macro.defineLocation();
mu.useLocation = *usageIt;
macroUsage->emplace_back(std::move(mu));
}
}
}
}
void simplecpp::cleanup(FileDataCache &cache)
{
cache.clear();
}
simplecpp::cstd_t simplecpp::getCStd(const std::string &std)
{
if (std == "c90" || std == "c89" || std == "iso9899:1990" || std == "iso9899:199409" || std == "gnu90" || std == "gnu89")
return C89;
if (std == "c99" || std == "c9x" || std == "iso9899:1999" || std == "iso9899:199x" || std == "gnu99" || std == "gnu9x")
return C99;
if (std == "c11" || std == "c1x" || std == "iso9899:2011" || std == "gnu11" || std == "gnu1x")
return C11;
if (std == "c17" || std == "c18" || std == "iso9899:2017" || std == "iso9899:2018" || std == "gnu17" || std == "gnu18")
return C17;
if (std == "c23" || std == "gnu23" || std == "c2x" || std == "gnu2x")
return C23;
if (std == "c2y" || std == "gnu2y")
return C2Y;
return CUnknown;
}
std::string simplecpp::getCStdString(cstd_t std)
{
switch (std) {
case C89:
// __STDC_VERSION__ is not set for C90 although the macro was added in the 1994 amendments
return "";
case C99:
return "199901L";
case C11:
return "201112L";
case C17:
return "201710L";
case C23:
// supported by GCC 9+ and Clang 9+
// Clang 9, 10, 11, 12, 13 return "201710L"
// Clang 14, 15, 16, 17 return "202000L"
// Clang 9, 10, 11, 12, 13, 14, 15, 16, 17 do not support "c23" and "gnu23"
return "202311L";
case C2Y:
// supported by GCC 15+ and Clang 19+
// Clang 19, 20, 21, 22 return "202400L"
return "202500L";
case CUnknown:
return "";
}
return "";
}
std::string simplecpp::getCStdString(const std::string &std)
{
return getCStdString(getCStd(std));
}
simplecpp::cppstd_t simplecpp::getCppStd(const std::string &std)
{
if (std == "c++98" || std == "c++03" || std == "gnu++98" || std == "gnu++03")
return CPP03;
if (std == "c++11" || std == "gnu++11" || std == "c++0x" || std == "gnu++0x")
return CPP11;
if (std == "c++14" || std == "c++1y" || std == "gnu++14" || std == "gnu++1y")
return CPP14;
if (std == "c++17" || std == "c++1z" || std == "gnu++17" || std == "gnu++1z")
return CPP17;
if (std == "c++20" || std == "c++2a" || std == "gnu++20" || std == "gnu++2a")
return CPP20;
if (std == "c++23" || std == "c++2b" || std == "gnu++23" || std == "gnu++2b")
return CPP23;
if (std == "c++26" || std == "c++2c" || std == "gnu++26" || std == "gnu++2c")
return CPP26;
return CPPUnknown;
}
std::string simplecpp::getCppStdString(cppstd_t std)
{
switch (std) {
case CPP03:
return "199711L";
case CPP11:
return "201103L";
case CPP14:
return "201402L";
case CPP17:
return "201703L";
case CPP20:
// GCC 10 returns "201703L" - correct in 11+
return "202002L";
case CPP23:
// supported by GCC 11+ and Clang 12+
// GCC 11, 12, 13 return "202100L"
// Clang 12, 13, 14, 15, 16 do not support "c++23" and "gnu++23" and return "202101L"
// Clang 17, 18 return "202302L"
return "202302L";
case CPP26:
// supported by GCC 14+ and Clang 17+
return "202400L";
case CPPUnknown:
return "";
}
return "";
}
std::string simplecpp::getCppStdString(const std::string &std)
{
return getCppStdString(getCppStd(std));
}