cppcheck/lib/preprocessor.cpp

3337 lines
119 KiB
C++
Raw Normal View History

/*
* Cppcheck - A tool for static C/C++ code analysis
2016-01-01 14:34:45 +01:00
* Copyright (C) 2007-2016 Cppcheck team.
*
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*/
#include "preprocessor.h"
#include "tokenize.h"
#include "token.h"
#include "path.h"
#include "errorlogger.h"
#include "settings.h"
#include <algorithm>
#include <sstream>
#include <fstream>
#include <cstdlib>
#include <cctype>
#include <vector>
#include <set>
#include <stack>
/**
* Remove heading and trailing whitespaces from the input parameter.
* If string is all spaces/tabs, return empty string.
* @param s The string to trim.
*/
static std::string trim(const std::string& s)
{
const std::string::size_type beg = s.find_first_not_of(" \t");
if (beg == std::string::npos)
return "";
const std::string::size_type end = s.find_last_not_of(" \t");
return s.substr(beg, end - beg + 1);
}
Directive::Directive(const std::string &_file, const int _linenr, const std::string &_str):
file(_file),
linenr(_linenr),
str(_str)
{
// strip C++ comment if there is one
std::size_t pos = str.find("//");
if (pos != std::string::npos)
str.erase(pos);
// strip any C comments
while ((pos = str.find("/*")) != std::string::npos) {
std::size_t end = str.find("*/", pos+2);
if (end != std::string::npos) {
str.erase(pos, end + 2 - pos);
} else { // treat '/*' as '//' if '*/' is missing
str.erase(pos);
}
}
str = trim(str);
}
bool Preprocessor::missingIncludeFlag;
bool Preprocessor::missingSystemIncludeFlag;
char Preprocessor::macroChar = char(1);
Preprocessor::Preprocessor(Settings& settings, ErrorLogger *errorLogger) : _settings(settings), _errorLogger(errorLogger)
{
}
void Preprocessor::writeError(const std::string &fileName, const unsigned int linenr, ErrorLogger *errorLogger, const std::string &errorType, const std::string &errorText)
{
if (!errorLogger)
return;
std::list<ErrorLogger::ErrorMessage::FileLocation> locationList;
ErrorLogger::ErrorMessage::FileLocation loc(fileName, linenr);
locationList.push_back(loc);
errorLogger->reportErr(ErrorLogger::ErrorMessage(locationList,
Severity::error,
errorText,
errorType,
false));
}
static unsigned char readChar(std::istream &istr, unsigned int bom)
{
unsigned char ch = (unsigned char)istr.get();
// For UTF-16 encoded files the BOM is 0xfeff/0xfffe. If the
// character is non-ASCII character then replace it with 0xff
if (bom == 0xfeff || bom == 0xfffe) {
const unsigned char ch2 = (unsigned char)istr.get();
const int ch16 = (bom == 0xfeff) ? (ch<<8 | ch2) : (ch2<<8 | ch);
ch = (unsigned char)((ch16 >= 0x80) ? 0xff : ch16);
}
// Handling of newlines..
2011-10-13 20:53:06 +02:00
if (ch == '\r') {
ch = '\n';
if (bom == 0 && (char)istr.peek() == '\n')
(void)istr.get();
else if (bom == 0xfeff || bom == 0xfffe) {
int c1 = istr.get();
int c2 = istr.get();
int ch16 = (bom == 0xfeff) ? (c1<<8 | c2) : (c2<<8 | c1);
if (ch16 != '\n') {
istr.unget();
istr.unget();
}
}
}
return ch;
}
2012-01-02 15:52:19 +01:00
// Concatenates a list of strings, inserting a separator between parts
static std::string join(const std::set<std::string>& list, char separator)
{
std::string s;
for (std::set<std::string>::const_iterator it = list.begin(); it != list.end(); ++it) {
if (!s.empty())
s += separator;
s += *it;
}
return s;
}
// Removes duplicate string portions separated by the specified separator
static std::string unify(const std::string &s, char separator)
{
2012-01-02 15:52:19 +01:00
std::set<std::string> parts;
std::string::size_type prevPos = 0;
2011-10-13 20:53:06 +02:00
for (std::string::size_type pos = 0; pos < s.length(); ++pos) {
if (s[pos] == separator) {
if (pos > prevPos)
2012-01-02 15:52:19 +01:00
parts.insert(s.substr(prevPos, pos - prevPos));
prevPos = pos + 1;
}
}
if (prevPos < s.length())
2012-01-02 15:52:19 +01:00
parts.insert(s.substr(prevPos));
2012-01-02 15:52:19 +01:00
return join(parts, separator);
}
bool Preprocessor::cplusplus(const Settings *settings, const std::string &filename)
{
const bool undef = settings && settings->userUndefs.find("__cplusplus") != settings->userUndefs.end();
const bool cpplang = settings && settings->enforcedLang == Settings::CPP;
const bool cppfile = (!settings || settings->enforcedLang == Settings::None) && Path::isCPP(filename);
return (!undef && (cpplang || cppfile));
}
/**
* Get cfgmap - a map of macro names and values
*/
static std::map<std::string,std::string> getcfgmap(const std::string &cfg, const Settings *settings, const std::string &filename)
{
std::map<std::string, std::string> cfgmap;
if (!cfg.empty()) {
std::string::size_type pos = 0;
for (;;) {
std::string::size_type pos2 = cfg.find_first_of(";=", pos);
if (pos2 == std::string::npos) {
cfgmap[cfg.substr(pos)] = "";
break;
}
if (cfg[pos2] == ';') {
cfgmap[cfg.substr(pos, pos2-pos)] = "";
} else {
std::string::size_type pos3 = pos2;
pos2 = cfg.find(';', pos2);
if (pos2 == std::string::npos) {
cfgmap[cfg.substr(pos, pos3-pos)] = cfg.substr(pos3 + 1);
break;
} else {
cfgmap[cfg.substr(pos, pos3-pos)] = cfg.substr(pos3 + 1, pos2 - pos3 - 1);
}
}
pos = pos2 + 1;
}
}
if (cfgmap.find("__cplusplus") == cfgmap.end() && Preprocessor::cplusplus(settings,filename))
cfgmap["__cplusplus"] = "1";
return cfgmap;
}
/** Just read the code into a string. Perform simple cleanup of the code */
2012-01-05 18:37:15 +01:00
std::string Preprocessor::read(std::istream &istr, const std::string &filename)
{
// The UTF-16 BOM is 0xfffe or 0xfeff.
unsigned int bom = 0;
if (istr.peek() >= 0xfe) {
2015-11-28 10:11:07 +01:00
bom = ((unsigned char)istr.get() << 8);
if (istr.peek() >= 0xfe)
2015-11-28 10:11:07 +01:00
bom |= (unsigned char)istr.get();
else
bom = 0; // allowed boms are 0/0xfffe/0xfeff
}
if (_settings.terminated())
return "";
if (_settings.checkConfiguration)
return readpreprocessor(istr,bom);
// ------------------------------------------------------------------------------------------
//
// handling <backslash><newline>
// when this is encountered the <backslash><newline> will be "skipped".
// on the next <newline>, extra newlines will be added
std::ostringstream code;
unsigned int newlines = 0;
for (unsigned char ch = readChar(istr,bom); istr.good(); ch = readChar(istr,bom)) {
// Replace assorted special chars with spaces..
if (((ch & 0x80) == 0) && (ch != '\n') && (std::isspace(ch) || std::iscntrl(ch)))
ch = ' ';
// <backslash><newline>..
// for gcc-compatibility the trailing spaces should be ignored
// for vs-compatibility the trailing spaces should be kept
// See tickets #640 and #1869
// The solution for now is to have a compiler-dependent behaviour.
2011-10-13 20:53:06 +02:00
if (ch == '\\') {
unsigned char chNext;
2012-12-02 20:17:25 +01:00
std::string spaces;
#ifdef __GNUC__
// gcc-compatibility: ignore spaces
2012-12-02 20:17:25 +01:00
for (;; spaces += ' ') {
chNext = (unsigned char)istr.peek();
if (chNext != '\n' && chNext != '\r' &&
2011-10-13 20:53:06 +02:00
(std::isspace(chNext) || std::iscntrl(chNext))) {
// Skip whitespace between <backslash> and <newline>
(void)readChar(istr,bom);
continue;
}
break;
}
#else
// keep spaces
chNext = (unsigned char)istr.peek();
#endif
2011-10-13 20:53:06 +02:00
if (chNext == '\n' || chNext == '\r') {
++newlines;
(void)readChar(istr,bom); // Skip the "<backslash><newline>"
2012-12-02 20:17:25 +01:00
} else {
code << "\\" << spaces;
}
2011-10-13 20:53:06 +02:00
} else {
code << char(ch);
// if there has been <backslash><newline> sequences, add extra newlines..
2011-10-13 20:53:06 +02:00
if (ch == '\n' && newlines > 0) {
code << std::string(newlines, '\n');
newlines = 0;
}
}
}
// ------------------------------------------------------------------------------------------
//
// Remove all comments..
std::string result = removeComments(code.str(), filename);
if (_settings.terminated())
return "";
code.str("");
// ------------------------------------------------------------------------------------------
//
// Clean up all preprocessor statements
result = preprocessCleanupDirectives(result);
if (_settings.terminated())
return "";
// ------------------------------------------------------------------------------------------
//
2011-03-30 16:45:31 +02:00
// Clean up preprocessor #if statements with Parentheses
result = removeParentheses(result);
if (_settings.terminated())
return "";
// Remove '#if 0' blocks
if (result.find("#if 0\n") != std::string::npos)
result = removeIf0(result);
if (_settings.terminated())
return "";
return result;
}
/** read preprocessor statements */
std::string Preprocessor::readpreprocessor(std::istream &istr, const unsigned int bom)
{
enum { NEWLINE, SPACE, PREPROCESSOR, BACKSLASH, OTHER } state = NEWLINE;
std::ostringstream code;
unsigned int newlines = 1;
unsigned char chPrev = ' ';
for (unsigned char ch = readChar(istr,bom); istr.good(); ch = readChar(istr,bom)) {
// Replace assorted special chars with spaces..
if (((ch & 0x80) == 0) && (ch != '\n') && (std::isspace(ch) || std::iscntrl(ch)))
ch = ' ';
if (ch == ' ' && chPrev == ' ')
continue;
if (state == PREPROCESSOR && chPrev == '/' && (ch == '/' || ch == '*'))
state = OTHER;
chPrev = ch;
if (ch == '\n') {
if (state != BACKSLASH) {
state = NEWLINE;
code << std::string(newlines, '\n');
newlines = 1;
} else {
++newlines;
state = PREPROCESSOR;
}
continue;
}
switch (state) {
case NEWLINE:
if (ch==' ')
state = SPACE;
else if (ch == '#') {
state = PREPROCESSOR;
code << ch;
} else
state = OTHER;
break;
case SPACE:
if (ch == '#') {
state = PREPROCESSOR;
code << ch;
} else if (ch != ' ')
state = OTHER;
break;
case PREPROCESSOR:
code << ch;
if (ch == '\\')
state = BACKSLASH;
break;
case BACKSLASH:
code << ch;
if (ch != ' ')
state = PREPROCESSOR;
break;
case OTHER:
break;
};
}
std::string result = preprocessCleanupDirectives(code.str());
result = removeParentheses(result);
return removeIf0(result);
}
std::string Preprocessor::preprocessCleanupDirectives(const std::string &processedFile)
{
std::ostringstream code;
std::istringstream sstr(processedFile);
std::string line;
2011-10-13 20:53:06 +02:00
while (std::getline(sstr, line)) {
// Trim lines..
if (!line.empty() && line[0] == ' ')
line.erase(0, line.find_first_not_of(" "));
if (!line.empty() && line.back() == ' ')
line.erase(line.find_last_not_of(" ") + 1);
// Preprocessor
2011-10-13 20:53:06 +02:00
if (!line.empty() && line[0] == '#') {
enum {
ESC_NONE,
ESC_SINGLE,
ESC_DOUBLE
} escapeStatus = ESC_NONE;
char prev = ' '; // hack to make it skip spaces between # and the directive
code << "#";
std::string::const_iterator i = line.begin();
++i;
// need space.. #if( => #if (
bool needSpace = true;
2011-10-13 20:53:06 +02:00
while (i != line.end()) {
// disable esc-mode
2011-10-13 20:53:06 +02:00
if (escapeStatus != ESC_NONE) {
if (prev != '\\' && escapeStatus == ESC_SINGLE && *i == '\'') {
escapeStatus = ESC_NONE;
}
2011-10-13 20:53:06 +02:00
if (prev != '\\' && escapeStatus == ESC_DOUBLE && *i == '"') {
escapeStatus = ESC_NONE;
}
2011-10-13 20:53:06 +02:00
} else {
// enable esc-mode
if (escapeStatus == ESC_NONE && *i == '"')
escapeStatus = ESC_DOUBLE;
if (escapeStatus == ESC_NONE && *i == '\'')
escapeStatus = ESC_SINGLE;
}
// skip double whitespace between arguments
2011-10-13 20:53:06 +02:00
if (escapeStatus == ESC_NONE && prev == ' ' && *i == ' ') {
++i;
continue;
}
// Convert #if( to "#if ("
2011-10-13 20:53:06 +02:00
if (escapeStatus == ESC_NONE) {
if (needSpace) {
if (*i == '(' || *i == '!')
code << " ";
else if (!std::isalpha((unsigned char)*i))
needSpace = false;
}
if (*i == '#')
needSpace = true;
}
code << *i;
2011-10-13 20:53:06 +02:00
if (escapeStatus != ESC_NONE && prev == '\\' && *i == '\\') {
prev = ' ';
2011-10-13 20:53:06 +02:00
} else {
prev = *i;
}
++i;
}
2011-10-13 20:53:06 +02:00
if (escapeStatus != ESC_NONE) {
// unmatched quotes.. compiler should probably complain about this..
}
2011-10-13 20:53:06 +02:00
} else {
// Do not mess with regular code..
code << line;
}
code << (sstr.eof()?"":"\n");
}
return code.str();
}
2009-10-10 09:29:06 +02:00
static bool hasbom(const std::string &str)
{
return bool(str.size() >= 3 &&
static_cast<unsigned char>(str[0]) == 0xef &&
static_cast<unsigned char>(str[1]) == 0xbb &&
static_cast<unsigned char>(str[2]) == 0xbf);
2009-10-10 09:29:06 +02:00
}
// This wrapper exists because Sun's CC does not allow a static_cast
// from extern "C" int(*)(int) to int(*)(int).
static int tolowerWrapper(int c)
{
return std::tolower(c);
}
static bool isFallThroughComment(std::string comment)
{
// convert comment to lower case without whitespace
2011-10-13 20:53:06 +02:00
for (std::string::iterator i = comment.begin(); i != comment.end();) {
2011-10-16 07:52:54 +02:00
if (std::isspace(static_cast<unsigned char>(*i)))
i = comment.erase(i);
else
++i;
}
std::transform(comment.begin(), comment.end(), comment.begin(), tolowerWrapper);
return comment.find("fallthr") != std::string::npos ||
comment.find("fallsthr") != std::string::npos ||
comment.find("fall-thr") != std::string::npos ||
comment.find("dropthr") != std::string::npos ||
comment.find("passthr") != std::string::npos ||
comment.find("nobreak") != std::string::npos ||
comment == "fall";
}
std::string Preprocessor::removeComments(const std::string &str, const std::string &filename)
{
// For the error report
unsigned int lineno = 1;
// handling <backslash><newline>
// when this is encountered the <backslash><newline> will be "skipped".
// on the next <newline>, extra newlines will be added
unsigned int newlines = 0;
std::ostringstream code;
unsigned char previous = 0;
bool inPreprocessorLine = false;
std::vector<std::string> suppressionIDs;
const bool detectFallThroughComments = _settings.experimental && _settings.isEnabled("style");
bool fallThroughComment = false;
2011-10-13 20:53:06 +02:00
for (std::string::size_type i = hasbom(str) ? 3U : 0U; i < str.length(); ++i) {
unsigned char ch = static_cast<unsigned char>(str[i]);
2011-10-13 20:53:06 +02:00
if (ch & 0x80) {
2010-02-27 17:06:33 +01:00
std::ostringstream errmsg;
errmsg << "(character code = 0x" << std::hex << (int(ch) & 0xff) << ")";
std::string info = errmsg.str();
errmsg.str("");
errmsg << "The code contains unhandled characters " << info << ". Checking continues, but do not expect valid results.\n"
<< "The code contains characters that are unhandled " << info << ". Neither unicode nor extended ASCII are supported. Checking continues, but do not expect valid results.";
writeError(Path::simplifyPath(filename), lineno, _errorLogger, "unhandledCharacters", errmsg.str());
}
if (_settings.terminated())
return "";
// First skip over any whitespace that may be present
2011-10-13 20:53:06 +02:00
if (std::isspace(ch)) {
if (ch == ' ' && previous == ' ') {
// Skip double white space
2011-10-13 20:53:06 +02:00
} else {
code << char(ch);
previous = ch;
}
// if there has been <backslash><newline> sequences, add extra newlines..
2011-10-13 20:53:06 +02:00
if (ch == '\n') {
if (previous != '\\')
inPreprocessorLine = false;
++lineno;
2011-10-13 20:53:06 +02:00
if (newlines > 0) {
code << std::string(newlines, '\n');
newlines = 0;
previous = '\n';
}
}
continue;
}
if ((ch == '#') && (str.compare(i+1, 6, "error ") == 0 || str.compare(i+1, 8, "warning ") == 0)) {
if (str.compare(i, 6, "#error") == 0)
code << "#error";
i = str.find('\n', i);
if (i == std::string::npos)
break;
--i;
continue;
}
// Remove comments..
if (str.compare(i, 2, "//") == 0) {
const std::size_t commentStart = i + 2;
i = str.find('\n', i);
if (i == std::string::npos)
break;
std::string comment(str, commentStart, i - commentStart);
if (_settings.inlineSuppressions) {
std::istringstream iss(comment);
std::string word;
iss >> word;
2011-10-13 20:53:06 +02:00
if (word == "cppcheck-suppress") {
iss >> word;
if (iss)
suppressionIDs.push_back(word);
}
}
if (detectFallThroughComments && isFallThroughComment(comment)) {
fallThroughComment = true;
}
code << "\n";
previous = '\n';
++lineno;
} else if (str.compare(i, 2, "/*") == 0) {
const std::size_t commentStart = i + 2;
unsigned char chPrev = 0;
++i;
2011-10-13 20:53:06 +02:00
while (i < str.length() && (chPrev != '*' || ch != '/')) {
chPrev = ch;
++i;
ch = static_cast<unsigned char>(str[i]);
2011-10-13 20:53:06 +02:00
if (ch == '\n') {
++newlines;
++lineno;
}
}
std::string comment(str, commentStart, i - commentStart - 1);
if (detectFallThroughComments && isFallThroughComment(comment)) {
fallThroughComment = true;
}
if (_settings.inlineSuppressions) {
std::istringstream iss(comment);
std::string word;
iss >> word;
2011-10-13 20:53:06 +02:00
if (word == "cppcheck-suppress") {
iss >> word;
if (iss)
suppressionIDs.push_back(word);
}
}
} else if ((i == 0 || std::isspace((unsigned char)str[i-1])) && str.compare(i, 5, "__asm") == 0) {
while (i < str.size() && (std::isalpha((unsigned char)str[i]) || str[i] == '_'))
code << str[i++];
while (i < str.size() && std::isspace((unsigned char)str[i])) {
if (str[i] == '\n')
lineno++;
code << str[i++];
}
if (str[i] == '{') {
// Ticket 4873: Extract comments from the __asm / __asm__'s content
std::string asmBody;
while (i < str.size() && str[i] != '}') {
if (str[i] == ';') {
std::string::size_type backslashN = str.find('\n', i);
if (backslashN != std::string::npos) // Ticket #4922: Don't go in infinite loop or crash if there is no '\n'
i = backslashN;
}
if (str[i] == '\n')
lineno++;
asmBody += str[i++];
}
code << removeComments(asmBody, filename);
code << '}';
} else
--i;
2011-10-13 20:53:06 +02:00
} else if (ch == '#' && previous == '\n') {
code << ch;
previous = ch;
inPreprocessorLine = true;
// Add any pending inline suppressions that have accumulated.
2011-10-13 20:53:06 +02:00
if (!suppressionIDs.empty()) {
// Add the suppressions.
for (std::size_t j = 0; j < suppressionIDs.size(); ++j) {
const std::string errmsg(_settings.nomsg.addSuppression(suppressionIDs[j], filename, lineno));
if (!errmsg.empty()) {
writeError(filename, lineno, _errorLogger, "cppcheckError", errmsg);
}
}
suppressionIDs.clear();
}
2011-10-13 20:53:06 +02:00
} else {
if (!inPreprocessorLine) {
// Not whitespace, not a comment, and not preprocessor.
// Must be code here!
// First check for a "fall through" comment match, but only
// add a suppression if the next token is 'case' or 'default'
if (detectFallThroughComments && fallThroughComment) {
const std::string::size_type j = str.find_first_not_of("abcdefghijklmnopqrstuvwxyz", i);
2015-10-14 11:57:20 +02:00
if (str.compare(i, j-i, "case") == 0 || str.compare(i, j-i, "default") == 0)
suppressionIDs.push_back("switchCaseFallThrough");
fallThroughComment = false;
}
// Add any pending inline suppressions that have accumulated.
2011-10-13 20:53:06 +02:00
if (!suppressionIDs.empty()) {
// Relative filename
std::string relativeFilename(filename);
if (_settings.relativePaths) {
for (std::size_t j = 0U; j < _settings.basePaths.size(); ++j) {
const std::string bp = _settings.basePaths[j] + "/";
if (relativeFilename.compare(0,bp.size(),bp)==0) {
relativeFilename = relativeFilename.substr(bp.size());
}
}
}
// Add the suppressions.
for (std::size_t j = 0; j < suppressionIDs.size(); ++j) {
const std::string errmsg(_settings.nomsg.addSuppression(suppressionIDs[j], relativeFilename, lineno));
if (!errmsg.empty()) {
writeError(filename, lineno, _errorLogger, "cppcheckError", errmsg);
}
}
suppressionIDs.clear();
}
}
2015-10-14 11:57:20 +02:00
// C++14 digit separators
if (ch == '\'' && std::isxdigit(previous))
2015-10-14 11:57:20 +02:00
; // Just skip it.
// String or char constants..
2015-10-14 11:57:20 +02:00
else if (ch == '\"' || ch == '\'') {
code << char(ch);
char chNext;
2011-10-13 20:53:06 +02:00
do {
++i;
chNext = str[i];
2011-10-13 20:53:06 +02:00
if (chNext == '\\') {
++i;
const char chSeq = str[i];
if (chSeq == '\n')
++newlines;
2011-10-13 20:53:06 +02:00
else {
code << chNext;
code << chSeq;
previous = static_cast<unsigned char>(chSeq);
}
2011-10-13 20:53:06 +02:00
} else {
code << chNext;
previous = static_cast<unsigned char>(chNext);
}
2011-10-13 20:53:06 +02:00
} while (i < str.length() && chNext != ch && chNext != '\n');
}
// Rawstring..
2011-10-13 20:53:06 +02:00
else if (str.compare(i,2,"R\"")==0) {
std::string delim;
2011-10-13 20:53:06 +02:00
for (std::string::size_type i2 = i+2; i2 < str.length(); ++i2) {
if (i2 > 16 + i ||
std::isspace(str[i2]) ||
std::iscntrl(str[i2]) ||
str[i2] == ')' ||
2011-10-13 20:53:06 +02:00
str[i2] == '\\') {
delim = " ";
break;
2011-10-13 20:53:06 +02:00
} else if (str[i2] == '(')
break;
delim += str[i2];
}
const std::string::size_type endpos = str.find(")" + delim + "\"", i);
2011-10-13 20:53:06 +02:00
if (delim != " " && endpos != std::string::npos) {
unsigned int rawstringnewlines = 0;
code << '\"';
2011-10-13 20:53:06 +02:00
for (std::string::size_type p = i + 3 + delim.size(); p < endpos; ++p) {
if (str[p] == '\n') {
rawstringnewlines++;
code << "\\n";
2011-10-13 20:53:06 +02:00
} else if (std::iscntrl((unsigned char)str[p]) ||
std::isspace((unsigned char)str[p])) {
code << " ";
} else if (str[p] == '\\') {
code << "\\\\";
} else if (str[p] == '\"') {
code << "\\" << (char)str[p];
2011-10-13 20:53:06 +02:00
} else {
code << (char)str[p];
}
}
code << "\"";
if (rawstringnewlines > 0)
code << std::string(rawstringnewlines, '\n');
i = endpos + delim.size() + 1;
2011-10-13 20:53:06 +02:00
} else {
code << "R";
previous = 'R';
}
2011-10-13 20:53:06 +02:00
} else {
code << char(ch);
previous = ch;
}
}
}
return code.str();
}
std::string Preprocessor::removeIf0(const std::string &code)
{
std::ostringstream ret;
std::istringstream istr(code);
std::string line;
2011-10-13 20:53:06 +02:00
while (std::getline(istr,line)) {
2012-01-02 15:52:19 +01:00
ret << line << "\n";
if (line == "#if 0") {
// goto the end of the '#if 0' block
unsigned int level = 1;
bool in = false;
2011-10-13 20:53:06 +02:00
while (level > 0 && std::getline(istr,line)) {
if (line.compare(0,3,"#if") == 0)
++level;
else if (line == "#endif")
--level;
2011-10-13 20:53:06 +02:00
else if ((line == "#else") || (line.compare(0, 5, "#elif") == 0)) {
if (level == 1)
in = true;
2011-10-13 20:53:06 +02:00
} else {
if (in)
ret << line << "\n";
else
// replace code within '#if 0' block with empty lines
ret << "\n";
continue;
}
2011-02-26 20:09:14 +01:00
ret << line << "\n";
}
}
}
return ret.str();
}
2011-03-30 16:45:31 +02:00
std::string Preprocessor::removeParentheses(const std::string &str)
{
if (str.find("\n#if") == std::string::npos && str.compare(0, 3, "#if") != 0)
return str;
2011-12-08 21:28:34 +01:00
std::istringstream istr(str);
std::ostringstream ret;
std::string line;
2011-10-13 20:53:06 +02:00
while (std::getline(istr, line)) {
if (line.compare(0, 3, "#if") == 0 || line.compare(0, 5, "#elif") == 0) {
2009-10-04 15:41:50 +02:00
std::string::size_type pos;
pos = 0;
while ((pos = line.find(" (", pos)) != std::string::npos)
2009-10-04 15:41:50 +02:00
line.erase(pos, 1);
pos = 0;
while ((pos = line.find("( ", pos)) != std::string::npos)
2009-10-04 15:41:50 +02:00
line.erase(pos + 1, 1);
pos = 0;
while ((pos = line.find(" )", pos)) != std::string::npos)
2009-10-04 15:41:50 +02:00
line.erase(pos, 1);
pos = 0;
while ((pos = line.find(") ", pos)) != std::string::npos)
2009-10-04 15:41:50 +02:00
line.erase(pos + 1, 1);
// Remove inner parentheses "((..))"..
2009-10-04 15:41:50 +02:00
pos = 0;
2011-10-13 20:53:06 +02:00
while ((pos = line.find("((", pos)) != std::string::npos) {
2009-10-04 15:41:50 +02:00
++pos;
std::string::size_type pos2 = line.find_first_of("()", pos + 1);
2011-10-13 20:53:06 +02:00
if (pos2 != std::string::npos && line[pos2] == ')') {
2009-10-04 15:41:50 +02:00
line.erase(pos2, 1);
line.erase(pos, 1);
}
}
// "#if(A) => #if A", but avoid "#if (defined A) || defined (B)"
if ((line.compare(0, 4, "#if(") == 0 || line.compare(0, 6, "#elif(") == 0) &&
2011-10-13 20:53:06 +02:00
line[line.length() - 1] == ')') {
int ind = 0;
2011-10-13 20:53:06 +02:00
for (std::string::size_type i = 0; i < line.length(); ++i) {
if (line[i] == '(')
++ind;
2011-10-13 20:53:06 +02:00
else if (line[i] == ')') {
--ind;
2011-10-13 20:53:06 +02:00
if (ind == 0) {
if (i == line.length() - 1) {
line[line.find('(')] = ' ';
line.erase(line.length() - 1);
}
break;
}
}
}
}
if (line.compare(0, 4, "#if(") == 0)
line.insert(3, " ");
else if (line.compare(0, 6, "#elif(") == 0)
line.insert(5, " ");
}
ret << line << "\n";
}
return ret.str();
}
void Preprocessor::removeAsm(std::string &str)
{
std::string::size_type pos = 0;
2011-10-13 20:53:06 +02:00
while ((pos = str.find("#asm\n", pos)) != std::string::npos) {
str.replace(pos, 4, "asm(");
2011-05-11 19:27:19 +02:00
std::string::size_type pos2 = str.find("#endasm", pos);
if (pos2 != std::string::npos) {
str.replace(pos2, 7, ");");
pos = pos2;
2011-05-11 19:27:19 +02:00
}
}
}
void Preprocessor::preprocess(std::istream &istr, std::map<std::string, std::string> &result, const std::string &filename, const std::list<std::string> &includePaths)
{
std::list<std::string> configs;
std::string data;
preprocess(istr, data, configs, filename, includePaths);
for (std::list<std::string>::const_iterator it = configs.begin(); it != configs.end(); ++it) {
if (_settings.userUndefs.find(*it) == _settings.userUndefs.end()) {
result[ *it ] = getcode(data, *it, filename);
}
}
}
std::string Preprocessor::removeSpaceNearNL(const std::string &str)
{
std::string tmp;
char prev = '\n'; // treat start of file as newline
for (std::size_t i = 0; i < str.size(); i++) {
if (str[i] == ' ' &&
(prev == '\n' ||
i + 1 >= str.size() || // treat end of file as newline
str[i+1] == '\n'
)
2011-10-13 20:53:06 +02:00
) {
// Ignore space that has new line in either side of it
2011-10-13 20:53:06 +02:00
} else {
tmp.append(1, str[i]);
prev = str[i];
}
}
return tmp;
}
void Preprocessor::replaceIfDefined(std::string &str) const
{
std::string::size_type pos = 0;
while ((pos = str.find("#if defined(", pos)) != std::string::npos) {
std::string::size_type pos2 = str.find(')', pos + 9);
if (pos2 > str.length() - 1)
break;
if (str[pos2+1] == '\n') {
str.erase(pos2, 1);
str.erase(pos + 3, 9);
str.insert(pos + 3, "def ");
}
++pos;
if (_settings.terminated())
return;
}
pos = 0;
while ((pos = str.find("#if !defined(", pos)) != std::string::npos) {
std::string::size_type pos2 = str.find(')', pos + 9);
if (pos2 > str.length() - 1)
break;
if (str[pos2+1] == '\n') {
str.erase(pos2, 1);
str.erase(pos + 3, 10);
str.insert(pos + 3, "ndef ");
}
++pos;
if (_settings.terminated())
return;
}
2009-10-04 15:41:50 +02:00
pos = 0;
while ((pos = str.find("#elif defined(", pos)) != std::string::npos) {
std::string::size_type pos2 = str.find(')', pos + 9);
if (pos2 > str.length() - 1)
2009-10-04 15:41:50 +02:00
break;
if (str[pos2+1] == '\n') {
str.erase(pos2, 1);
str.erase(pos + 6, 8);
2009-10-04 15:41:50 +02:00
}
++pos;
if (_settings.terminated())
return;
2009-10-04 15:41:50 +02:00
}
}
void Preprocessor::preprocessWhitespaces(std::string &processedFile)
{
// Replace all tabs with spaces..
std::replace(processedFile.begin(), processedFile.end(), '\t', ' ');
// Remove space characters that are after or before new line character
processedFile = removeSpaceNearNL(processedFile);
}
void Preprocessor::preprocess(std::istream &srcCodeStream, std::string &processedFile, std::list<std::string> &resultConfigurations, const std::string &filename, const std::list<std::string> &includePaths)
{
std::string forcedIncludes;
2010-09-03 13:30:49 +02:00
if (file0.empty())
file0 = filename;
2012-01-05 18:37:15 +01:00
processedFile = read(srcCodeStream, filename);
for (std::list<std::string>::iterator it = _settings.userIncludes.begin();
it != _settings.userIncludes.end();
++it) {
const std::string& cur = *it;
// try to open file
std::ifstream fin;
fin.open(cur.c_str());
if (!fin.is_open()) {
missingInclude(cur,
1,
cur,
UserHeader
);
continue;
}
const std::string fileData = read(fin, filename);
2014-02-08 08:51:38 +01:00
fin.close();
forcedIncludes +=
"#file \"" + cur + "\"\n" +
"#line 1\n" +
fileData + "\n" +
"#endfile\n"
;
}
for (std::vector<std::string>::iterator it = _settings.library.defines.begin();
it != _settings.library.defines.end();
++it) {
forcedIncludes += *it;
}
if (!forcedIncludes.empty()) {
processedFile =
2012-12-27 16:59:30 +01:00
forcedIncludes +
"#file \"" + filename + "\"\n" +
"#line 1\n" +
processedFile +
"#endfile\n"
;
}
// Remove asm(...)
removeAsm(processedFile);
// Replace "defined A" with "defined(A)"
{
2011-12-08 21:28:34 +01:00
std::istringstream istr(processedFile);
std::ostringstream ostr;
std::string line;
2011-10-13 20:53:06 +02:00
while (std::getline(istr, line)) {
if (line.compare(0, 4, "#if ") == 0 || line.compare(0, 6, "#elif ") == 0) {
std::string::size_type pos = 0;
2011-10-13 20:53:06 +02:00
while ((pos = line.find(" defined ")) != std::string::npos) {
line[pos+8] = '(';
pos = line.find_first_of(" |&", pos + 8);
if (pos == std::string::npos)
line += ")";
else
line.insert(pos, ")");
if (_settings.terminated())
return;
}
}
ostr << line << "\n";
}
processedFile = ostr.str();
}
std::map<std::string, std::string> defs(getcfgmap(_settings.userDefines, &_settings, filename));
if (_settings.maxConfigs == 1U) {
std::set<std::string> pragmaOnce;
std::list<std::string> includes;
processedFile = handleIncludes(processedFile, filename, includePaths, defs, pragmaOnce, includes);
resultConfigurations = getcfgs(processedFile, filename, defs);
} else {
handleIncludes(processedFile, filename, includePaths);
replaceIfDefined(processedFile);
// Get all possible configurations..
resultConfigurations = getcfgs(processedFile, filename, defs);
// Remove configurations that are disabled by -U
handleUndef(resultConfigurations);
}
}
void Preprocessor::handleUndef(std::list<std::string> &configurations) const
{
if (!_settings.userUndefs.empty()) {
for (std::list<std::string>::iterator cfg = configurations.begin(); cfg != configurations.end();) {
bool undef = false;
for (std::set<std::string>::const_iterator it = _settings.userUndefs.begin(); it != _settings.userUndefs.end(); ++it) {
if (*it == *cfg)
undef = true;
else if (cfg->compare(0,it->length(),*it)==0 && cfg->find_first_of(";=") == it->length())
undef = true;
else if (cfg->find(";" + *it) == std::string::npos)
2015-09-10 11:54:35 +02:00
continue;
else if (cfg->find(";" + *it + ";") != std::string::npos)
undef = true;
else if (cfg->find(";" + *it + "=") != std::string::npos)
undef = true;
else if (cfg->find(";" + *it) + it->size() + 1U == cfg->size())
undef = true;
2015-09-10 11:54:35 +02:00
if (undef)
break;
}
if (undef)
configurations.erase(cfg++);
else
++cfg;
}
}
}
// Get the DEF in this line: "#ifdef DEF"
std::string Preprocessor::getdef(std::string line, bool def)
{
if (line.empty() || line[0] != '#')
2011-01-31 03:37:37 +01:00
return "";
// If def is true, the line must start with "#ifdef"
if (def && line.compare(1, 6, "ifdef ") != 0 && line.compare(1, 3, "if ") != 0
&& (line.compare(1, 5, "elif ") != 0 || line.compare(1, 6, "elif !") == 0)) {
return "";
}
// If def is false, the line must start with "#ifndef"
if (!def && line.compare(1, 7, "ifndef ") != 0 && line.compare(1, 6, "elif !") != 0) {
return "";
}
// Remove the "#ifdef" or "#ifndef"
if (line.compare(1, 11, "if defined ") == 0)
line.erase(0, 11);
else if (line.compare(1, 14, "elif !defined(") == 0) {
2011-01-31 00:33:44 +01:00
line.erase(0, 15);
std::string::size_type pos = line.find(')');
2011-01-31 00:33:44 +01:00
// if pos == ::npos then another part of the code will complain
// about the mismatch
if (pos != std::string::npos)
line.erase(pos, 1);
2011-10-13 20:53:06 +02:00
} else
line.erase(0, line.find(' '));
// Remove all spaces.
std::string::size_type pos = 0;
while ((pos = line.find(' ', pos)) != std::string::npos) {
const unsigned char chprev(static_cast<unsigned char>((pos > 0) ? line[pos-1] : 0));
const unsigned char chnext(static_cast<unsigned char>((pos + 1 < line.length()) ? line[pos+1] : 0));
if ((std::isalnum(chprev) || chprev == '_') && (std::isalnum(chnext) || chnext == '_'))
++pos;
else
line.erase(pos, 1);
}
// The remaining string is our result.
return line;
}
/** Simplify variable in variable map. */
static Token *simplifyVarMapExpandValue(Token *tok, const std::map<std::string, std::string> &variables, std::set<std::string> seenVariables, const Settings& settings)
{
// TODO: handle function-macros too.
// Prevent infinite recursion..
if (seenVariables.find(tok->str()) != seenVariables.end())
return tok;
seenVariables.insert(tok->str());
const std::map<std::string, std::string>::const_iterator it = variables.find(tok->str());
if (it != variables.end()) {
TokenList tokenList(&settings);
std::istringstream istr(it->second);
if (tokenList.createTokens(istr)) {
// expand token list
for (Token *tok2 = tokenList.front(); tok2; tok2 = tok2->next()) {
2013-01-26 18:45:18 +01:00
if (tok2->isName()) {
2015-05-12 14:00:43 +02:00
tok2 = simplifyVarMapExpandValue(tok2, variables, seenVariables, settings);
}
}
// insert token list into "parent" token list
for (const Token *tok2 = tokenList.front(); tok2; tok2 = tok2->next()) {
if (tok2->previous()) {
tok->insertToken(tok2->str());
tok = tok->next();
} else
tok->str(tok2->str());
}
}
}
return tok;
}
/**
* Simplifies the variable map. For example if the map contains A=>B, B=>1, then A=>B is simplified to A=>1.
* @param [in,out] variables - a map of variable name to variable value. This map will be modified.
2015-08-30 20:12:02 +02:00
* @param [in] settings Current settings being used
*/
static void simplifyVarMap(std::map<std::string, std::string> &variables, const Settings& settings)
{
for (std::map<std::string, std::string>::iterator i = variables.begin(); i != variables.end(); ++i) {
TokenList tokenList(&settings);
std::istringstream istr(i->second);
if (tokenList.createTokens(istr)) {
for (Token *tok = tokenList.front(); tok; tok = tok->next()) {
if (tok->isName()) {
std::set<std::string> seenVariables;
2015-05-12 14:00:43 +02:00
tok = simplifyVarMapExpandValue(tok, variables, seenVariables, settings);
}
}
std::string str;
for (const Token *tok = tokenList.front(); tok; tok = tok->next())
str.append((tok->previous() ? " " : "") + tok->str());
i->second = str;
}
}
}
std::list<std::string> Preprocessor::getcfgs(const std::string &filedata, const std::string &filename, const std::map<std::string, std::string> &defs)
{
std::list<std::string> ret;
ret.push_back("");
std::list<std::string> deflist, ndeflist;
// constants defined through "#define" in the code..
std::set<std::string> defines;
std::map<std::string, std::string> alldefinesmap(defs);
std::stack<std::pair<std::string,bool> > includeStack;
includeStack.push(std::pair<std::string,bool>(filename,false));
// How deep into included files are we currently parsing?
// 0=>Source file, 1=>Included by source file, 2=>included by header that was included by source file, etc
int filelevel = 0;
bool includeguard = false;
unsigned int linenr = 0;
std::istringstream istr(filedata);
std::string line;
const bool printDebug = _settings.debugwarnings;
while (std::getline(istr, line)) {
++linenr;
if (_settings.terminated())
return ret;
if (_errorLogger)
_errorLogger->reportProgress(filename, "Preprocessing (get configurations 1)", 0);
if (line.empty())
2011-01-31 03:37:37 +01:00
continue;
2011-10-13 20:53:06 +02:00
if (line.compare(0, 6, "#file ") == 0) {
includeguard = true;
const std::string::size_type start=line.find('\"');
const std::string::size_type end=line.find('\"',start+1);
const std::string includeFile=line.substr(start+1,end-start-1);
++filelevel;
bool fileExcluded = _settings.configurationExcluded(includeFile);
includeStack.push(std::pair<std::string,bool>(includeFile,fileExcluded));
continue;
}
2011-10-13 20:53:06 +02:00
else if (line == "#endfile") {
includeguard = false;
includeStack.pop();
if (filelevel > 0)
--filelevel;
continue;
}
2011-10-13 20:53:06 +02:00
if (line.compare(0, 8, "#define ") == 0) {
bool valid = false;
for (std::string::size_type pos = 8; pos < line.size(); ++pos) {
const char ch = line[pos];
if (ch=='_' || (ch>='a' && ch<='z') || (ch>='A' && ch<='Z') || (pos>8 && ch>='0' && ch<='9')) {
valid = true;
continue;
}
if (ch==' ' || ch=='(') {
if (valid)
break;
}
valid = false;
break;
}
if (!valid)
line.clear();
2011-10-13 20:53:06 +02:00
else {
std::string definestr = line.substr(8);
const std::string::size_type spacepos = definestr.find(' ');
if (spacepos != std::string::npos)
definestr[spacepos] = '=';
defines.insert(definestr);
const std::string::size_type separatorpos = definestr.find_first_of("=(");
if (separatorpos != std::string::npos && definestr[separatorpos] == '=') {
const std::string varname(definestr.substr(0, separatorpos));
const std::string value(definestr.substr(separatorpos + 1));
alldefinesmap[varname] = value;
}
}
}
if (!line.empty() && line.compare(0, 3, "#if") != 0)
includeguard = false;
if (line.empty() || line[0] != '#')
2011-01-31 03:37:37 +01:00
continue;
if (includeguard)
continue;
if (line.compare(0, 5, "#line") == 0)
continue;
bool from_negation = false;
std::string def = getdef(line, true);
2011-10-13 20:53:06 +02:00
if (def.empty()) {
def = getdef(line, false);
// sub conditionals of ndef blocks need to be
// constructed _without_ the negated define
if (!def.empty())
from_negation = true;
}
2011-10-13 20:53:06 +02:00
if (!def.empty()) {
int par = 0;
2011-10-13 20:53:06 +02:00
for (std::string::size_type pos = 0; pos < def.length(); ++pos) {
if (def[pos] == '(')
++par;
2011-10-13 20:53:06 +02:00
else if (def[pos] == ')') {
--par;
if (par < 0)
break;
}
}
2011-10-13 20:53:06 +02:00
if (par != 0) {
2010-04-09 21:40:37 +02:00
std::ostringstream lineStream;
lineStream << __LINE__;
const std::string errorId = "preprocessor" + lineStream.str();
const std::string errorText = "mismatching number of '(' and ')' in this line: " + def;
2014-08-06 09:20:04 +02:00
writeError(filename, linenr, _errorLogger, errorId, errorText);
ret.clear();
return ret;
}
// Replace defined constants
simplifyCondition(alldefinesmap, def, false);
2011-01-31 01:40:59 +01:00
if (! deflist.empty() && line.compare(0, 6, "#elif ") == 0)
deflist.pop_back();
// translate A==1 condition to A=1 configuration
if (def.find("==") != std::string::npos) {
// Check if condition match pattern "%name% == %num%"
// %name%
std::string::size_type pos = 0;
if (std::isalpha((unsigned char)def[pos]) || def[pos] == '_') {
++pos;
while (std::isalnum((unsigned char)def[pos]) || def[pos] == '_')
++pos;
}
// ==
if (def.compare(pos,2,"==")==0)
pos += 2;
// %num%
if (pos<def.size() && std::isdigit(def[pos])) {
if (def.compare(pos,2,"0x")==0) {
pos += 2;
if (pos >= def.size())
pos = 0;
while (pos < def.size() && std::isxdigit((unsigned char)def[pos]))
++pos;
} else {
while (pos < def.size() && std::isdigit((unsigned char)def[pos]))
++pos;
}
// Does the condition match the pattern "%name% == %num%"?
if (pos == def.size()) {
def.erase(def.find("=="),1);
}
}
}
deflist.push_back(def);
def = "";
2011-10-13 20:53:06 +02:00
for (std::list<std::string>::const_iterator it = deflist.begin(); it != deflist.end(); ++it) {
if (*it == "0")
break;
if (*it == "1" || *it == "!")
continue;
// don't add "T;T":
// treat two and more similar nested conditions as one
2011-10-13 20:53:06 +02:00
if (def != *it) {
if (! def.empty())
def += ";";
def += *it;
}
/* TODO: Fix TestPreprocessor::test7e (#2552)
else
{
std::ostringstream lineStream;
lineStream << __LINE__;
ErrorLogger::ErrorMessage errmsg;
ErrorLogger::ErrorMessage::FileLocation loc;
loc.setfile(filename);
loc.line = linenr;
errmsg._callStack.push_back(loc);
errmsg._severity = Severity::error;
errmsg.setmsg(*it+" is already guaranteed to be defined");
errmsg._id = "preprocessor" + lineStream.str();
_errorLogger->reportErr(errmsg);
}
*/
}
2011-10-13 20:53:06 +02:00
if (from_negation) {
ndeflist.push_back(deflist.back());
2011-11-26 21:02:04 +01:00
deflist.back() = "!";
}
2011-10-13 20:53:06 +02:00
if (std::find(ret.begin(), ret.end(), def) == ret.end()) {
if (!includeStack.top().second) {
ret.push_back(def);
} else {
if (_errorLogger && printDebug) {
std::list<ErrorLogger::ErrorMessage::FileLocation> locationList;
const ErrorLogger::ErrorMessage errmsg(locationList, Severity::debug,
"Configuration not considered: " + def +" for file:"+includeStack.top().first, "debug", false);
_errorLogger->reportErr(errmsg);
}
}
}
}
2011-10-13 20:53:06 +02:00
else if (line.compare(0, 5, "#else") == 0 && ! deflist.empty()) {
if (deflist.back() == "!" && !ndeflist.empty()) {
2011-11-26 21:02:04 +01:00
deflist.back() = ndeflist.back();
ndeflist.pop_back();
2011-10-13 20:53:06 +02:00
} else {
2010-04-09 21:40:37 +02:00
std::string tempDef((deflist.back() == "1") ? "0" : "1");
2011-11-26 21:02:04 +01:00
deflist.back() = tempDef;
}
}
2011-10-13 20:53:06 +02:00
else if (line.compare(0, 6, "#endif") == 0 && ! deflist.empty()) {
if (deflist.back() == "!" && !ndeflist.empty())
ndeflist.pop_back();
deflist.pop_back();
}
}
// Remove defined constants from ifdef configurations..
std::size_t count = 0;
2011-10-13 20:53:06 +02:00
for (std::list<std::string>::iterator it = ret.begin(); it != ret.end(); ++it) {
if (_errorLogger)
_errorLogger->reportProgress(filename, "Preprocessing (get configurations 2)", (100 * count++) / ret.size());
std::string cfg(*it);
2011-10-13 20:53:06 +02:00
for (std::set<std::string>::const_iterator it2 = defines.begin(); it2 != defines.end(); ++it2) {
std::string::size_type pos = 0;
// Get name of define
std::string defineName(*it2);
const std::string::size_type end = defineName.find_first_of("=(");
if (end != std::string::npos)
defineName.erase(end);
// Remove ifdef configurations that match the defineName
2011-10-13 20:53:06 +02:00
while ((pos = cfg.find(defineName, pos)) != std::string::npos) {
const std::string::size_type pos1 = pos;
++pos;
if (pos1 > 0 && cfg[pos1-1] != ';')
continue;
const std::string::size_type pos2 = pos1 + defineName.length();
if (pos2 < cfg.length() && cfg[pos2] != ';')
continue;
--pos;
cfg.erase(pos, defineName.length());
}
}
2011-10-13 20:53:06 +02:00
if (cfg.length() != it->length()) {
while (cfg.length() > 0 && cfg[0] == ';')
cfg.erase(0, 1);
while (cfg.length() > 0 && cfg.back() == ';')
cfg.erase(cfg.length() - 1);
std::string::size_type pos = 0;
while ((pos = cfg.find(";;", pos)) != std::string::npos)
cfg.erase(pos, 1);
*it = cfg;
}
}
// convert configurations: "defined(A) && defined(B)" => "A;B"
2011-10-13 20:53:06 +02:00
for (std::list<std::string>::iterator it = ret.begin(); it != ret.end(); ++it) {
std::string s(*it);
2011-10-13 20:53:06 +02:00
if (s.find("&&") != std::string::npos) {
Tokenizer tokenizer(&_settings, _errorLogger);
if (!tokenizer.tokenizeCondition(s)) {
2010-04-09 21:40:37 +02:00
std::ostringstream lineStream;
lineStream << __LINE__;
ErrorLogger::ErrorMessage errmsg;
ErrorLogger::ErrorMessage::FileLocation loc;
loc.setfile(filename);
loc.line = 1;
errmsg._callStack.push_back(loc);
errmsg._severity = Severity::error;
errmsg.setmsg("Error parsing this: " + s);
2010-04-09 21:40:37 +02:00
errmsg._id = "preprocessor" + lineStream.str();
_errorLogger->reportErr(errmsg);
}
const Token *tok = tokenizer.tokens();
2012-01-02 15:52:19 +01:00
std::set<std::string> varList;
2011-10-13 20:53:06 +02:00
while (tok) {
if (Token::Match(tok, "defined ( %name% )")) {
2012-01-02 15:52:19 +01:00
varList.insert(tok->strAt(2));
tok = tok->tokAt(4);
2011-10-13 20:53:06 +02:00
if (tok && tok->str() == "&&") {
tok = tok->next();
}
} else if (Token::Match(tok, "%name% ;")) {
2012-01-02 15:52:19 +01:00
varList.insert(tok->str());
tok = tok->tokAt(2);
2011-10-13 20:53:06 +02:00
} else {
break;
}
}
s = join(varList, ';');
if (!s.empty())
*it = s;
}
}
// Convert configurations into a canonical form: B;C;A or C;A;B => A;B;C
2012-01-02 15:52:19 +01:00
for (std::list<std::string>::iterator it = ret.begin(); it != ret.end(); ++it)
*it = unify(*it, ';');
// Remove duplicates from the ret list..
ret.sort();
ret.unique();
// cleanup unhandled configurations..
2011-10-13 20:53:06 +02:00
for (std::list<std::string>::iterator it = ret.begin(); it != ret.end();) {
const std::string s(*it + ";");
bool unhandled = false;
2011-10-13 20:53:06 +02:00
for (std::string::size_type pos = 0; pos < s.length(); ++pos) {
const unsigned char c = static_cast<unsigned char>(s[pos]);
// ok with ";"
if (c == ';')
continue;
// identifier..
2011-10-13 20:53:06 +02:00
if (std::isalpha(c) || c == '_') {
while (std::isalnum((unsigned char)s[pos]) || s[pos] == '_')
++pos;
2011-10-13 20:53:06 +02:00
if (s[pos] == '=') {
++pos;
while (std::isdigit((unsigned char)s[pos]))
++pos;
2011-10-13 20:53:06 +02:00
if (s[pos] != ';') {
unhandled = true;
break;
}
}
--pos;
continue;
}
// not ok..
2011-10-13 20:53:06 +02:00
else {
unhandled = true;
break;
}
}
2011-10-13 20:53:06 +02:00
if (unhandled) {
// unhandled ifdef configuration..
if (_errorLogger && printDebug) {
std::list<ErrorLogger::ErrorMessage::FileLocation> locationList;
const ErrorLogger::ErrorMessage errmsg(locationList, Severity::debug, "unhandled configuration: " + *it, "debug", false);
_errorLogger->reportErr(errmsg);
}
ret.erase(it++);
2011-10-13 20:53:06 +02:00
} else {
++it;
}
}
return ret;
}
void Preprocessor::simplifyCondition(const std::map<std::string, std::string> &cfg, std::string &condition, bool match)
{
Tokenizer tokenizer(&_settings, _errorLogger);
if (!tokenizer.tokenizeCondition("(" + condition + ")")) {
// If tokenize returns false, then there is syntax error in the
// code which we can't handle. So stop here.
return;
}
if (Token::Match(tokenizer.tokens(), "( %name% )")) {
std::map<std::string,std::string>::const_iterator var = cfg.find(tokenizer.tokens()->strAt(1));
2011-10-13 20:53:06 +02:00
if (var != cfg.end()) {
2015-12-24 08:21:48 +01:00
const std::string &value = var->second;
condition = (value == "0") ? "0" : "1";
2011-10-13 20:53:06 +02:00
} else if (match)
condition = "0";
return;
}
if (Token::Match(tokenizer.tokens(), "( ! %name% )")) {
std::map<std::string,std::string>::const_iterator var = cfg.find(tokenizer.tokens()->strAt(2));
if (var == cfg.end())
condition = "1";
else if (var->second == "0")
condition = "1";
else if (match)
condition = "0";
return;
}
// replace variable names with values..
2011-10-13 20:53:06 +02:00
for (Token *tok = const_cast<Token *>(tokenizer.tokens()); tok; tok = tok->next()) {
if (!tok->isName())
continue;
if (Token::Match(tok, "defined ( %name% )")) {
if (cfg.find(tok->strAt(2)) != cfg.end())
tok->str("1");
else if (match)
tok->str("0");
else
continue;
tok->deleteNext(3);
continue;
}
if (Token::Match(tok, "defined %name%")) {
if (cfg.find(tok->strAt(1)) != cfg.end())
tok->str("1");
else if (match)
tok->str("0");
else
continue;
tok->deleteNext();
continue;
}
const std::map<std::string, std::string>::const_iterator it = cfg.find(tok->str());
2011-10-13 20:53:06 +02:00
if (it != cfg.end()) {
if (!it->second.empty()) {
// Tokenize the value
Tokenizer tokenizer2(&_settings, _errorLogger);
tokenizer2.tokenizeCondition(it->second);
2011-02-20 21:00:03 +01:00
// Copy the value tokens
std::stack<Token *> link;
2011-10-13 20:53:06 +02:00
for (const Token *tok2 = tokenizer2.tokens(); tok2; tok2 = tok2->next()) {
tok->str(tok2->str());
if (Token::Match(tok2,"[{([]"))
link.push(tok);
2011-10-13 20:53:06 +02:00
else if (!link.empty() && Token::Match(tok2,"[})]]")) {
Token::createMutualLinks(link.top(), tok);
link.pop();
}
2011-10-13 20:53:06 +02:00
if (tok2->next()) {
tok->insertToken("");
tok = tok->next();
}
}
2011-10-13 20:53:06 +02:00
} else if ((!tok->previous() || Token::Match(tok->previous(), "&&|%oror%|(")) &&
(!tok->next() || Token::Match(tok->next(), "&&|%oror%|)")))
tok->str("1");
else
tok->deleteThis();
2009-11-28 11:47:44 +01:00
}
}
// simplify calculations..
2012-12-01 02:02:45 +01:00
tokenizer.concatenateNegativeNumberAndAnyPositive();
2009-12-11 22:32:44 +01:00
bool modified = true;
2011-10-13 20:53:06 +02:00
while (modified) {
2009-12-11 22:32:44 +01:00
modified = false;
modified |= tokenizer.simplifySizeof();
2011-07-18 21:44:23 +02:00
modified |= tokenizer.simplifyCalculations();
2013-06-19 21:29:39 +02:00
modified |= tokenizer.simplifyConstTernaryOp();
modified |= tokenizer.simplifyRedundantParentheses();
2011-10-13 20:53:06 +02:00
for (Token *tok = const_cast<Token *>(tokenizer.tokens()); tok; tok = tok->next()) {
if (Token::Match(tok, "! %num%")) {
2009-12-11 22:32:44 +01:00
tok->deleteThis();
tok->str(tok->str() == "0" ? "1" : "0");
modified = true;
}
}
}
2011-10-13 20:53:06 +02:00
for (Token *tok = const_cast<Token *>(tokenizer.tokens()); tok; tok = tok->next()) {
if (Token::Match(tok, "(|%oror%|&& %num% &&|%oror%|)")) {
if (tok->next()->str() != "0") {
tok->next()->str("1");
}
}
}
2011-10-13 20:53:06 +02:00
for (Token *tok = const_cast<Token *>(tokenizer.tokens()); tok; tok = tok->next()) {
while (Token::Match(tok, "(|%oror% %any% %oror% 1")) {
tok->deleteNext(2);
if (tok->tokAt(-3))
tok = tok->tokAt(-3);
}
}
if (Token::simpleMatch(tokenizer.tokens(), "( 1 )") ||
Token::simpleMatch(tokenizer.tokens(), "( 1 ||"))
condition = "1";
else if (Token::simpleMatch(tokenizer.tokens(), "( 0 )"))
condition = "0";
}
bool Preprocessor::match_cfg_def(std::map<std::string, std::string> cfg, std::string def)
{
2011-07-16 09:32:35 +02:00
/*
std::cout << "cfg: \"";
for (std::map<std::string, std::string>::const_iterator it = cfg.begin(); it != cfg.end(); ++it)
{
std::cout << it->first;
if (!it->second.empty())
std::cout << "=" << it->second;
std::cout << ";";
}
std::cout << "\" ";
std::cout << "def: \"" << def << "\"\n";
*/
2015-05-12 14:00:43 +02:00
simplifyVarMap(cfg, _settings);
simplifyCondition(cfg, def, true);
if (cfg.find(def) != cfg.end())
return true;
if (def == "0")
return false;
if (def == "1")
return true;
return false;
}
std::string Preprocessor::getcode(const std::string &filedata, const std::string &cfg, const std::string &filename)
{
// For the error report and preprocessor dump:
// line number relative to current (included) file
// (may decrease when popping back from an included file)
unsigned int lineno = 0;
std::ostringstream ret;
bool match = true;
std::list<bool> matching_ifdef;
std::list<bool> matched_ifdef;
// Create a map for the cfg for faster access to defines
std::map<std::string, std::string> cfgmap(getcfgmap(cfg, &_settings, filename));
std::stack<std::string> filenames;
filenames.push(filename);
std::stack<unsigned int> lineNumbers;
std::istringstream istr(filedata);
std::string line;
directives.clear();
while (std::getline(istr, line)) {
++lineno;
if (line.empty()) {
ret << '\n';
continue;
}
// record directive for addons / checkers
if ((line[0] == '#')
&& (line.compare(0, 6, "#line ") != 0)
&& (line.compare(0, 8, "#endfile") != 0)) {
// for clarity, turn "#file ..." back into "#include ..."
std::string orig_line = line;
if (orig_line.compare(0, 6, "#file ")==0)
orig_line.replace(1, 4, "include");
// record directive and extra
directives.push_back(Directive(filenames.top(), lineno,
orig_line));
}
if (_settings.terminated())
return "";
2011-10-13 20:53:06 +02:00
if (line.compare(0, 11, "#pragma asm") == 0) {
ret << "\n";
bool found_end = false;
2011-10-13 20:53:06 +02:00
while (getline(istr, line)) {
if (line.compare(0, 14, "#pragma endasm") == 0) {
found_end = true;
break;
}
ret << "\n";
}
if (!found_end)
break;
if (line.find('=') != std::string::npos) {
Tokenizer tokenizer(&_settings, _errorLogger);
line.erase(0, sizeof("#pragma endasm"));
2011-12-08 21:28:34 +01:00
std::istringstream tempIstr(line);
tokenizer.tokenize(tempIstr, "", "", true);
if (Token::Match(tokenizer.tokens(), "( %name% = %any% )")) {
ret << "asm(" << tokenizer.tokens()->strAt(1) << ");";
}
}
ret << "\n";
continue;
}
const std::string def = getdef(line, true);
const std::string ndef = getdef(line, false);
const bool emptymatch = matching_ifdef.empty() || matched_ifdef.empty();
2011-10-13 20:53:06 +02:00
if (line.compare(0, 8, "#define ") == 0) {
match = true;
typedef std::set<std::string>::const_iterator It;
for (It it = _settings.userUndefs.begin(); it != _settings.userUndefs.end(); ++it) {
const std::string::size_type symbolPos = line.find_first_not_of(' ', 8);
if (symbolPos == std::string::npos)
continue;
const std::string::size_type undefMatchPos = line.find(*it, symbolPos);
if (undefMatchPos == std::string::npos)
continue;
const std::string::size_type behindUndefPos = undefMatchPos + (*it).size();
if ((line.size() == behindUndefPos) ||
(line[behindUndefPos] == ' ') ||
(line[behindUndefPos] == '(')) {
match = false;
break;
}
}
if (match) {
for (std::list<bool>::const_iterator it = matching_ifdef.begin(); it != matching_ifdef.end(); ++it) {
if (!bool(*it)) {
match = false;
break;
}
}
}
2011-10-13 20:53:06 +02:00
if (match) {
const std::string::size_type pos = line.find_first_of(" (", 8);
if (pos == std::string::npos)
cfgmap[line.substr(8)] = "";
2011-10-13 20:53:06 +02:00
else if (line[pos] == ' ') {
std::string value(line.substr(pos + 1));
2015-12-03 09:03:18 +01:00
std::map<std::string, std::string>::const_iterator cfgpos = cfgmap.find(value);
if (cfgpos != cfgmap.end())
value = cfgpos->second;
cfgmap[line.substr(8, pos - 8)] = value;
2011-10-13 20:53:06 +02:00
} else
cfgmap[line.substr(8, pos - 8)] = "";
}
}
else if (line.compare(0, 7, "#undef ") == 0) {
const std::string name(line.substr(7));
cfgmap.erase(name);
}
2011-10-13 20:53:06 +02:00
else if (!emptymatch && line.compare(0, 7, "#elif !") == 0) {
if (matched_ifdef.back()) {
2011-01-31 00:33:44 +01:00
matching_ifdef.back() = false;
2011-10-13 20:53:06 +02:00
} else {
if (!match_cfg_def(cfgmap, ndef)) {
2011-01-31 00:33:44 +01:00
matching_ifdef.back() = true;
matched_ifdef.back() = true;
}
}
}
2011-10-13 20:53:06 +02:00
else if (!emptymatch && line.compare(0, 6, "#elif ") == 0) {
if (matched_ifdef.back()) {
matching_ifdef.back() = false;
2011-10-13 20:53:06 +02:00
} else {
if (match_cfg_def(cfgmap, def)) {
matching_ifdef.back() = true;
matched_ifdef.back() = true;
}
}
}
2013-06-19 21:29:39 +02:00
else if (line.compare(0,4,"#if ") == 0) {
matching_ifdef.push_back(match_cfg_def(cfgmap, line.substr(4)));
matched_ifdef.push_back(matching_ifdef.back());
}
2011-10-13 20:53:06 +02:00
else if (! def.empty()) {
2013-06-19 21:29:39 +02:00
matching_ifdef.push_back(cfgmap.find(def) != cfgmap.end());
matched_ifdef.push_back(matching_ifdef.back());
}
2011-10-13 20:53:06 +02:00
else if (! ndef.empty()) {
2013-06-19 21:29:39 +02:00
matching_ifdef.push_back(cfgmap.find(ndef) == cfgmap.end());
matched_ifdef.push_back(matching_ifdef.back());
}
2011-10-13 20:53:06 +02:00
else if (!emptymatch && line == "#else") {
if (! matched_ifdef.empty())
matching_ifdef.back() = ! matched_ifdef.back();
}
2011-10-13 20:53:06 +02:00
else if (line.compare(0, 6, "#endif") == 0) {
if (! matched_ifdef.empty())
matched_ifdef.pop_back();
if (! matching_ifdef.empty())
matching_ifdef.pop_back();
}
2011-10-13 20:53:06 +02:00
if (!line.empty() && line[0] == '#') {
match = true;
for (std::list<bool>::const_iterator it = matching_ifdef.begin(); it != matching_ifdef.end(); ++it) {
if (!bool(*it)) {
match = false;
break;
}
}
}
// #error => return ""
2011-10-13 20:53:06 +02:00
if (match && line.compare(0, 6, "#error") == 0) {
if (!_settings.userDefines.empty() && !_settings.force) {
error(Path::simplifyPath(filenames.top()), lineno, line);
}
return "";
}
if (!match && (line.compare(0, 8, "#define ") == 0 ||
2011-10-13 20:53:06 +02:00
line.compare(0, 6, "#undef") == 0)) {
// Remove define that is not part of this configuration
line = "";
2011-10-13 20:53:06 +02:00
} else if (line.compare(0, 7, "#file \"") == 0 ||
line.compare(0, 8, "#endfile") == 0 ||
line.compare(0, 8, "#define ") == 0 ||
line.compare(0, 6, "#line ") == 0 ||
2011-10-13 20:53:06 +02:00
line.compare(0, 6, "#undef") == 0) {
// We must not remove #file tags or line numbers
// are corrupted. File tags are removed by the tokenizer.
// Keep location info updated
2011-10-13 20:53:06 +02:00
if (line.compare(0, 7, "#file \"") == 0) {
filenames.push(line.substr(7, line.size() - 8));
lineNumbers.push(lineno);
lineno = 0;
2011-10-13 20:53:06 +02:00
} else if (line.compare(0, 8, "#endfile") == 0) {
if (filenames.size() > 1U)
filenames.pop();
2011-10-13 20:53:06 +02:00
if (!lineNumbers.empty()) {
lineno = lineNumbers.top();
lineNumbers.pop();
}
}
2011-10-13 20:53:06 +02:00
} else if (!match || line.compare(0, 1, "#") == 0) {
// Remove #if, #else, #pragma etc, leaving only
// #define, #undef, #file and #endfile. and also lines
// which are not part of this configuration.
line = "";
}
ret << line << "\n";
}
if (!validateCfg(ret.str(), cfg)) {
return "";
}
return expandMacros(ret.str(), filename, cfg, _errorLogger);
}
void Preprocessor::error(const std::string &filename, unsigned int linenr, const std::string &msg)
{
std::list<ErrorLogger::ErrorMessage::FileLocation> locationList;
2011-10-13 20:53:06 +02:00
if (!filename.empty()) {
2015-10-14 11:57:20 +02:00
ErrorLogger::ErrorMessage::FileLocation loc(filename, linenr);
locationList.push_back(loc);
}
_errorLogger->reportErr(ErrorLogger::ErrorMessage(locationList,
Severity::error,
msg,
"preprocessorErrorDirective",
false));
}
2010-07-21 11:50:04 +02:00
Preprocessor::HeaderTypes Preprocessor::getHeaderFileName(std::string &str)
{
std::string::size_type i = str.find_first_of("<\"");
2011-10-13 20:53:06 +02:00
if (i == std::string::npos) {
str = "";
2010-07-21 11:50:04 +02:00
return NoHeader;
}
char c = str[i];
if (c == '<')
c = '>';
std::string result;
2011-10-13 20:53:06 +02:00
for (i = i + 1; i < str.length(); ++i) {
if (str[i] == c)
break;
result.append(1, str[i]);
}
// Linux can't open include paths with \ separator, so fix them
std::replace(result.begin(), result.end(), '\\', '/');
str = result;
return (c == '\"') ? UserHeader : SystemHeader;
}
/**
* Try to open header
* @param filename header name (in/out)
* @param includePaths paths where to look for the file
2012-10-24 01:20:14 +02:00
* @param filePath path to the header file
* @param fin file input stream (in/out)
* @return if file is opened then true is returned
*/
static bool openHeader(std::string &filename, const std::list<std::string> &includePaths, const std::string &filePath, std::ifstream &fin)
{
fin.open((filePath + filename).c_str());
if (fin.is_open()) {
filename = filePath + filename;
return true;
}
fin.open(filename.c_str());
if (fin.is_open())
return true;
for (std::list<std::string>::const_iterator iter = includePaths.begin(); iter != includePaths.end(); ++iter) {
const std::string nativePath(Path::toNativeSeparators(*iter));
fin.open((nativePath + filename).c_str());
if (fin.is_open()) {
filename = nativePath + filename;
return true;
}
fin.clear();
}
return false;
}
std::string Preprocessor::handleIncludes(const std::string &code, const std::string &filePath, const std::list<std::string> &includePaths, std::map<std::string,std::string> &defs, std::set<std::string> &pragmaOnce, std::list<std::string> includes)
{
std::string path;
std::string::size_type sep_pos = filePath.find_last_of("\\/");
if (sep_pos != std::string::npos)
path = filePath.substr(0, 1 + sep_pos);
// current #if indent level.
std::stack<bool>::size_type indent = 0;
// how deep does the #if match? this can never be bigger than "indent".
std::stack<bool>::size_type indentmatch = 0;
// has there been a true #if condition at the current indentmatch level?
// then no more #elif or #else can be true before the #endif is seen.
std::stack<bool> elseIsTrueStack;
unsigned int linenr = 0;
const std::set<std::string> &undefs = _settings.userUndefs;
2013-01-26 19:11:58 +01:00
if (_errorLogger)
_errorLogger->reportProgress(filePath, "Preprocessor (handleIncludes)", 0);
std::ostringstream ostr;
std::istringstream istr(code);
std::string line;
bool suppressCurrentCodePath = false;
while (std::getline(istr,line)) {
++linenr;
if (_settings.terminated())
return "";
// has there been a true #if condition at the current indentmatch level?
// then no more #elif or #else can be true before the #endif is seen.
while (elseIsTrueStack.size() != indentmatch + 1) {
if (elseIsTrueStack.size() < indentmatch + 1) {
elseIsTrueStack.push(true);
} else {
elseIsTrueStack.pop();
}
}
if (elseIsTrueStack.empty()) {
2013-10-19 09:52:19 +02:00
writeError(filePath, linenr, _errorLogger, "syntaxError", "Syntax error in preprocessor code");
return "";
}
std::stack<bool>::reference elseIsTrue = elseIsTrueStack.top();
if (line == "#pragma once") {
pragmaOnce.insert(filePath);
} else if (line.compare(0,7,"#ifdef ") == 0) {
if (indent == indentmatch) {
const std::string tag = getdef(line,true);
if (defs.find(tag) != defs.end()) {
elseIsTrue = false;
indentmatch++;
} else if (undefs.find(tag) != undefs.end()) {
elseIsTrue = true;
indentmatch++;
suppressCurrentCodePath = true;
}
}
++indent;
if (indent == indentmatch + 1)
elseIsTrue = true;
} else if (line.compare(0,8,"#ifndef ") == 0) {
if (indent == indentmatch) {
const std::string tag = getdef(line,false);
if (defs.find(tag) == defs.end()) {
elseIsTrue = false;
indentmatch++;
} else if (undefs.find(tag) != undefs.end()) {
elseIsTrue = false;
indentmatch++;
suppressCurrentCodePath = false;
}
}
++indent;
if (indent == indentmatch + 1)
elseIsTrue = true;
} else if (line.compare(0,4,"#if ") == 0) {
if (!suppressCurrentCodePath && indent == indentmatch && match_cfg_def(defs, line.substr(4))) {
elseIsTrue = false;
indentmatch++;
}
++indent;
if (indent == indentmatch + 1)
elseIsTrue = true; // this value doesn't matter when suppressCurrentCodePath is true
} else if (line.compare(0,6,"#elif ") == 0 || line.compare(0,5,"#else") == 0) {
if (!elseIsTrue) {
if ((indentmatch > 0) && (indentmatch == indent)) {
indentmatch = indent - 1;
}
} else {
if ((indentmatch > 0) && (indentmatch == indent)) {
indentmatch = indent - 1;
} else if ((indent > 0) && indentmatch == indent - 1) {
if (line.compare(0,5,"#else")==0 || match_cfg_def(defs,line.substr(6))) {
indentmatch = indent;
elseIsTrue = false;
}
}
}
} else if (line.compare(0, 6, "#endif") == 0) {
if (indent > 0)
--indent;
if (indentmatch > indent || indent == 0) {
indentmatch = indent;
elseIsTrue = false;
suppressCurrentCodePath = false;
}
} else if (indentmatch == indent) {
if (!suppressCurrentCodePath && line.compare(0, 8, "#define ") == 0) {
const unsigned int endOfDefine = 8;
std::string::size_type endOfTag = line.find_first_of("( ", endOfDefine);
std::string tag;
// define a symbol
if (endOfTag == std::string::npos) {
tag = line.substr(endOfDefine);
defs[tag] = "";
} else {
tag = line.substr(endOfDefine, endOfTag-endOfDefine);
// define a function-macro
if (line[endOfTag] == '(') {
defs[tag] = "";
}
// define value
else {
++endOfTag;
const std::string& value = line.substr(endOfTag, line.size()-endOfTag);
if (defs.find(value) != defs.end())
defs[tag] = defs[value];
else
defs[tag] = value;
}
}
if (undefs.find(tag) != undefs.end()) {
defs.erase(tag);
}
}
else if (!suppressCurrentCodePath && line.compare(0,7,"#undef ") == 0) {
defs.erase(line.substr(7));
}
else if (!suppressCurrentCodePath && line.compare(0,9,"#include ")==0) {
2011-11-03 19:05:48 +01:00
std::string filename(line.substr(9));
const HeaderTypes headerType = getHeaderFileName(filename);
if (headerType == NoHeader) {
ostr << std::endl;
continue;
}
// try to open file
2011-12-08 22:17:50 +01:00
std::string filepath;
if (headerType == UserHeader)
filepath = path;
2011-11-03 19:05:48 +01:00
std::ifstream fin;
2011-12-08 22:17:50 +01:00
if (!openHeader(filename, includePaths, filepath, fin)) {
missingInclude(Path::toNativeSeparators(filePath),
linenr,
filename,
headerType
);
2011-11-03 19:05:48 +01:00
ostr << std::endl;
continue;
}
// Prevent that files are recursively included
if (std::find(includes.begin(), includes.end(), filename) != includes.end()) {
ostr << std::endl;
continue;
}
// #6913 - simplify Path to avoid strange recursion
includes.push_back(Path::simplifyPath(filename));
2011-11-03 19:05:48 +01:00
// Don't include header if it's already included and contains #pragma once
if (pragmaOnce.find(filename) != pragmaOnce.end()) {
ostr << std::endl;
continue;
}
2011-11-03 19:05:48 +01:00
ostr << "#file \"" << filename << "\"\n"
<< handleIncludes(read(fin, filename), filename, includePaths, defs, pragmaOnce, includes) << std::endl
2011-11-03 19:05:48 +01:00
<< "#endfile\n";
continue;
}
if (!suppressCurrentCodePath)
ostr << line;
}
// A line has been read..
ostr << "\n";
}
return ostr.str();
}
void Preprocessor::handleIncludes(std::string &code, const std::string &filePath, const std::list<std::string> &includePaths)
{
std::list<std::string> paths;
std::string path = filePath;
const std::string::size_type sep_pos = path.find_last_of("\\/");
if (sep_pos != std::string::npos)
path.erase(1 + sep_pos);
paths.push_back(path);
std::string::size_type pos = 0;
std::string::size_type endfilePos = 0;
if (code.compare(0,7U,"#file \"")==0) {
const std::string::size_type start = code.find("#file \"" + filePath, 7U);
if (start != std::string::npos)
endfilePos = start;
}
std::set<std::string> handledFiles;
2011-10-13 20:53:06 +02:00
while ((pos = code.find("#include", pos)) != std::string::npos) {
if (_settings.terminated())
return;
// Accept only includes that are at the start of a line
2011-10-13 20:53:06 +02:00
if (pos > 0 && code[pos-1] != '\n') {
pos += 8; // length of "#include"
continue;
}
// If endfile is encountered, we have moved to a next file in our stack,
// so remove last path in our list.
while (!paths.empty() && (endfilePos = code.find("\n#endfile", endfilePos)) != std::string::npos && endfilePos < pos) {
paths.pop_back();
endfilePos += 9; // size of #endfile
}
endfilePos = pos;
std::string::size_type end = code.find('\n', pos);
std::string filename = code.substr(pos, end - pos);
// Remove #include clause
code.erase(pos, end - pos);
2010-07-21 11:50:04 +02:00
HeaderTypes headerType = getHeaderFileName(filename);
if (headerType == NoHeader)
continue;
// filename contains now a file name e.g. "menu.h"
std::string processedFile;
2011-12-08 22:17:50 +01:00
std::string filepath;
if (headerType == UserHeader && !paths.empty())
2011-12-08 22:17:50 +01:00
filepath = paths.back();
std::ifstream fin;
2011-12-08 22:17:50 +01:00
const bool fileOpened(openHeader(filename, includePaths, filepath, fin));
2011-10-13 20:53:06 +02:00
if (fileOpened) {
filename = Path::simplifyPath(filename);
std::string tempFile = filename;
std::transform(tempFile.begin(), tempFile.end(), tempFile.begin(), tolowerWrapper);
if (handledFiles.find(tempFile) != handledFiles.end()) {
// We have processed this file already once, skip
// it this time to avoid eternal loop.
fin.close();
continue;
}
handledFiles.insert(tempFile);
2012-01-05 18:37:15 +01:00
processedFile = Preprocessor::read(fin, filename);
fin.close();
}
2011-10-13 20:53:06 +02:00
if (!processedFile.empty()) {
// Remove space characters that are after or before new line character
processedFile = "#file \"" + Path::fromNativeSeparators(filename) + "\"\n" + processedFile + "\n#endfile";
code.insert(pos, processedFile);
path = filename;
path.erase(1 + path.find_last_of("\\/"));
paths.push_back(path);
} else if (!fileOpened) {
std::string f = filePath;
// Determine line number of include
unsigned int linenr = 1;
unsigned int level = 0;
for (std::string::size_type p = 1; p <= pos; ++p) {
if (level == 0 && code[pos-p] == '\n')
++linenr;
else if (code.compare(pos-p, 9, "#endfile\n") == 0) {
++level;
} else if (code.compare(pos-p, 6, "#file ") == 0) {
if (level == 0) {
linenr--;
const std::string::size_type pos1 = pos - p + 7;
const std::string::size_type pos2 = code.find_first_of("\"\n", pos1);
f = code.substr(pos1, (pos2 == std::string::npos) ? pos2 : (pos2 - pos1));
break;
2010-09-03 13:30:49 +02:00
}
--level;
2010-09-03 13:30:49 +02:00
}
}
2010-09-03 13:30:49 +02:00
missingInclude(Path::toNativeSeparators(f),
linenr,
filename,
headerType);
}
}
}
// Report that include is missing
void Preprocessor::missingInclude(const std::string &filename, unsigned int linenr, const std::string &header, HeaderTypes headerType)
{
const std::string fname = Path::fromNativeSeparators(filename);
if (_settings.nomsg.isSuppressed("missingInclude", fname, linenr))
return;
if (headerType == SystemHeader && _settings.nomsg.isSuppressed("missingIncludeSystem", fname, linenr))
return;
if (headerType == SystemHeader)
missingSystemIncludeFlag = true;
else
missingIncludeFlag = true;
if (_errorLogger && _settings.checkConfiguration) {
std::list<ErrorLogger::ErrorMessage::FileLocation> locationList;
if (!filename.empty()) {
ErrorLogger::ErrorMessage::FileLocation loc;
loc.line = linenr;
loc.setfile(Path::toNativeSeparators(filename));
locationList.push_back(loc);
}
ErrorLogger::ErrorMessage errmsg(locationList, Severity::information,
(headerType==SystemHeader) ?
"Include file: <" + header + "> not found. Please note: Cppcheck does not need standard library headers to get proper results." :
"Include file: \"" + header + "\" not found.",
(headerType==SystemHeader) ? "missingIncludeSystem" : "missingInclude",
false);
errmsg.file0 = file0;
_errorLogger->reportInfo(errmsg);
}
}
/**
* Skip string in line. A string begins and ends with either a &quot; or a &apos;
* @param line the string
* @param pos in=start position of string, out=end position of string
*/
static void skipstring(const std::string &line, std::string::size_type &pos)
{
const char ch = line[pos];
++pos;
2011-10-13 20:53:06 +02:00
while (pos < line.size() && line[pos] != ch) {
if (line[pos] == '\\')
++pos;
++pos;
}
}
/**
* @brief get parameters from code. For example 'foo(1,2)' => '1','2'
* @param line in: The code
* @param pos in: Position to the '('. out: Position to the ')'
* @param params out: The extracted parameters
* @param numberOfNewlines out: number of newlines in the macro call
* @param endFound out: was the end parentheses found?
*/
static void getparams(const std::string &line,
std::string::size_type &pos,
std::vector<std::string> &params,
unsigned int &numberOfNewlines,
bool &endFound)
{
params.clear();
numberOfNewlines = 0;
endFound = false;
if (line[pos] == ' ')
pos++;
if (line[pos] != '(')
return;
2010-12-15 18:45:53 +01:00
// parentheses level
int parlevel = 0;
// current parameter data
std::string par;
// scan for parameters..
2011-10-13 20:53:06 +02:00
for (; pos < line.length(); ++pos) {
// increase parentheses level
2011-10-13 20:53:06 +02:00
if (line[pos] == '(') {
++parlevel;
if (parlevel == 1)
continue;
}
// decrease parentheses level
2011-10-13 20:53:06 +02:00
else if (line[pos] == ')') {
--parlevel;
2011-10-13 20:53:06 +02:00
if (parlevel <= 0) {
endFound = true;
2013-08-10 23:28:02 +02:00
params.push_back(trim(par));
break;
}
}
// string
2011-10-13 20:53:06 +02:00
else if (line[pos] == '\"' || line[pos] == '\'') {
const std::string::size_type p = pos;
skipstring(line, pos);
if (pos == line.length())
break;
par += line.substr(p, pos + 1 - p);
continue;
}
// count newlines. the expanded macro must have the same number of newlines
2011-10-13 20:53:06 +02:00
else if (line[pos] == '\n') {
++numberOfNewlines;
continue;
}
// new parameter
2011-10-13 20:53:06 +02:00
if (parlevel == 1 && line[pos] == ',') {
2013-08-10 23:28:02 +02:00
params.push_back(trim(par));
par = "";
}
// spaces are only added if needed
2011-10-13 20:53:06 +02:00
else if (line[pos] == ' ') {
// Add space only if it is needed
if (par.size() && std::isalnum((unsigned char)par.back())) {
par += ' ';
}
}
// add character to current parameter
else if (parlevel >= 1 && line[pos] != Preprocessor::macroChar) {
par.append(1, line[pos]);
}
}
}
/** @brief Class that the preprocessor uses when it expands macros. This class represents a preprocessor macro */
2011-10-13 20:53:06 +02:00
class PreprocessorMacro {
private:
2010-01-13 21:50:44 +01:00
/** tokens of this macro */
TokenList tokenlist;
2010-01-13 21:50:44 +01:00
/** macro parameters */
std::vector<std::string> _params;
2010-01-13 21:50:44 +01:00
/** macro definition in plain text */
const std::string _macro;
/** does this macro take a variable number of parameters? */
bool _variadic;
2010-12-15 18:45:53 +01:00
/** The macro has parentheses but no parameters.. "AAA()" */
bool _nopar;
/** disabled assignment operator */
void operator=(const PreprocessorMacro &);
/** @brief expand inner macro */
std::vector<std::string> expandInnerMacros(const std::vector<std::string> &params1,
2014-11-20 14:20:09 +01:00
const std::map<std::string, PreprocessorMacro *> &macros) const {
std::string innerMacroName;
// Is there an inner macro..
{
const Token *tok = Token::findsimplematch(tokens(), ")");
if (!Token::Match(tok, ") %name% ("))
return params1;
innerMacroName = tok->strAt(1);
tok = tok->tokAt(3);
unsigned int par = 0;
while (Token::Match(tok, "%name% ,|)")) {
tok = tok->tokAt(2);
par++;
}
if (tok || par != params1.size())
return params1;
}
std::vector<std::string> params2(params1);
for (std::size_t ipar = 0; ipar < params1.size(); ++ipar) {
const std::string s(innerMacroName + "(");
const std::string param(params1[ipar]);
if (param.compare(0,s.length(),s)==0 && param.back() == ')') {
std::vector<std::string> innerparams;
std::string::size_type pos = s.length() - 1;
unsigned int num = 0;
bool endFound = false;
getparams(param, pos, innerparams, num, endFound);
2011-10-13 20:53:06 +02:00
if (pos == param.length()-1 && num==0 && endFound && innerparams.size() == params1.size()) {
// Is inner macro defined?
std::map<std::string, PreprocessorMacro *>::const_iterator it = macros.find(innerMacroName);
2011-10-13 20:53:06 +02:00
if (it != macros.end()) {
// expand the inner macro
const PreprocessorMacro *innerMacro = it->second;
std::string innercode;
std::map<std::string,PreprocessorMacro *> innermacros = macros;
innermacros.erase(innerMacroName);
innerMacro->code(innerparams, innermacros, innercode);
params2[ipar] = innercode;
}
}
}
}
return params2;
}
public:
/**
2010-03-14 07:48:44 +01:00
* @brief Constructor for PreprocessorMacro. This is the "setter"
* for this class - everything is setup here.
2015-08-30 20:12:02 +02:00
* @param [in] macro The code after define, until end of line,
* e.g. "A(x) foo(x);"
2015-08-30 20:12:02 +02:00
* @param [in] settings Current settings being used
*/
PreprocessorMacro(const std::string &macro, const Settings* settings)
: tokenlist(settings), _macro(macro) {
// Tokenize the macro to make it easier to handle
2011-12-08 21:28:34 +01:00
std::istringstream istr(macro);
tokenlist.createTokens(istr);
2010-01-13 21:50:44 +01:00
// initialize parameters to default values
_variadic = _nopar = false;
const std::string::size_type pos = macro.find_first_of(" (");
2011-10-13 20:53:06 +02:00
if (pos != std::string::npos && macro[pos] == '(') {
// Extract macro parameters
if (Token::Match(tokens(), "%name% ( %name%")) {
2011-10-13 20:53:06 +02:00
for (const Token *tok = tokens()->tokAt(2); tok; tok = tok->next()) {
if (tok->str() == ")")
break;
2011-10-13 20:53:06 +02:00
if (Token::simpleMatch(tok, ". . . )")) {
if (tok->previous()->str() == ",")
_params.push_back("__VA_ARGS__");
_variadic = true;
break;
}
if (tok->isName())
_params.push_back(tok->str());
}
}
else if (Token::Match(tokens(), "%name% ( . . . )"))
_variadic = true;
else if (Token::Match(tokens(), "%name% ( )"))
_nopar = true;
}
}
2010-01-13 21:50:44 +01:00
/** return tokens of this macro */
2014-11-20 14:20:09 +01:00
const Token *tokens() const {
return tokenlist.front();
}
2010-01-13 21:50:44 +01:00
/** read parameters of this macro */
2014-11-20 14:20:09 +01:00
const std::vector<std::string> &params() const {
return _params;
}
2010-01-13 21:50:44 +01:00
/** check if this is macro has a variable number of parameters */
2014-11-20 14:20:09 +01:00
bool variadic() const {
return _variadic;
}
2010-12-15 18:45:53 +01:00
/** Check if this macro has parentheses but no parameters */
2014-11-20 14:20:09 +01:00
bool nopar() const {
return _nopar;
}
2010-01-13 21:50:44 +01:00
/** name of macro */
2014-11-20 14:20:09 +01:00
const std::string &name() const {
return tokens() ? tokens()->str() : emptyString;
}
2010-01-13 21:50:44 +01:00
/**
* get expanded code for this macro
* @param params2 macro parameters
* @param macros macro definitions (recursion)
2010-01-13 21:50:44 +01:00
* @param macrocode output string
* @return true if the expanding was successful
*/
2014-11-20 14:20:09 +01:00
bool code(const std::vector<std::string> &params2, const std::map<std::string, PreprocessorMacro *> &macros, std::string &macrocode) const {
2011-10-13 20:53:06 +02:00
if (_nopar || (_params.empty() && _variadic)) {
macrocode = _macro.substr(1 + _macro.find(')'));
if (macrocode.empty())
return true;
std::string::size_type pos = 0;
// Remove leading spaces
if ((pos = macrocode.find_first_not_of(" ")) > 0)
macrocode.erase(0, pos);
// Remove ending newline
if ((pos = macrocode.find_first_of("\r\n")) != std::string::npos)
macrocode.erase(pos);
// Replace "__VA_ARGS__" with parameters
2011-10-13 20:53:06 +02:00
if (!_nopar) {
2010-06-20 18:41:45 +02:00
std::string s;
for (std::size_t i = 0; i < params2.size(); ++i) {
2010-06-20 18:41:45 +02:00
if (i > 0)
s += ",";
s += params2[i];
}
pos = 0;
2011-10-13 20:53:06 +02:00
while ((pos = macrocode.find("__VA_ARGS__", pos)) != std::string::npos) {
2010-06-20 18:41:45 +02:00
macrocode.erase(pos, 11);
macrocode.insert(pos, s);
pos += s.length();
}
}
}
2011-10-13 20:53:06 +02:00
else if (_params.empty()) {
std::string::size_type pos = _macro.find_first_of(" \"");
if (pos == std::string::npos)
macrocode = "";
2011-10-13 20:53:06 +02:00
else {
if (_macro[pos] == ' ')
pos++;
macrocode = _macro.substr(pos);
if ((pos = macrocode.find_first_of("\r\n")) != std::string::npos)
macrocode.erase(pos);
}
}
2011-10-13 20:53:06 +02:00
else {
const std::vector<std::string> givenparams = expandInnerMacros(params2, macros);
const Token *tok = tokens();
while (tok && tok->str() != ")")
tok = tok->next();
2011-10-13 20:53:06 +02:00
if (tok) {
bool optcomma = false;
2014-02-15 08:37:57 +01:00
while (nullptr != (tok = tok->next())) {
std::string str = tok->str();
2011-10-13 20:53:06 +02:00
if (str[0] == '#' || tok->isName()) {
if (str == "##")
continue;
const bool stringify(str[0] == '#');
2011-10-13 20:53:06 +02:00
if (stringify) {
2009-01-25 14:30:15 +01:00
str = str.erase(0, 1);
}
for (std::size_t i = 0; i < _params.size(); ++i) {
2011-10-13 20:53:06 +02:00
if (str == _params[i]) {
if (_variadic &&
(i == _params.size() - 1 ||
2011-10-13 20:53:06 +02:00
(givenparams.size() + 2 == _params.size() && i + 1 == _params.size() - 1))) {
str = "";
for (std::size_t j = _params.size() - 1; j < givenparams.size(); ++j) {
if (optcomma || j > _params.size() - 1)
str += ",";
optcomma = false;
str += givenparams[j];
}
2011-10-13 20:53:06 +02:00
} else if (i >= givenparams.size()) {
// Macro had more parameters than caller used.
macrocode = "";
return false;
2011-10-13 20:53:06 +02:00
} else if (stringify) {
const std::string &s(givenparams[i]);
std::ostringstream ostr;
ostr << "\"";
2011-10-13 20:53:06 +02:00
for (std::string::size_type j = 0; j < s.size(); ++j) {
2010-04-09 21:40:37 +02:00
if (s[j] == '\\' || s[j] == '\"')
ostr << '\\';
2010-04-09 21:40:37 +02:00
ostr << s[j];
}
str = ostr.str() + "\"";
2011-10-13 20:53:06 +02:00
} else
str = givenparams[i];
break;
}
}
// expand nopar macro
2011-10-13 20:53:06 +02:00
if (tok->strAt(-1) != "##") {
const std::map<std::string, PreprocessorMacro *>::const_iterator it = macros.find(str);
if (it != macros.end() && it->second->_macro.find('(') == std::string::npos) {
const std::string& macro = it->second->_macro;
const std::string::size_type whitespacePos = macro.find(' ');
2015-12-23 14:10:39 +01:00
if (whitespacePos != std::string::npos)
str = macro.substr(whitespacePos);
else
str = "";
}
}
}
2011-10-13 20:53:06 +02:00
if (_variadic && tok->str() == "," && tok->next() && tok->next()->str() == "##") {
optcomma = true;
continue;
}
optcomma = false;
// separate ++ and -- with space
if (!str.empty() && !macrocode.empty() && (str[0] == '+' || str[0] == '-') && str[0] == macrocode[macrocode.size()-1U])
macrocode += ' ';
macrocode += str;
2015-08-11 11:19:59 +02:00
if (Token::Match(tok, "%name% %name%|%num%") ||
Token::Match(tok, "%num% %name%") ||
Token::simpleMatch(tok, "> >"))
macrocode += ' ';
}
}
}
return true;
}
};
2010-01-13 21:50:44 +01:00
/**
* Get data from a input string. This is an extended version of std::getline.
* The std::getline only get a single line at a time. It can therefore happen that it
* contains a partial statement. This function ensures that the returned data
* doesn't end in the middle of a statement. The "getlines" name indicate that
* this function will return multiple lines if needed.
* @param istr input stream
* @param line output data
* @return success
*/
static bool getlines(std::istream &istr, std::string &line)
{
if (!istr.good())
return false;
line = "";
int parlevel = 0;
bool directive = false;
2011-10-13 20:53:06 +02:00
for (char ch = (char)istr.get(); istr.good(); ch = (char)istr.get()) {
if (ch == '\'' || ch == '\"') {
line += ch;
char c = 0;
2011-10-13 20:53:06 +02:00
while (istr.good() && c != ch) {
if (c == '\\') {
c = (char)istr.get();
if (!istr.good())
return true;
line += c;
}
2009-12-06 20:10:10 +01:00
c = (char)istr.get();
if (!istr.good())
return true;
if (c == '\n' && directive)
return true;
line += c;
}
2009-12-06 20:10:10 +01:00
continue;
}
if (ch == '(')
++parlevel;
else if (ch == ')')
--parlevel;
2011-10-13 20:53:06 +02:00
else if (ch == '\n') {
if (directive)
return true;
2011-10-13 20:53:06 +02:00
if (istr.peek() == '#') {
line += ch;
return true;
}
} else if (!directive && parlevel <= 0 && ch == ';') {
line += ";";
return true;
2009-12-06 20:10:10 +01:00
}
if (ch == '#' && line.empty())
directive = true;
line += ch;
}
return true;
}
bool Preprocessor::validateCfg(const std::string &code, const std::string &cfg)
{
const bool printInformation = _settings.isEnabled("information");
// fill up "macros" with empty configuration macros
std::set<std::string> macros;
for (std::string::size_type pos = 0; pos < cfg.size();) {
const std::string::size_type pos2 = cfg.find_first_of(";=", pos);
if (pos2 == std::string::npos) {
macros.insert(cfg.substr(pos));
break;
}
if (cfg[pos2] == ';')
macros.insert(cfg.substr(pos, pos2-pos));
pos = cfg.find(';', pos2);
if (pos != std::string::npos)
++pos;
}
// check if any empty macros are used in code
for (std::set<std::string>::const_iterator it = macros.begin(); it != macros.end(); ++it) {
const std::string &macro = *it;
std::string::size_type pos = 0;
while ((pos = code.find_first_of(std::string("#\"'")+macro[0], pos)) != std::string::npos) {
const std::string::size_type pos1 = pos;
const std::string::size_type pos2 = pos + macro.size();
pos++;
// skip string..
if (code[pos1] == '\"' || code[pos1] == '\'') {
while (pos < code.size() && code[pos] != code[pos1]) {
if (code[pos] == '\\')
++pos;
++pos;
}
++pos;
}
// skip preprocessor statement..
else if (code[pos1] == '#') {
if (pos1 == 0 || code[pos1-1] == '\n')
pos = code.find('\n', pos);
}
// is macro used in code?
else if (code.compare(pos1,macro.size(),macro) == 0) {
if (pos1 > 0 && (std::isalnum((unsigned char)code[pos1-1U]) || code[pos1-1U] == '_'))
continue;
if (pos2 < code.size() && (std::isalnum((unsigned char)code[pos2]) || code[pos2] == '_'))
continue;
// macro is used in code, return false
if (printInformation)
validateCfgError(cfg, macro);
return false;
}
}
}
return true;
}
void Preprocessor::validateCfgError(const std::string &cfg, const std::string &macro)
{
const std::string id = "ConfigurationNotChecked";
2012-11-03 11:25:40 +01:00
std::list<ErrorLogger::ErrorMessage::FileLocation> locationList;
ErrorLogger::ErrorMessage::FileLocation loc(file0, 1);
2012-11-03 11:25:40 +01:00
locationList.push_back(loc);
ErrorLogger::ErrorMessage errmsg(locationList, Severity::information, "Skipping configuration '" + cfg + "' since the value of '" + macro + "' is unknown. Use -D if you want to check it. You can use -U to skip it explicitly.", id, false);
_errorLogger->reportInfo(errmsg);
}
std::string Preprocessor::expandMacros(const std::string &code, std::string filename, const std::string &cfg, ErrorLogger *errorLogger)
{
// Search for macros and expand them..
2010-01-13 21:50:44 +01:00
// --------------------------------------------
2010-01-13 21:50:44 +01:00
// Available macros (key=macroname, value=macro).
std::map<std::string, PreprocessorMacro *> macros;
const Settings settings;
{
// fill up "macros" with user defined macros
2014-02-15 08:37:57 +01:00
const std::map<std::string,std::string> cfgmap(getcfgmap(cfg,nullptr,""));
std::map<std::string, std::string>::const_iterator it;
for (it = cfgmap.begin(); it != cfgmap.end(); ++it) {
std::string s = it->first;
if (!it->second.empty())
s += " " + it->second;
PreprocessorMacro *macro = new PreprocessorMacro(s, &settings);
macros[it->first] = macro;
}
}
2010-01-13 21:50:44 +01:00
// Current line number
unsigned int linenr = 1;
// linenr, filename
std::stack< std::pair<unsigned int, std::string> > fileinfo;
2010-01-13 21:50:44 +01:00
// output stream
std::ostringstream ostr;
2010-01-13 21:50:44 +01:00
// read code..
2011-12-08 21:28:34 +01:00
std::istringstream istr(code);
std::string line;
2011-10-13 20:53:06 +02:00
while (getlines(istr, line)) {
if (line.empty())
continue;
// Preprocessor directive
if (line[0] == '#') {
// defining a macro..
if (line.compare(1, 7, "define ") == 0) {
PreprocessorMacro *macro = new PreprocessorMacro(line.substr(8), &settings);
if (macro->name().empty() || macro->name() == "NULL") {
delete macro;
} else if (macro->name() == "BOOST_FOREACH") {
// BOOST_FOREACH is currently too complex to parse, so skip it.
delete macro;
} else {
std::map<std::string, PreprocessorMacro *>::iterator it;
it = macros.find(macro->name());
if (it != macros.end())
delete it->second;
macros[macro->name()] = macro;
}
line = "\n";
}
// undefining a macro..
else if (line.compare(1, 6, "undef ") == 0) {
std::map<std::string, PreprocessorMacro *>::iterator it;
it = macros.find(line.substr(7));
if (it != macros.end()) {
delete it->second;
macros.erase(it);
}
line = "\n";
}
// entering a file, update position..
else if (line.compare(1, 6, "file \"") == 0) {
fileinfo.push(std::pair<unsigned int, std::string>(linenr, filename));
filename = line.substr(7, line.length() - 8);
linenr = 0;
line += "\n";
2009-12-06 20:10:10 +01:00
}
// leaving a file, update position..
else if (line == "#endfile") {
if (!fileinfo.empty()) {
linenr = fileinfo.top().first;
filename = fileinfo.top().second;
fileinfo.pop();
}
line += "\n";
2009-12-06 20:10:10 +01:00
}
// all other preprocessor directives are just replaced with a newline
else
line += "\n";
}
// expand macros..
2011-10-13 20:53:06 +02:00
else {
// Limit for each macro.
// The limit specify a position in the "line" variable.
// For a "recursive macro" where the expanded text contains
// the macro again, the macro should not be expanded again.
// The limits are used to prevent recursive expanding.
// * When a macro is expanded its limit position is set to
// the last expanded character.
// * macros are only allowed to be expanded when the
// the position is beyond the limit.
// * The limit is relative to the end of the "line"
// variable. Inserting and deleting text before the limit
// without updating the limit is safe.
// * when pos goes beyond a limit the limit needs to be
// deleted because it is unsafe to insert/delete text
2010-01-13 21:50:44 +01:00
// after the limit otherwise
std::map<const PreprocessorMacro *, std::size_t> limits;
// pos is the current position in line
std::string::size_type pos = 0;
2010-01-13 21:50:44 +01:00
// scan line to see if there are any macros to expand..
unsigned int tmpLinenr = 0;
2011-10-13 20:53:06 +02:00
while (pos < line.size()) {
if (line[pos] == '\n')
++tmpLinenr;
// skip strings..
2011-10-13 20:53:06 +02:00
if (line[pos] == '\"' || line[pos] == '\'') {
const char ch = line[pos];
skipstring(line, pos);
++pos;
2011-10-13 20:53:06 +02:00
if (pos >= line.size()) {
writeError(Path::simplifyPath(filename),
linenr + tmpLinenr,
errorLogger,
"noQuoteCharPair",
std::string("No pair for character (") + ch + "). Can't process file. File is either invalid or unicode, which is currently not supported.");
std::map<std::string, PreprocessorMacro *>::iterator it;
for (it = macros.begin(); it != macros.end(); ++it)
delete it->second;
macros.clear();
return "";
}
2009-12-06 20:10:10 +01:00
continue;
}
if (!std::isalpha((unsigned char)line[pos]) && line[pos] != '_')
++pos;
// found an identifier..
2010-01-13 21:50:44 +01:00
// the "while" is used in case the expanded macro will immediately call another macro
while (pos < line.length() && (std::isalpha((unsigned char)line[pos]) || line[pos] == '_')) {
2010-01-13 21:50:44 +01:00
// pos1 = start position of macro
const std::string::size_type pos1 = pos++;
2010-01-13 21:50:44 +01:00
// find the end of the identifier
while (pos < line.size() && (std::isalnum((unsigned char)line[pos]) || line[pos] == '_'))
++pos;
2010-01-13 21:50:44 +01:00
// get identifier
const std::string id = line.substr(pos1, pos - pos1);
// is there a macro with this name?
std::map<std::string, PreprocessorMacro *>::const_iterator it;
it = macros.find(id);
if (it == macros.end())
2010-01-13 21:50:44 +01:00
break; // no macro with this name exist
const PreprocessorMacro * const macro = it->second;
2009-12-06 20:10:10 +01:00
2010-01-13 21:50:44 +01:00
// check that pos is within allowed limits for this
// macro
2009-12-06 20:10:10 +01:00
{
const std::map<const PreprocessorMacro *, std::size_t>::const_iterator it2 = limits.find(macro);
if (it2 != limits.end() && pos <= line.length() - it2->second)
2009-12-06 20:10:10 +01:00
break;
}
2010-01-13 21:50:44 +01:00
// get parameters from line..
2014-11-18 06:38:19 +01:00
if (macro->params().size() && pos >= line.length())
break;
std::vector<std::string> params;
std::string::size_type pos2 = pos;
2010-01-13 21:50:44 +01:00
// number of newlines within macro use
unsigned int numberOfNewlines = 0;
2010-12-15 18:45:53 +01:00
// if the macro has parentheses, get parameters
2011-10-13 20:53:06 +02:00
if (macro->variadic() || macro->nopar() || macro->params().size()) {
// is the end parentheses found?
bool endFound = false;
2010-01-13 21:50:44 +01:00
getparams(line,pos2,params,numberOfNewlines,endFound);
2010-01-13 21:50:44 +01:00
// something went wrong so bail out
if (!endFound)
break;
}
2010-01-13 21:50:44 +01:00
// Just an empty parameter => clear
if (params.size() == 1 && params[0] == "")
params.clear();
2010-01-13 21:50:44 +01:00
// Check that it's the same number of parameters..
if (!macro->variadic() && params.size() != macro->params().size())
break;
// Create macro code..
std::string tempMacro;
2011-10-13 20:53:06 +02:00
if (!macro->code(params, macros, tempMacro)) {
// Syntax error in code
writeError(filename,
linenr + tmpLinenr,
errorLogger,
"syntaxError",
std::string("Syntax error. Not enough parameters for macro '") + macro->name() + "'.");
2010-04-09 21:40:37 +02:00
std::map<std::string, PreprocessorMacro *>::iterator iter;
for (iter = macros.begin(); iter != macros.end(); ++iter)
delete iter->second;
macros.clear();
return "";
}
2009-12-06 20:10:10 +01:00
2010-01-13 21:50:44 +01:00
// make sure number of newlines remain the same..
std::string macrocode(std::string(numberOfNewlines, '\n') + tempMacro);
2009-12-06 20:10:10 +01:00
// Insert macro code..
if (macro->variadic() || macro->nopar() || !macro->params().empty())
++pos2;
2009-12-06 20:10:10 +01:00
// Remove old limits
for (std::map<const PreprocessorMacro *, std::size_t>::iterator iter = limits.begin();
2011-10-13 20:53:06 +02:00
iter != limits.end();) {
if ((line.length() - pos1) < iter->second) {
// We have gone past this limit, so just delete it
limits.erase(iter++);
2011-10-13 20:53:06 +02:00
} else {
++iter;
}
}
2009-12-06 20:10:10 +01:00
// don't allow this macro to be expanded again before pos2
limits[macro] = line.length() - pos2;
// erase macro
line.erase(pos1, pos2 - pos1);
// Don't glue this macro into variable or number after it
if (!line.empty() && (std::isalnum((unsigned char)line[pos1]) || line[pos1] == '_'))
macrocode.append(1,' ');
// insert macrochar before each symbol/nr/operator
bool str = false;
bool chr = false;
for (std::size_t i = 0U; i < macrocode.size(); ++i) {
if (macrocode[i] == '\\') {
i++;
continue;
} else if (macrocode[i] == '\"')
str = !str;
else if (macrocode[i] == '\'')
chr = !chr;
else if (str || chr)
continue;
else if (macrocode[i] == '.') { // 5. / .5
if ((i > 0U && std::isdigit((unsigned char)macrocode[i-1])) ||
(i+1 < macrocode.size() && std::isdigit((unsigned char)macrocode[i+1]))) {
if (i > 0U && !std::isdigit((unsigned char)macrocode[i-1])) {
macrocode.insert(i, 1U, macroChar);
i++;
}
i++;
if (i<macrocode.size() && std::isdigit((unsigned char)macrocode[i]))
i++;
2014-01-08 18:05:14 +01:00
if (i+1U < macrocode.size() &&
(macrocode[i] == 'e' || macrocode[i] == 'E') &&
(macrocode[i+1] == '+' || macrocode[i+1] == '-')) {
i+=2;
}
}
} else if (std::isalnum((unsigned char)macrocode[i]) || macrocode[i] == '_') {
if ((i > 0U) &&
(!std::isalnum((unsigned char)macrocode[i-1])) &&
(macrocode[i-1] != '_') &&
(macrocode[i-1] != macroChar)) {
macrocode.insert(i, 1U, macroChar);
}
// 1e-7 / 1e+7
if (i+3U < macrocode.size() &&
(std::isdigit((unsigned char)macrocode[i]) || macrocode[i]=='.') &&
(macrocode[i+1] == 'e' || macrocode[i+1] == 'E') &&
(macrocode[i+2] == '-' || macrocode[i+2] == '+') &&
std::isdigit((unsigned char)macrocode[i+3])) {
i += 3U;
}
// 1.f / 1.e7
if (i+2U < macrocode.size() &&
std::isdigit((unsigned char)macrocode[i]) &&
macrocode[i+1] == '.' &&
std::isalpha((unsigned char)macrocode[i+2])) {
i += 2U;
if (i+2U < macrocode.size() &&
(macrocode[i+0] == 'e' || macrocode[i+0] == 'E') &&
(macrocode[i+1] == '-' || macrocode[i+1] == '+') &&
std::isdigit((unsigned char)macrocode[i+2])) {
i += 2U;
}
}
}
}
line.insert(pos1, macroChar + macrocode);
// position = start position.
pos = pos1;
}
}
}
2010-01-13 21:50:44 +01:00
// the line has been processed in various ways. Now add it to the output stream
ostr << line;
2010-01-13 21:50:44 +01:00
// update linenr
2011-10-13 20:53:06 +02:00
for (std::string::size_type p = 0; p < line.length(); ++p) {
if (line[p] == '\n')
++linenr;
}
}
2011-11-26 21:02:04 +01:00
for (std::map<std::string, PreprocessorMacro *>::iterator it = macros.begin(); it != macros.end(); ++it)
delete it->second;
macros.clear();
return ostr.str();
}
void Preprocessor::getErrorMessages(ErrorLogger *errorLogger, const Settings *settings)
2010-09-03 13:30:49 +02:00
{
Settings settings2(*settings);
Preprocessor preprocessor(settings2, errorLogger);
settings2.checkConfiguration=true;
preprocessor.missingInclude("", 1, "", UserHeader);
preprocessor.missingInclude("", 1, "", SystemHeader);
preprocessor.validateCfgError("X", "X");
preprocessor.error("", 1, "#error message"); // #error ..
2010-09-03 13:30:49 +02:00
}
void Preprocessor::dump(std::ostream &out) const
{
// Create a xml directive dump.
// The idea is not that this will be readable for humans. It's a
// data dump that 3rd party tools could load and get useful info from.
std::list<Directive>::const_iterator it;
out << " <directivelist>" << std::endl;
for (it = directives.begin(); it != directives.end(); ++it) {
out << " <directive "
<< "file=\"" << it->file << "\" "
<< "linenr=\"" << it->linenr << "\" "
// str might contain characters such as '"', '<' or '>' which
// could result in invalid XML, so run it through toxml().
<< "str=\"" << ErrorLogger::toxml(it->str) << "\"/>" << std::endl;
}
out << " </directivelist>" << std::endl;
}