|
| 1 | +/* |
| 2 | + * Cppcheck - A tool for static C/C++ code analysis |
| 3 | + * Copyright (C) 2007-2019 Cppcheck team. |
| 4 | + * |
| 5 | + * This program is free software: you can redistribute it and/or modify |
| 6 | + * it under the terms of the GNU General Public License as published by |
| 7 | + * the Free Software Foundation, either version 3 of the License, or |
| 8 | + * (at your option) any later version. |
| 9 | + * |
| 10 | + * This program is distributed in the hope that it will be useful, |
| 11 | + * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 12 | + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 13 | + * GNU General Public License for more details. |
| 14 | + * |
| 15 | + * You should have received a copy of the GNU General Public License |
| 16 | + * along with this program. If not, see <http://www.gnu.org/licenses/>. |
| 17 | + */ |
| 18 | + |
| 19 | +#include "clangastdump.h" |
| 20 | +#include "symboldatabase.h" |
| 21 | +#include "tokenize.h" |
| 22 | + |
| 23 | +#include <memory> |
| 24 | +#include <vector> |
| 25 | +#include <iostream> |
| 26 | + |
| 27 | +static const std::string BinaryOperator = "BinaryOperator"; |
| 28 | +static const std::string CallExpr = "CallExpr"; |
| 29 | +static const std::string CompoundStmt = "CompoundStmt"; |
| 30 | +static const std::string DeclRefExpr = "DeclRefExpr"; |
| 31 | +static const std::string FunctionDecl = "FunctionDecl"; |
| 32 | +static const std::string ImplicitCastExpr = "ImplicitCastExpr"; |
| 33 | +static const std::string IntegerLiteral = "IntegerLiteral"; |
| 34 | +static const std::string ParmVarDecl = "ParmVarDecl"; |
| 35 | +static const std::string ReturnStmt = "ReturnStmt"; |
| 36 | +static const std::string UnaryOperator = "UnaryOperator"; |
| 37 | + |
| 38 | +static std::string unquote(const std::string &s) |
| 39 | +{ |
| 40 | + return (s[0] == '\'') ? s.substr(1, s.size() - 2) : s; |
| 41 | +} |
| 42 | + |
| 43 | +static std::vector<std::string> splitString(const std::string &line) |
| 44 | +{ |
| 45 | + std::vector<std::string> ret; |
| 46 | + std::string::size_type pos1 = line.find_first_not_of(" "); |
| 47 | + while (pos1 != std::string::npos) { |
| 48 | + std::string::size_type pos2; |
| 49 | + if (line[pos1] == '<') |
| 50 | + pos2 = line.find(">", pos1); |
| 51 | + else if (line[pos1] == '\'') |
| 52 | + pos2 = line.find("\'", pos1+1); |
| 53 | + else |
| 54 | + pos2 = line.find(" ", pos1) - 1; |
| 55 | + ret.push_back(line.substr(pos1, pos2+1-pos1)); |
| 56 | + if (pos2 == std::string::npos) |
| 57 | + break; |
| 58 | + pos1 = line.find_first_not_of(" ", pos2 + 1); |
| 59 | + } |
| 60 | + return ret; |
| 61 | +} |
| 62 | + |
| 63 | +namespace clangastdump { |
| 64 | + struct Data { |
| 65 | + std::map<std::string, int> varId; |
| 66 | + std::map<std::string, Variable *> variableMap; |
| 67 | + }; |
| 68 | + |
| 69 | + class AstNode { |
| 70 | + public: |
| 71 | + AstNode(const std::string &nodeType, const std::string &ext, Data *data, SymbolDatabase *symbolDatabase) |
| 72 | + : nodeType(nodeType), mExtTokens(splitString(ext)), mData(data), mSymbolDatabase(symbolDatabase) |
| 73 | + {} |
| 74 | + std::string nodeType; |
| 75 | + std::vector<std::shared_ptr<AstNode>> children; |
| 76 | + |
| 77 | + void setLocations(TokenList *tokenList, int file, int line, int col); |
| 78 | + |
| 79 | + void dumpAst(int num = 0, int indent = 0) const; |
| 80 | + Token *createTokens(TokenList *tokenList); |
| 81 | + private: |
| 82 | + Token *addtoken(TokenList *tokenList, const std::string &str); |
| 83 | + Token *addTypeTokens(TokenList *tokenList, const std::string &str); |
| 84 | + std::string getSpelling() const; |
| 85 | + std::string getType() const; |
| 86 | + |
| 87 | + int mFile = 0; |
| 88 | + int mLine = 1; |
| 89 | + int mCol = 1; |
| 90 | + int mVarId = 0; |
| 91 | + std::vector<std::string> mExtTokens; |
| 92 | + Data *mData; |
| 93 | + SymbolDatabase *mSymbolDatabase; |
| 94 | + }; |
| 95 | + |
| 96 | + typedef std::shared_ptr<AstNode> AstNodePtr; |
| 97 | +} |
| 98 | + |
| 99 | +std::string clangastdump::AstNode::getSpelling() const |
| 100 | +{ |
| 101 | + if (nodeType == ParmVarDecl) |
| 102 | + return mExtTokens[mExtTokens.size() - 2]; |
| 103 | + return ""; |
| 104 | +} |
| 105 | + |
| 106 | +std::string clangastdump::AstNode::getType() const |
| 107 | +{ |
| 108 | + if (nodeType == DeclRefExpr) |
| 109 | + return unquote(mExtTokens.back()); |
| 110 | + if (nodeType == BinaryOperator) |
| 111 | + return unquote(mExtTokens[mExtTokens.size() - 2]); |
| 112 | + if (nodeType == IntegerLiteral) |
| 113 | + return unquote(mExtTokens[mExtTokens.size() - 2]); |
| 114 | + return ""; |
| 115 | +} |
| 116 | + |
| 117 | +void clangastdump::AstNode::dumpAst(int num, int indent) const |
| 118 | +{ |
| 119 | + (void)num; |
| 120 | + std::cout << std::string(indent, ' ') << nodeType; |
| 121 | + for (auto tok: mExtTokens) |
| 122 | + std::cout << " " << tok; |
| 123 | + std::cout << std::endl; |
| 124 | + for (int c = 0; c < children.size(); ++c) |
| 125 | + children[c]->dumpAst(c, indent + 2); |
| 126 | +} |
| 127 | + |
| 128 | +void clangastdump::AstNode::setLocations(TokenList *tokenList, int file, int line, int col) |
| 129 | +{ |
| 130 | + for (const std::string &ext: mExtTokens) { |
| 131 | + if (ext.compare(0,5,"<col:") == 0) |
| 132 | + col = std::atoi(ext.substr(5).c_str()); |
| 133 | + else if (ext.compare(0,6,"<line:") == 0) |
| 134 | + line = std::atoi(ext.substr(6).c_str()); |
| 135 | + else if (ext[0] == '<') |
| 136 | + file = tokenList->appendFileIfNew(ext.substr(1,ext.find(":") - 1)); |
| 137 | + } |
| 138 | + mFile = file; |
| 139 | + mLine = line; |
| 140 | + mCol = col; |
| 141 | + for (auto child: children) |
| 142 | + child->setLocations(tokenList, file, line, col); |
| 143 | +} |
| 144 | + |
| 145 | +Token *clangastdump::AstNode::addtoken(TokenList *tokenList, const std::string &str) |
| 146 | +{ |
| 147 | + tokenList->addtoken(str, mLine, mFile); |
| 148 | + if (getType() == "int") |
| 149 | + tokenList->back()->setValueType(new ValueType(ValueType::Sign::SIGNED, ValueType::Type::INT, 0)); |
| 150 | + return tokenList->back(); |
| 151 | +} |
| 152 | + |
| 153 | +Token *clangastdump::AstNode::addTypeTokens(TokenList *tokenList, const std::string &str) |
| 154 | +{ |
| 155 | + if (str.find(" (") == std::string::npos) |
| 156 | + return addtoken(tokenList, unquote(str)); |
| 157 | + return addtoken(tokenList, str.substr(1,str.find(" (")-1)); |
| 158 | +} |
| 159 | + |
| 160 | +Token *clangastdump::AstNode::createTokens(TokenList *tokenList) |
| 161 | +{ |
| 162 | + if (nodeType == BinaryOperator) { |
| 163 | + Token *tok1 = children[0]->createTokens(tokenList); |
| 164 | + Token *binop = addtoken(tokenList, unquote(mExtTokens.back())); |
| 165 | + Token *tok2 = children[1]->createTokens(tokenList); |
| 166 | + binop->astOperand1(tok1); |
| 167 | + binop->astOperand2(tok2); |
| 168 | + return binop; |
| 169 | + } |
| 170 | + if (nodeType == CallExpr) { |
| 171 | + Token *op1 = children[0]->createTokens(tokenList); |
| 172 | + Token *call = addtoken(tokenList, "("); |
| 173 | + call->astOperand1(op1); |
| 174 | + for (int c = 1; c < children.size(); ++c) |
| 175 | + call->astOperand2(children[c]->createTokens(tokenList)); |
| 176 | + call->link(addtoken(tokenList, ")")); |
| 177 | + return call; |
| 178 | + } |
| 179 | + if (nodeType == CompoundStmt) { |
| 180 | + Token *start = addtoken(tokenList, "{"); |
| 181 | + for (AstNodePtr child: children) { |
| 182 | + child->createTokens(tokenList); |
| 183 | + child->addtoken(tokenList, ";"); |
| 184 | + } |
| 185 | + Token *end = addtoken(tokenList, "}"); |
| 186 | + start->link(end); |
| 187 | + return start; |
| 188 | + } |
| 189 | + if (nodeType == DeclRefExpr) { |
| 190 | + Token *vartok = addtoken(tokenList, unquote(mExtTokens[mExtTokens.size() - 2])); |
| 191 | + std::string addr = mExtTokens[mExtTokens.size() - 3]; |
| 192 | + vartok->varId(mData->varId[addr]); |
| 193 | + vartok->variable(mData->variableMap[addr]); |
| 194 | + return vartok; |
| 195 | + } |
| 196 | + if (nodeType == FunctionDecl) { |
| 197 | + addTypeTokens(tokenList, mExtTokens.back()); |
| 198 | + Token *nameToken = addtoken(tokenList, mExtTokens[mExtTokens.size() - 2]); |
| 199 | + Scope &globalScope = mSymbolDatabase->scopeList.front(); |
| 200 | + mSymbolDatabase->scopeList.push_back(Scope(nullptr, nullptr, &globalScope)); |
| 201 | + Scope &scope = mSymbolDatabase->scopeList.back(); |
| 202 | + mSymbolDatabase->functionScopes.push_back(&scope); |
| 203 | + globalScope.functionList.push_back(Function(nameToken)); |
| 204 | + scope.function = &globalScope.functionList.back(); |
| 205 | + Token *par1 = addtoken(tokenList, "("); |
| 206 | + for (AstNodePtr child: children) { |
| 207 | + if (child->nodeType != ParmVarDecl) |
| 208 | + continue; |
| 209 | + if (tokenList->back() != par1) |
| 210 | + addtoken(tokenList, ","); |
| 211 | + addTypeTokens(tokenList, child->mExtTokens.back()); |
| 212 | + const std::string spelling = child->getSpelling(); |
| 213 | + if (!spelling.empty()) { |
| 214 | + Token *vartok = addtoken(tokenList, spelling); |
| 215 | + std::string addr = child->mExtTokens[0]; |
| 216 | + int varId = mData->varId.size() + 1; |
| 217 | + mData->varId[addr] = varId; |
| 218 | + vartok->varId(varId); |
| 219 | + scope.function->argumentList.push_back(Variable(vartok, nullptr, nullptr, varId, AccessControl::Argument, nullptr, nullptr, nullptr)); |
| 220 | + Variable *var = &scope.function->argumentList.back(); |
| 221 | + mData->variableMap[addr] = var; |
| 222 | + vartok->variable(var); |
| 223 | + var->setValueType(ValueType(ValueType::Sign::SIGNED, ValueType::Type::INT, 0)); |
| 224 | + } |
| 225 | + } |
| 226 | + Token *par2 = addtoken(tokenList, ")"); |
| 227 | + par1->link(par2); |
| 228 | + children.back()->createTokens(tokenList); |
| 229 | + if (Token::simpleMatch(par2, ") {")) { |
| 230 | + Token *bodyStart = par2->next(); |
| 231 | + scope.bodyStart = bodyStart; |
| 232 | + scope.bodyEnd = bodyStart->link(); |
| 233 | + } |
| 234 | + return nullptr; |
| 235 | + } |
| 236 | + if (nodeType == ImplicitCastExpr) |
| 237 | + return children[0]->createTokens(tokenList); |
| 238 | + if (nodeType == IntegerLiteral) |
| 239 | + return addtoken(tokenList, mExtTokens.back()); |
| 240 | + if (nodeType == ReturnStmt) { |
| 241 | + Token *tok1 = addtoken(tokenList, "return"); |
| 242 | + if (!children.empty()) |
| 243 | + tok1->astOperand1(children[0]->createTokens(tokenList)); |
| 244 | + return tok1; |
| 245 | + } |
| 246 | + if (nodeType == UnaryOperator) { |
| 247 | + Token *unop = addtoken(tokenList, unquote(mExtTokens.back())); |
| 248 | + unop->astOperand1(children[0]->createTokens(tokenList)); |
| 249 | + return unop; |
| 250 | + } |
| 251 | + return addtoken(tokenList, "?" + nodeType + "?"); |
| 252 | +} |
| 253 | + |
| 254 | +void clangastdump::parseClangAstDump(Tokenizer *tokenizer, std::istream &f) |
| 255 | +{ |
| 256 | + TokenList *tokenList = &tokenizer->list; |
| 257 | + clangastdump::Data data; |
| 258 | + |
| 259 | + tokenizer->createSymbolDatabase(); |
| 260 | + SymbolDatabase *symbolDatabase = const_cast<SymbolDatabase *>(tokenizer->getSymbolDatabase()); |
| 261 | + symbolDatabase->scopeList.push_back(Scope(nullptr, nullptr, nullptr)); |
| 262 | + symbolDatabase->scopeList.back().type = Scope::ScopeType::eGlobal; |
| 263 | + |
| 264 | + std::string line; |
| 265 | + std::vector<AstNodePtr> tree; |
| 266 | + while (std::getline(f,line)) { |
| 267 | + const std::string::size_type pos1 = line.find("-"); |
| 268 | + if (pos1 == std::string::npos) |
| 269 | + continue; |
| 270 | + const std::string::size_type pos2 = line.find(" ", pos1); |
| 271 | + if (pos2 < pos1 + 4 || pos2 == std::string::npos) |
| 272 | + continue; |
| 273 | + const std::string nodeType = line.substr(pos1+1, pos2 - pos1 - 1); |
| 274 | + const std::string ext = line.substr(pos2); |
| 275 | + |
| 276 | + if (nodeType == FunctionDecl) { |
| 277 | + if (!tree.empty()) { |
| 278 | + tree[0]->setLocations(tokenList, 0, 1, 1); |
| 279 | + tree[0]->createTokens(tokenList); |
| 280 | + } |
| 281 | + tree.clear(); |
| 282 | + tree. push_back(std::make_shared<AstNode>(nodeType, ext, &data, symbolDatabase)); |
| 283 | + continue; |
| 284 | + } |
| 285 | + |
| 286 | + const int level = (pos1 - 1) / 2; |
| 287 | + if (level == 0 || tree.empty()) |
| 288 | + continue; |
| 289 | + |
| 290 | + AstNodePtr newNode = std::make_shared<AstNode>(nodeType, ext, &data, symbolDatabase); |
| 291 | + tree[level - 1]->children.push_back(newNode); |
| 292 | + if (level >= tree.size()) |
| 293 | + tree.push_back(newNode); |
| 294 | + else |
| 295 | + tree[level] = newNode; |
| 296 | + } |
| 297 | + |
| 298 | + if (!tree.empty()) { |
| 299 | + tree[0]->setLocations(tokenList, 0, 1, 1); |
| 300 | + tree[0]->createTokens(tokenList); |
| 301 | + } |
| 302 | + |
| 303 | + symbolDatabase->clangSetVariables(data.variableMap); |
| 304 | + tokenList->clangSetOrigFiles(); |
| 305 | +} |
| 306 | + |
0 commit comments