Analyze.cpp

   1 /* Cycript - Optimizing JavaScript Compiler/Runtime
   2  * Copyright (C) 2009-2015  Jay Freeman (saurik)
   3 */
   4
   5 /* GNU Affero General Public License, Version 3 {{{ */
   6 /*
   7  * This program is free software: you can redistribute it and/or modify
   8  * it under the terms of the GNU Affero General Public License as published by
   9  * the Free Software Foundation, either version 3 of the License, or
  10  * (at your option) any later version.
  11
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  * GNU Affero General Public License for more details.
  16
  17  * You should have received a copy of the GNU Affero General Public License
  18  * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  19 **/
  20 /* }}} */
  21
  22 #include <cmath>
  23 #include <cstring>
  24 #include <iostream>
  25 #include <map>
  26 #include <sstream>
  27 #include <string>
  28
  29 #include <clang-c/Index.h>
  30
  31 #include "Functor.hpp"
  32 #include "Replace.hpp"
  33 #include "Syntax.hpp"
  34
  35 static CXChildVisitResult CYVisit(CXCursor cursor, CXCursor parent, CXClientData arg) {
  36     (*reinterpret_cast<const Functor<void (CXCursor)> *>(arg))(cursor);
  37     return CXChildVisit_Continue;
  38 }
  39
  40 static unsigned CYForChild(CXCursor cursor, const Functor<void (CXCursor)> &visitor) {
  41     return clang_visitChildren(cursor, &CYVisit, const_cast<void *>(static_cast<const void *>(&visitor)));
  42 }
  43
  44 static bool CYOneChild(CXCursor cursor, const Functor<void (CXCursor)> &visitor) {
  45     bool visited(false);
  46     CYForChild(cursor, fun([&](CXCursor child) {
  47         _assert(!visited);
  48         visited = true;
  49         visitor(child);
  50     }));
  51     return visited;
  52 }
  53
  54 struct CYCXString {
  55     CXString value_;
  56
  57     CYCXString(CXString value) :
  58         value_(value)
  59     {
  60     }
  61
  62     CYCXString(CXCursor cursor) :
  63         value_(clang_getCursorSpelling(cursor))
  64     {
  65     }
  66
  67     CYCXString(CXCursorKind kind) :
  68         value_(clang_getCursorKindSpelling(kind))
  69     {
  70     }
  71
  72     CYCXString(CXFile file) :
  73         value_(clang_getFileName(file))
  74     {
  75     }
  76
  77     CYCXString(CXTranslationUnit unit, CXToken token) :
  78         value_(clang_getTokenSpelling(unit, token))
  79     {
  80     }
  81
  82     ~CYCXString() {
  83         clang_disposeString(value_);
  84     }
  85
  86     operator const char *() const {
  87         return clang_getCString(value_);
  88     }
  89
  90     const char *Pool(CYPool &pool) const {
  91         return pool.strdup(*this);
  92     }
  93
  94     bool operator ==(const char *rhs) const {
  95         const char *lhs(*this);
  96         return lhs == rhs || strcmp(lhs, rhs) == 0;
  97     }
  98 };
  99
 100 template <void (&clang_get_Location)(CXSourceLocation, CXFile *, unsigned *, unsigned *, unsigned *) = clang_getSpellingLocation>
 101 struct CYCXPosition {
 102     CXFile file_;
 103     unsigned line_;
 104     unsigned column_;
 105     unsigned offset_;
 106
 107     CYCXPosition(CXSourceLocation location) {
 108         clang_get_Location(location, &file_, &line_, &column_, &offset_);
 109     }
 110
 111     CYCXPosition(CXTranslationUnit unit, CXToken token) :
 112         CYCXPosition(clang_getTokenLocation(unit, token))
 113     {
 114     }
 115
 116     CXSourceLocation Get(CXTranslationUnit unit) const {
 117         return clang_getLocation(unit, file_, line_, column_);
 118     }
 119 };
 120
 121 template <void (&clang_get_Location)(CXSourceLocation, CXFile *, unsigned *, unsigned *, unsigned *)>
 122 std::ostream &operator <<(std::ostream &out, const CYCXPosition<clang_get_Location> &position) {
 123     if (position.file_ != NULL)
 124         out << "[" << CYCXString(position.file_) << "]:";
 125     out << position.line_ << ":" << position.column_ << "@" << position.offset_;
 126     return out;
 127 }
 128
 129 typedef std::map<std::string, std::string> CYKeyMap;
 130
 131 struct CYChildBaton {
 132     CXTranslationUnit unit;
 133     CYKeyMap &keys;
 134
 135     CYChildBaton(CXTranslationUnit unit, CYKeyMap &keys) :
 136         unit(unit),
 137         keys(keys)
 138     {
 139     }
 140 };
 141
 142 struct CYTokens {
 143   private:
 144     CXTranslationUnit unit_;
 145     CXToken *tokens_;
 146     unsigned count_;
 147     unsigned valid_;
 148
 149   public:
 150     CYTokens(CXTranslationUnit unit, CXSourceRange range) :
 151         unit_(unit)
 152     {
 153         clang_tokenize(unit_, range, &tokens_, &count_);
 154
 155
 156         // libclang's tokenizer is horribly broken and returns "extra" tokens.
 157         // this code goes back through the tokens and filters for good ones :/
 158
 159         CYCXPosition<> end(clang_getRangeEnd(range));
 160         CYCXString file(end.file_);
 161
 162         for (valid_ = 0; valid_ != count_; ++valid_) {
 163             CYCXPosition<> position(unit, tokens_[valid_]);
 164             _assert(CYCXString(position.file_) == file);
 165             if (position.offset_ >= end.offset_)
 166                 break;
 167         }
 168     }
 169
 170     CYTokens(CXTranslationUnit unit, CXCursor cursor) :
 171         CYTokens(unit, clang_getCursorExtent(cursor))
 172     {
 173     }
 174
 175     ~CYTokens() {
 176         clang_disposeTokens(unit_, tokens_, count_);
 177     }
 178
 179     operator CXToken *() const {
 180         return tokens_;
 181     }
 182
 183     size_t size() const {
 184         return valid_;
 185     }
 186 };
 187
 188 static CYUTF8String CYCXPoolUTF8Range(CYPool &pool, CXSourceRange range) {
 189     CYCXPosition<> start(clang_getRangeStart(range));
 190     CYCXPosition<> end(clang_getRangeEnd(range));
 191     CYCXString file(start.file_);
 192     _assert(file == CYCXString(end.file_));
 193
 194     CYPool temp;
 195     size_t size;
 196     char *data(static_cast<char *>(CYPoolFile(temp, file, &size)));
 197     _assert(start.offset_ <= size && end.offset_ <= size && start.offset_ <= end.offset_);
 198
 199     CYUTF8String code;
 200     code.size = end.offset_ - start.offset_;
 201     code.data = pool.strndup(data + start.offset_, code.size);
 202     return code;
 203 }
 204
 205 static CYExpression *CYTranslateExpression(CXTranslationUnit unit, CXCursor cursor) {
 206     switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
 207         case CXCursor_CallExpr: {
 208             CYExpression *function(NULL);
 209             CYList<CYArgument> arguments;
 210             CYForChild(cursor, fun([&](CXCursor child) {
 211                 CYExpression *expression(CYTranslateExpression(unit, child));
 212                 if (function == NULL)
 213                     function = expression;
 214                 else
 215                     arguments->*$C_(expression);
 216             }));
 217             return $C(function, arguments);
 218         } break;
 219
 220         case CXCursor_DeclRefExpr: {
 221             return $V(CYCXString(cursor).Pool($pool));
 222         } break;
 223
 224         case CXCursor_IntegerLiteral: {
 225             // libclang doesn't provide any reasonable way to do this
 226             // note: clang_tokenize doesn't work if this is a macro
 227             // the token range starts inside the macro but ends after it
 228             // the tokenizer freaks out and either fails with 0 tokens
 229             // or returns some massive number of tokens ending here :/
 230
 231             CYUTF8String token(CYCXPoolUTF8Range($pool, clang_getCursorExtent(cursor)));
 232             double value(CYCastDouble(token));
 233             if (std::isnan(value))
 234                 return $V(token.data);
 235             return $ CYNumber(value);
 236         } break;
 237
 238         case CXCursor_CStyleCastExpr:
 239             // XXX: most of the time, this is a "NoOp" integer cast; but we should check it
 240
 241         case CXCursor_UnexposedExpr:
 242             // there is a very high probability that this is actually an "ImplicitCastExpr"
 243             // "Douglas Gregor" <dgregor@apple.com> err'd on the incorrect side of this one
 244             // http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20110926/046998.html
 245
 246         case CXCursor_ParenExpr: {
 247             CYExpression *pass(NULL);
 248             CYOneChild(cursor, fun([&](CXCursor child) {
 249                 pass = CYTranslateExpression(unit, child);
 250             }));
 251             return pass;
 252         } break;
 253
 254         default:
 255             //std::cerr << "E:" << CYCXString(kind) << std::endl;
 256             _assert(false);
 257     }
 258 }
 259
 260 static CYStatement *CYTranslateStatement(CXTranslationUnit unit, CXCursor cursor) {
 261     switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
 262         case CXCursor_ReturnStmt: {
 263             CYExpression *value(NULL);
 264             CYOneChild(cursor, fun([&](CXCursor child) {
 265                 value = CYTranslateExpression(unit, child);
 266             }));
 267             return $ CYReturn(value);
 268         } break;
 269
 270         default:
 271             //std::cerr << "S:" << CYCXString(kind) << std::endl;
 272             _assert(false);
 273     }
 274 }
 275
 276 static CYStatement *CYTranslateBlock(CXTranslationUnit unit, CXCursor cursor) {
 277     CYList<CYStatement> statements;
 278     CYForChild(cursor, fun([&](CXCursor child) {
 279         statements->*CYTranslateStatement(unit, child);
 280     }));
 281     return $ CYBlock(statements);
 282 }
 283
 284 static CXChildVisitResult CYChildVisit(CXCursor cursor, CXCursor parent, CXClientData arg) {
 285     CYChildBaton &baton(*static_cast<CYChildBaton *>(arg));
 286     CXTranslationUnit &unit(baton.unit);
 287
 288     CYCXString spelling(cursor);
 289     std::string name(spelling);
 290     std::ostringstream value;
 291
 292     /*CXSourceLocation location(clang_getCursorLocation(cursor));
 293     CYCXPosition<> position(location);
 294     std::cout << spelling << " " << position << std::endl;*/
 295
 296     switch (CXCursorKind kind = clang_getCursorKind(cursor)) {
 297         case CXCursor_EnumConstantDecl: {
 298             value << clang_getEnumConstantDeclValue(cursor);
 299         } break;
 300
 301         case CXCursor_MacroDefinition: try {
 302             CXSourceRange range(clang_getCursorExtent(cursor));
 303             CYTokens tokens(unit, range);
 304             _assert(tokens.size() != 0);
 305
 306             CXCursor cursors[tokens.size()];
 307             clang_annotateTokens(unit, tokens, tokens.size(), cursors);
 308
 309             CYLocalPool local;
 310             CYList<CYFunctionParameter> parameters;
 311             unsigned offset(1);
 312
 313             if (tokens.size() != 1) {
 314                 CYCXPosition<> start(clang_getRangeStart(range));
 315                 CYCXString first(unit, tokens[offset]);
 316                 if (first == "(") {
 317                     CYCXPosition<> paren(unit, tokens[offset]);
 318                     if (start.offset_ + strlen(spelling) == paren.offset_) {
 319                         for (;;) {
 320                             _assert(++offset != tokens.size());
 321                             CYCXString token(unit, tokens[offset]);
 322                             parameters->*$P($B($I(token.Pool($pool))));
 323                             _assert(++offset != tokens.size());
 324                             CYCXString comma(unit, tokens[offset]);
 325                             if (comma == ")")
 326                                 break;
 327                             _assert(comma == ",");
 328                         }
 329                         ++offset;
 330                     }
 331                 }
 332             }
 333
 334             std::ostringstream body;
 335             for (unsigned i(offset); i != tokens.size(); ++i) {
 336                 CYCXString token(unit, tokens[i]);
 337                 if (i != offset)
 338                     body << " ";
 339                 body << token;
 340             }
 341
 342             if (!parameters)
 343                 value << body.str();
 344             else {
 345                 CYOptions options;
 346                 CYOutput out(*value.rdbuf(), options);
 347                 out << '(' << "function" << '(';
 348                 out << parameters;
 349                 out << ')' << '{';
 350                 out << "return" << ' ';
 351                 value << body.str();
 352                 out << ';' << '}' << ')';
 353             }
 354         } catch (const CYException &error) {
 355             CYPool pool;
 356             //std::cerr << error.PoolCString(pool) << std::endl;
 357             goto skip;
 358         } break;
 359
 360         case CXCursor_StructDecl: {
 361             if (!clang_isCursorDefinition(cursor))
 362                 goto skip;
 363             if (spelling[0] == '\0')
 364                 goto skip;
 365
 366             std::ostringstream types;
 367             std::ostringstream names;
 368
 369             CYForChild(cursor, fun([&](CXCursor child) {
 370                 if (clang_getCursorKind(child) == CXCursor_FieldDecl) {
 371                     CXType type(clang_getCursorType(child));
 372                     types << "(typedef " << CYCXString(clang_getTypeSpelling(type)) << "),";
 373                     names << "'" << CYCXString(child) << "',";
 374                 }
 375             }));
 376
 377             name += "$cy";
 378             value << "new Type([" << types.str() << "],[" << names.str() << "])";
 379         } break;
 380
 381         case CXCursor_TypedefDecl: {
 382             CXType type(clang_getTypedefDeclUnderlyingType(cursor));
 383             value << "(typedef " << CYCXString(clang_getTypeSpelling(type)) << ")";
 384         } break;
 385
 386         case CXCursor_FunctionDecl:
 387         case CXCursor_VarDecl: try {
 388             std::string label;
 389
 390             CYList<CYFunctionParameter> parameters;
 391             CYStatement *code(NULL);
 392
 393             CYLocalPool local;
 394
 395             CYForChild(cursor, fun([&](CXCursor child) {
 396                 switch (CXCursorKind kind = clang_getCursorKind(child)) {
 397                     case CXCursor_AsmLabelAttr:
 398                         label = CYCXString(child);
 399                         break;
 400
 401                     case CXCursor_CompoundStmt:
 402                         code = CYTranslateBlock(unit, child);
 403                         break;
 404
 405                     case CXCursor_ParmDecl:
 406                         parameters->*$P($B($I(CYCXString(child).Pool($pool))));
 407                         break;
 408
 409                     case CXCursor_IntegerLiteral:
 410                     case CXCursor_ObjCClassRef:
 411                     case CXCursor_TypeRef:
 412                     case CXCursor_UnexposedAttr:
 413                         break;
 414
 415                     default:
 416                         //std::cerr << "A:" << CYCXString(child) << std::endl;
 417                         break;
 418                 }
 419             }));
 420
 421             if (label.empty()) {
 422                 label = spelling;
 423                 label = '_' + label;
 424             } else if (label[0] != '_')
 425                 goto skip;
 426
 427             if (code == NULL) {
 428                 CXType type(clang_getCursorType(cursor));
 429                 value << "*(typedef " << CYCXString(clang_getTypeSpelling(type)) << ").pointerTo()(dlsym(RTLD_DEFAULT,'" << label.substr(1) << "'))";
 430             } else {
 431                 CYOptions options;
 432                 CYOutput out(*value.rdbuf(), options);
 433                 CYFunctionExpression *function($ CYFunctionExpression(NULL, parameters, code));
 434                 function->Output(out, CYNoBFC);
 435                 //std::cerr << value.str() << std::endl;
 436             }
 437         } catch (const CYException &error) {
 438             CYPool pool;
 439             //std::cerr << error.PoolCString(pool) << std::endl;
 440             goto skip;
 441         } break;
 442
 443         default: {
 444             return CXChildVisit_Recurse;
 445         } break;
 446     }
 447
 448     baton.keys[name] = value.str();
 449
 450   skip:
 451     return CXChildVisit_Continue;
 452 }
 453
 454 int main(int argc, const char *argv[]) {
 455     CXIndex index(clang_createIndex(0, 0));
 456
 457     const char *file(argv[1]);
 458
 459     unsigned offset(3);
 460 #if CY_OBJECTIVEC
 461     argv[--offset] = "-ObjC++";
 462 #endif
 463
 464     CXTranslationUnit unit(clang_parseTranslationUnit(index, file, argv + offset, argc - offset, NULL, 0, CXTranslationUnit_DetailedPreprocessingRecord));
 465
 466     for (unsigned i(0), e(clang_getNumDiagnostics(unit)); i != e; ++i) {
 467         CXDiagnostic diagnostic(clang_getDiagnostic(unit, i));
 468         CYCXString spelling(clang_getDiagnosticSpelling(diagnostic));
 469         std::cerr << spelling << std::endl;
 470     }
 471
 472     CYKeyMap keys;
 473     CYChildBaton baton(unit, keys);
 474     clang_visitChildren(clang_getTranslationUnitCursor(unit), &CYChildVisit, &baton);
 475
 476     for (CYKeyMap::const_iterator key(keys.begin()); key != keys.end(); ++key) {
 477         std::string value(key->second);
 478         for (size_t i(0), e(value.size()); i != e; ++i)
 479             if (value[i] <= 0 || value[i] >= 0x7f || value[i] == '\n')
 480                 goto skip;
 481         std::cout << key->first << "|\"" << value << "\"" << std::endl;
 482     skip:; }
 483
 484     clang_disposeTranslationUnit(unit);
 485     clang_disposeIndex(index);
 486
 487     return 0;
 488 }